Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Identifying possible bladder cancer ocupational carcinogens via a case-control study and JEM 2004

You don't seem to have a PDF reader installed, try download the pdf

Item Metadata

Download

Media
ubc_2004-0610.pdf
ubc_2004-0610.pdf [ 18.27MB ]
Metadata
JSON: 1.0092319.json
JSON-LD: 1.0092319+ld.json
RDF/XML (Pretty): 1.0092319.xml
RDF/JSON: 1.0092319+rdf.json
Turtle: 1.0092319+rdf-turtle.txt
N-Triples: 1.0092319+rdf-ntriples.txt
Citation
1.0092319.ris

Full Text

Identifying  Possible Bladder Cancer Occupational Carcinogens via a Case-Control Study and JEM by Kathryn Jane Richardson B.S.c, University of  Warwick, 2000 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF Master of  Science in THE FACULTY OF GRADUATE STUDIES (Department of  Statistics) We accept this thesis as conforming to the required standard The University of  British Columbia August 2004 © Kathryn Jane Richardson, 2004 THE UNIVERSITY OF BRITISH COLUMBIA FACULTY OF GRADUATE STUDIES B Library Authorization In presenting this thesis in partial fulfillment  of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference  and study. 1 further  agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. m P H f t V N  ^TA i g i C H / V £ P S c s ^ Name of Author (please  print) 2S/O  S /  3lOO Lj- Date (dd/mm/yyyy) Title of Thesis: /CxznSTi  f^SfrJG,  f^OSSJiS  L.EL CAk  / V ^ / Q •f—rTLx.Q'W  ZS^r m Degree: f U S g S T 7 T T ) s T T l Year: g L Q Q Department of ^ T P T r x J E ^ Y ^ The University of British Columbia Vancouver, BC Canada grad.ubc.ca/forms/?formlD=THS page 1 of 1 last  updated:  20-Jul-04 Abstract A significant  proportion of  cancer development is attributable to exposures to certain chemicals in the workplace. However, examining these occupational exposures is often  a difficult  and challenging task. In this thesis we use the relatively new approach of  applying an extensive JEM (Job Exposure Matrix) to estimate the occupational exposures of  1,062 bladder cancer cases and 8,057 matched other cancer controls. The subjects are all male, and were at least 20 years old and resident in British Columbia when diagnosed with cancer between 1983 and 1990. A self-administered  questionnaire provided the occupational histories and confounding  information  on the study subjects. The cumulative exposure (expected work-years of considerable exposure) to each of  11,132 occupational agents was estimated. The bladder cancer cases were matched to cancer controls of  other cancer sites (excluding lung) on age at diagnosis and year of  diagnosis. The analysis was performed  via conditional logistic regression and the following  confounders  were taken into account: ethnicity, who completed the questionnaire, smoking duration, and alcohol drinking status. Of  the 5,699 agents with at least 3 bladder cancer cases exposed, a significantly  increased (5% level) odds-ratio was seen for  ever exposure to 646 of  them. Of  the 3,450 agents with at least 9 bladder cancer cases exposed, 350 exhibited a significantly  (5% level) increasing dose-response relationship. After adjusting for  multiplicity, a subset of  30 agents was selected that demonstrated sufficient  evidence of  bladder carcinogeneity. Principal components analysis was performed  on the cumulative exposures of  these selected agents and 10 independent groups of  agents were identified.  The groups were mainly distinguished by job. The cumulative exposures to these 30 agents were mainly due to employment in logging, ship and boat building, and construction industries, and in occupations involving motor vehicles (e.g. gasoline service station attendant, mechanic, and truck driver). The selected 30 agents seem to mainly be of  petroleum or mineral oil base. Contents Abstract 1 1 Contents iii List of  Tables vii List of  Figures i x Acknowledgements x 1 Introduction 1 2 Methodology Review 3 2.1 Study Methods 3 2.1.1 Animal Experiments 3 2.1.2 Cohort Studies 4 2.1.3 Case-Control Studies 6 2.1.4 Proportionate Mortality Studies 7 2.1.5 Summary of  Study Designs 8 2.2 Further Issues with Case-Control Studies 8 2.2.1 Choice of  Cases and Controls 9 2.2.2 Bias 10 2.2.3 Confounding  11 2.2.4 Matching 11 2.2.5 Analysis of  Matched Case-Control Studies 12 2.3 Assessing Exposure 14 2.3.1 Cancer Biology 14 2.3.2 Measuring Exposure 15 2.3.3 Workplace Records , 15 2.3.4 Assessing Current Workplace 16 2.3.5 Self-Administered  Questionnaires 16 2.3.6 Personal Interviews 16 2.3.7 Job Exposure Matrix (JEM) 16 3 Risk Factors for  Bladder Cancer 18 3.1 Non-Occupational Bladder Cancer Risk Factors 18 3.2 Occupations Associated with Bladder Cancer 19 3.3 Occupational Bladder Carcinogens 20 4 Data Resources 23 4.1 The Study Area - British Columbia, Canada 23 4.2 BCCA Data 24 4.2.1 Data Editing 26 4.2.2 Analysis Inclusion Criteria 26 4.3 NIOSH JEM 27 4.3.1 Field Guidelines 28 4.3.2 Sampling Methodology 28 4.4 BCCA Canadian to US Job Translations 29 5 Exposure Assessment 31 5.1 Cumulative Exposure 31 5.2 Calculating Cumulative Exposure 33 5.2.1 Providing Adequate Job Information  33 5.2.2 Canadian Jobs and Latency 34 5.2.3 Coding the Canadian Jobs According to the CSOC/CSIC 34 5.2.4 Translating Canadian Jobs to US Equivalents 35 5.2.5 US Industries Studied by NIOSH 36 5.2.6 US Job-Translations on the JEM 36 5.2.7 Applying the JEM 37 5.2.8 Calculating Cumulative Exposure 37 5.2.9 Final Cumulative Exposure Estimates 38 6 Statistical Approach 39 6.1 Matching 39 6.2 Possible Confounder  Variables 40 6.2.1 Characteristics of  Cases and Controls 41 6.3 Developing the Base Model 42 6.3.1 The Base Model 43 6.3.2 Consequences of  the Missing Data Exclusions on the Base Model 43 6.4 Testing the Agents Individually 44 6.4.1 Agents with Exposed Cases 45 6.4.2 Ever/never 45 6.4.3 Dose-Response 45 6.4.4 Multiple Comparisons 46 6.5 Testing the Agents in Groups: Principal Components Analysis 48 6.5.1 Principal Component Analysis 48 6.5.2 Grouping of  Agents and Analysis Approach 49 7 Results 51 7.1 Individual Agent Results 51 7.1.1 Ever/never 51 7.1.2 Dose-Response 52 7.2 Selecting Significant  Associations 52 7.2.1 Linear Exposure 53 7.3 Grouped Agent Results: Principal Components Analysis 54 7.3.1 Component Groups - Dose-Response 54 7.3.2 Component Groups - Any Exposure 54 7.3.3 Component Groups - All Exposure 55 7.4 Properties of  the Selected Agents 55 7.4.1 Discussion of  the Selected Agents 56 7.5 IARC and Siemiatycki Results Comparison 59 8 Discussion 61 8.1 Summary of  Findings 61 8.2 Bias and Confounding  62 8.2.1 Comparability of  Source Populations for  Cases and Controls 62 8.2.2 Selection Bias 63 8.2.3 Information  Bias 63 8.2.4 Confounding  by Non-Occupational Variables 64 8.2.5 Confounding  by Occupational Exposure 64 8.3 Evaluation of  Methods v 64 8.3.1 Ever/never 0.5 65 8.3.2 Siemiatycki Comparison 65 8.4 Future Directions 67 9 Bibliography 6 8 10 Tables and Figures 72 Appendix A Appendix B 155 List of  Tables 2.1 Two by Two Contingency Table for  Cohort Studies 72 2.2 Two by Two Contingency Table for  Case-Control Studies 72 3.1 Occupations Associated with Bladder Cancer 73 3.2 Chemicals Associated with Bladder Cancer 74 3.3 Siemiatycki Chemicals Associated with Bladder Cancer 75 3.4 Siemiatycki Exposure Coding 75 5.1 Distribution of  Missing Job Code, and Start and End Year Data 77 5.2 Cancer Site Distribution of  Jobs in the Armed Forces 78 5.3 Distribution of  Patients' Canadian Industries 79 5.4 Distribution of  Patients' US Industry Translations 80 5.5 Distribution of  Patients' Canadian Occupations 81 5.6 Distribution of  Patients' US Occupation Translations 82 5.7 Distribution of  Patients' Industries and those on the JEM 83 5.8 Distribution of  Patients' Occupations and those on the JEM 84 5.9 Location of  Canadian Jobs Found on the JEM 85 6.1 Characterisitics of  Cases and Controls 86 6.2 Odds Ratios (OR) for  Potentially Important0 Confounding  Variables 87 6.3 Log Likelihood for  Various Base Models 87 6.4 Characterisitics of  Cases and Controls Before  and After  Exclusions 88 6.5 Distribution of  Control Cancer Sites Before  and After  Exclusions 89 6.6 Odds Ratios (OR) for  Potentially Important" Confounding  Variables Before  and After  Exclusions 90 6.7 Distribution of  Bladder Cases Exposed Across the 8,986 Agents 90 7.1 Distribution of  p-values for  Ever Exposure of  5,699 Agents 91 7.2 Agents Significant  After  Adjusting for  Multiplicity Using the Hochberg and Benjamini Pro- cedure from  5,699 91 7.3 Distribution of  p-values Across 3,450 Agents Tested For Dose-Response 92 7.4 Number of  Agents With Significant  Ever Exposure and Ordinal Trend Results 92 7.5 Selected 30 Agents with Significant  Associations 93 7.6 Results for  Linear Fit of  Transformed"  Cumulative Exposure for  Top 30 Agents 98 7.7 Component Scores0 for  PCA of  the 30 Selected Agents 99 7.8 Dose-Response Results for  Component Groups 100 7.9 Multivariate Ordinal Results0 for  Component Groups 100 7.10 Results for  Ever Exposure to Any of  the Members of  each Component Group 100 7.11 Multivariate Any Results0 for  Component Groups 101 7.12 Results for  Ever Exposure to All of  the Members of  each Component Group 101 7.13 Multivariate All Results0 for  Component Groups 101 7.14 Properties of  the Selected 30 Agents 103 7.15 Chemicals Associated with Bladder Cancer with Study Odds Ratios 104 7.16 Siemiatycki Chemicals Associated with Bladder Cancer with Study Odds Ratios 106 A.l Odds-Ratios of  Ever Exposure and Dose-Response Results for  3,450 Agents0 108 B.l NIOSH Agent Name Abbreviations 156 List of  Figures 5.1 Calculating Cumulative Exposure Analysis Design 76 7.1 Histograms of  Positive Cumulative Exposures for  Top 30 Agents 94 7.2 Histograms of  Transformed  Positive Cumulative Exposures for  Top 30 Agents 96 7.3 Proportion of  Cumulative Exposure Due to Employment in Each US Job for  the Top 30 Agents 102 Acknowledgements I would first  like to thank Dr. Nhu Le from  the British Columbia Cancer Agency. Nhu provided the idea for  this thesis and much appreciated support and encouragement as my supervisor throughout its duration. I would also like to thank my second reader, Dr. Paul Gustafson,  for  his useful  a'dvice and comments. In addition I would like to thank my colleagues at the BC Cancer Agency for  their support and help, in particular, Amy MacArthur, Sandy Shen, Barbara Lang, and Melissa Friesen. Finally, a huge thank-you goes to my boyfriend,  Andy Parker, for  all his support and patience throughout my entire master's degree. KATHRYN J A N E RICHARDSON The  University  of  British Columbia August 2004 Chapter 1 Introduction Occupational exposures to certain hazards have long been recognised as possible health concerns. During the 18th century concerns arose after  physicians noted debilitating and fatal  conditions occurring preferentially among workers in certain types of  jobs. Percival Pott provided the first  unambiguous evidence of  chemical carcinogenesis from  occupational exposure in 1775. He identified  soot as the cause of  scrotal cancer in London chimney sweeps based on clinical observations. Subsequently, associations between coal tar product exposures and cancer development have been seen in many studies. Cancer development is not fully  understood and a better appreciation of  the factors  involved is clearly important. The future  risk to the general population of  particular occupational exposures can be projected, and public health policies can be steered towards minimising this risk through for  example, workplace regulations. Occupational exposures are particularly important as many workers are exposed similarly at the same time and over a significant  portion of  their lifetime.  Large groups of  workers with similar exposures makes studying the exposures simpler, and it is also relatively straightforward  to minimise and regulate exposures to protect the workers. It is difficult  to estimate the percentage of  cancers attributable to occupational exposure. Estimates range from  1% to around 40% (Siemiatycki, 1991). The proportion of male bladder cancers in the United States attributed to occupational exposures is estimated as 10% by Doll and Peto (1981), and around 21% to 25% in white males by Silverman et al. (1989). Also, about half  of  all known carcinogens are primarily industrial chemicals (Tomatis, 1990). The main approaches to identifying  occupational carcinogens are introduced in section 2. Animal studies are scientifically  valid, however, the results cannot always be applied to humans. Epidemiological studies are more commonly used to investigate the complex effects  of  occupational exposure to cancer development in humans. However these studies are difficult  to implement when the disease is rare and has many complex and interrelated causes. The most widely used approach is that of  a case-control study as introduced in section 2.1.3 and discussed further  in section 2.2. Exposure to a carcinogen may contribute to the initiation of  tumour development, or it may hasten the onset of  a tumour. Nevertheless, exposure to a carcinogen does not usually make cancer inevitable. Carcinogenesis and cancer biology are introduced in section 2.3. As cancer has a long induction period between exposure and manifestation,  exposures need to be considered over most of  a subjects lifetime. Various methods to estimate these lifetime  occupational exposures are also described in section 2.3. Current research into bladder cancer carcinogens is discussed in section 3. Many animal studies have been undertaken, but conclusive evidence in humans is rare and further  research is required. The approach taken in this thesis to identify  possible bladder cancer occupational carcinogens is via a case-control study of  bladder cancer cases and cancer controls identified  in a BC Cancer Agency (BCCA) study. The approach taken could be repeated for  other cancer sites of  interest; however, bladder is the cancer site of  interest here. A Job Exposure Matrix (JEM) developed in the US was thought most appropriate to estimate the lifetime  occupational exposures to a range of  agents of  the cases and controls. Conditional logistic regression was implemented to estimate the risk of  exposure to the occupational agents. Principal component analysis was undertaken to find  groups of  agents that act synergistically to increase the risk of bladder cancer. The BCCA study used for  the case-control data is described in section 4.2 and the JEM used is introduced in section 4.3. The exposure assessment calculations are described in section 5. The statistical approach taken is detailed further  in section 6, and the overall results are given in section 7. Finally, a discussion of  the whole procedure is presented in section 8. Chapter 2 Methodology Review The goal of  this thesis is to identify  possible occupational carcinogens for  bladder cancer. There are many methods to try to accomplish this. The main study methods are discussed next in section 2.1. This thesis looks at this problem from  the perspective of  an epidemiological case-control study. Some important issues surrounding case-control studies are discussed in section 2.2. The ideas of  carcinogenic exposure are discussed in section 2.3. 2.1 Study Methods Many types of  study can be conducted to try and identify  possible occupational carcinogens. Ideally an experiment is performed  to see if  particular exposures give rise to cancer in humans where only the exposure varies between subjects. Firstly, this type of  experiment would be unethical in humans. Secondly, it would be near impossible to keep all other factors  constant among subjects. Thirdly, chronic diseases have complex etiology and a long study period would be required post exposure before  the disease manifested  itself.  Sim- ilar experiments can be undertaken in animals, as described in section 2.1.1, but the results cannot always be extended to humans. These limitations confine  most etiologic research to non-experimental epidemio- logic varieties. Epidemiological studies are designed to reduce variation from  extraneous factors  other than those under study. The non-experimental epidemiological studies of  cohort, case-control, and proportional mortality are described in sections 2.1.2, 2.1.3 and 2.1.4 respectively. 2.1.1 Animal Experiments Controlled scientific  experiments can be carried out on small animals such as rats and mice to investigate whether certain exposures lead to cancer development. Variation from  other factors  between animals can be kept minimal by the investigator. In this way hypotheses about particular potential carcinogens can be tested directly. Humans are genetically very similar to rodents; however, the results from  animal experiments cannot be applied to humans with complete confidence.  The animal studies are designed to test for  carcinogenicity of  the substance in that particular animal, not to emulate the human experience. The doses administered, routes of  exposure, lifestyle  maintained, and induction periods in animal studies are unrealistic compared to the human experience (Siemiatycki, 1991). The sensitivity of  detecting human carcinogens from  animal experiments is quite high, however at the expense of  the specificity.  Most identified  human carcinogens show carcinogenic activity in animal experiments, and there is often  correlation between the target organs affected  and carcinogenic potency (Siemiatycki, 1991). However, for  most carcinogens found  in animal experiments, equivalent associations have not been seen in human studies. Ashby and Tennant (1988) found  a relatively low correlation between those carcinogens found  from  experiments on rats and those on mice, suggesting weak ability to extrapolate from  rodents to humans. Many studies into potential carcinogens are performed  via animal experiments, as the experiments are relatively quick and inexpensive to perform,  and can be easily controlled. Experimental animal data on carcinogenicity exists for  many substances, while human study data exists for  relatively few.  From the point of  view of  deciding public health policies, the animal data cannot be ignored. The International Agency for  Research on Cancer (IARC) is the worlds leading authority on assessing evidence for  carcinogenicity of substances in humans. The IARC recommends that when there is limited evidence of  carcinogenicity in humans and sufficient  evidence of  carcinogenicity in experimental animals, the agent (mixture) be classified as probably carcinogenic to humans (IARC, 1987). 2.1.2 Cohort Studies In a cohort study design subjects free  of  disease are selected into groups, or cohorts, according to their exposure to a suspected cause of  the disease. The cohorts and then followed  over time and the rate of disease compared within each group. Cohort studies, among all of  the epidemiologic study designs, are the most accepted by the scientific  community, as they mimic the scientific  trial by observing disease in different exposure groups (Checkoway, 1989). The cohorts are selected independently to disease outcome, and the sequence of  events follows  naturally from  suspected cause to effect. The simplest type of  cohort study occurs when a group exposed to the suspect hazard, and a group not exposed to the hazard are studied for  the same period of  time. The amount of  diseased subjects is observed in each group. Table 2.1 displays the results for  this simple cohort design. Suppose Aj subjects develop the disease in the exposed group and Ao subjects develop the disease in the non-exposed group. Also, B\ subjects are not diseased at the end of  the study in the exposed group and Bq subjects are not diseased in the non-exposed group. So the original exposed cohort consists of  subjects Ai + B\, and the not exposed cohort consists of  subjects Aq + Bq. The overall measure of  risk in the cohort study is calculated as the relative risk (RR). The relative risk is simply the ratio of  the probability of  disease in the exposed to the probability of  disease in the non-exposed. Using the notation in table 2.1, the relative risk is calculated as: - AMo  + BQ) A0(A 1+B1) The relative risk will equal 1 when there is no difference  in the risk of  developing the disease between the exposed and the non-exposed. The relative risk will be greater than one when the risk of  developing the disease is greater for  those exposed. When the non-exposed are at greater risk, the relative risk will lie between 0 and 1. The relative risk cannot be negative as it is a ratio of  probabilities. Other factors  affecting the outcome and exposure, called confounders,  can be incorporated in the analysis and the resulting relative risk adjusted for  these. Also the analysis and thus relative risk can account for  the usual situation where subjects are followed  for  differing  periods of  time by comparing rates per person-years of  follow-up  rather than rates per person. Rates can also be compared to those in the general population, to see if  there is any increased risk in the particular cohort (Checkoway, 1989). However, the members of  occupational cohorts may differ  from  the general population in more respects than just the exposure. Cohort studies can either have a prospective or retrospective design. The intuitive design is prospec- tive where the cohorts are selected and then follow-up  proceeds into the future.  In this way, the exposure estimates and estimates of  other potential confounders  can be accurately measured directly. Diseases are often  rare in a population, meaning it would take time to see enough diseased cases develop in a cohort. This is even truer for  chronic diseases, such as cancers. Either a lengthy study period or a large cohort is required due to their long induction and latent periods. The prospective cohort study then becomes a costly and timely exercise (Checkoway, 1989). A common approach to minimise the cost and time commitments in a prospective cohort study is to perform  a retrospective (historical) cohort study. Here a cohort is enumerated as starting some time in the past, and follow-up  is observed until the present time. The difficulty  now comes in estimating exposures and confounding  factors  for  cohort members retrospectively in time (Checkoway, 1989). This information may not be available or complete, and will certainly be less accurate than in the prospective cohort study. Issues surrounding estimating exposure retrospectively are discussed in section 2.3. 2.1.3 Case-Control Studies A case-control study is not as intuitively appealing as a cohort study, but is more efficient.  A cohort study requires obtaining exposure data on a large number of  subjects of  which only a small proportion typically develop the disease (Checkoway, 1989). A case-control study gains efficiency  by sampling only a small proportion of  those that do not develop the disease. For a case-control study, one identifies  a representative group of  subjects with the disease, the cases, and a comparable group of  subjects, but that are not diagnosed with the disease, the controls. The exposure histories and confounding  factors  are then sought as in a retrospective cohort study, but this time for  the case and control groups. If  the cases were significantly  more exposed to a particular hazard than the controls, given that all other confounding  factors are taken into account, one could infer  that the hazard was associated with the occurrence of  the disease. Table 2.2 depicts the results of  the simplest type of  case-control study. Of  the cases, a j denotes the number exposed to a particular substance, and bi denotes the amount of  controls exposed to the particular substance. So the total cases sampled is ai + ao, and the total controls sampled is 6i -I- bo. As the proportion of  diseased to non-diseased subjects is a feature  of  the case-control design rather than representing the true proportion, the formula  used to calculate the relative risk in the cohort study does not produce the true relative risk here. The denominators used in the relative risk formula  are unknown; therefore  a quantity called the odds-ratio (OR)  is calculated instead. The odds-ratio is approximately equal to the relative risk when the disease is rare in the general population. This is true for  cancer (Siemiatycki, 1991), so the odds-ratio is considered equal to the relative risk in this thesis. Again, we are interested in a departure from  one in the odds-ratio. The odds-ratio is calculated as the ratio of  the odds that a case was exposed to the odds that a control was exposed. Using the notation in table 2.2, it is calculated as follows: Q R _ a i /Qo _ aifto bi/b 0 a0bi The case and control groups can be considered as being drawn from  the same hypothetical population as the cohort study groups were drawn (Rothman & Greenland, 1998). However, most of  the diseased are sampled as cases for  the case-control study, whilst only a small subset of  the non-diseased are sampled as controls. When this truly happens and a case-control study is selected from  cohort at end-point, it is referred to as a nested case-control study. The nested case-control.study then benefits  from  being able to estimate exposures and confounders  more accurately (Siemiatycki, 1991). Case-control studies can provide as valid results as cohort studies, although controls have to be carefully  selected, and attention directed to avoid possible biases (Checkoway, 1989). The control group should form  a representative sample from  the population in which the cases were drawn. The control group must also be sampled independently of  exposure status (Rothman & Greenland, 1998). A population-based disease registry provides a good source of  possible cases when the registry contains accurate and up to date basic information  about all possible patients with the disease in the population (Checkoway, 1989). More details about case and control selection are described in section 2.2.1. Case-control studies are susceptible to more biases than cohort studies (Rothman & Greenland, 1998). Selection bias may occur when controls are selected differentially  to cases. For example, the controls may suffer  from  non-response bias. Recall bias may occur due to cases being more willing to provide better quality data than controls (Siemiatycki, 1991). Further discussion of  possible biases in case-control studies can be found  in section 2.2.2. 2.1.4 Proportionate Mortality Studies Proportionate mortality studies compare the proportional distributions of  causes of  death in a subgroup to those in a reference  population. Death certificates  often  contain information  on cause of  death and main occupation, or a workplace may keep records of  the death certificates  of  its former  employees. This method has the advantages of  being a relatively quick and inexpensive approach to gauging information  about disease excess in certain subgroup populations. It can also be useful  when the full  information  required for  a cohort or case-control study is incomplete (Checkoway, 1989). However, there are many limitations to this type of  crude analysis (Siemiatycki, 1991). Information on cause of  death may not be complete or accurate. The deaths recorded for  a particular subgroup may not be representative of  all deaths in that subgroup. Only limited information  is available on exposure history and possible confounder  variables. The proportional mortality ratio (PMR) of  the ratio of  deaths in the subgroup due to the disease of  interest used as the measure of  the effect  is influenced  by the proportions of deaths due to other diseases in the subgroup. Thus, if  lung cancer was particularly prominent in a worker population, then mortality due to bladder cancer may look proportionately low when in fact  there was a greater than expected number of  absolute deaths due to bladder cancer. Finally, when interested in disease incidence, the proportionate mortality studies investigating associations with mortality will not always be indicative of  incidence associations. 2.1.5 Summary of  Study Designs Epidemiological research applies directly to human beings. Animal experiments may be valid at testing hypotheses about animal carcinogens, but these results may not transfer  to humans. Observing apparent clusters of  excess disease in subgroups has motivated epidemiological research. Occupational Epidemiology study designs are similar in that they attempt to study a group's occupational and disease experience over time and sample from  it (Checkoway, 1989). Proportionate mortality studies are quick and simple, but the conclusions regarding cancer incidence associations with occupational hazards may be imprecise and misleading. A more formal  approach of  a case-control or cohort study is required. Cohort and case-control studies mainly differ  in whom they compare; cohort studies compare the exposed to the non-exposed, whilst case-control studies compare the diseased to the non-diseased. The cohort groups are followed  from  carcinogenic exposure to manifestation  of  the disease, whilst retrospective information  about possible exposure is sought for  case and control groups. Cohort studies are thus more appealing for  hypothesis testing, and case-control studies for  hypothesis generating (Siemiatycki, 1991). Cohort studies are advantageous when investigating specific  associations as a workforce  cohort is generally exposed to a narrow range of  occupational agents. Case-control studies can evaluate a range of  different exposures in a range of  different  occupations and industries (Breslow & Day, 1980). Cohort studies can also be prohibitively expensive and time consuming for  studying a rare chronic disease, such as bladder cancer. However, case-control studies are susceptible to biases such as selection and recall bias, and care must be taken when planning a study. Occupational exposure assessment is also more difficult,  and possible approaches are described in section 2.3. A case-control study is an appropriate design for  investigating associations between occupational exposures and cancer, as required in this predominantly hypothesis generating thesis. More details about performing  a case-control study are discussed in the next section. 2.2 Further Issues with Case-Control Studies As the choice of  cases and controls for  a case-control study is important this is discussed further  in section 2.2.1. As it is also important to be aware of  possible sources of  bias, these are described in section 2.2.2. Con- founders  are a special form  of  bias discussed in section 2.2.3. Possible methods of  controlling confounders  are introduced as section 2.2.4 discusses matching and section 2.2.5 discusses the analysis method of  conditional logistic regression. 2.2.1 Choice of  Cases and Controls Firstly, a clear source population for  the cases needs defining.  The source population need not be the whole population, and can be restricted to improve information  quality, control for  possible confounders, and facilitate  valid selection of  controls (Checkoway, 1989). The group of  cases should be a homogeneous etiological entity (Breslow & Day, 1980). To be sure all cases have the particular disease, the diagnosis should also be histologically confirmed  (dos Santos Silva, 1999). In a registry-based study, the case group usually consists of  all incident cases appearing in the registry during a specified  period of  time (Checkoway, 1989). The controls should then be selected from  the source population; the same population that gave rise to the cases. The controls should be selected independently from  exposure status. This is to prevent the controls being unrepresentative of  the source population with respect to exposure. A control should be at risk of  becoming a case. So, the specified  period of  time when a subject is eligible to become a case should be the time during which a subject is eligible to become a control (Rothman & Greenland, 1998). There are various sources of  possible controls: population, neighbour, hospital, or other disease controls. Controls can be sampled from  the associated subset of  the general population. This is the ideal situation; enabling the study results to be generalised to the subset of  the general population. However, the response rate may be low, making the results liable to selection bias. Also, those that respond may not provide as accurate information  as they might, introducing possible recall bias. An extension of  this method is to sample controls from  the neighbourhood where the cases live or friends  of  the cases. These people will be more similar to the cases, may be easier to contact and more willing to participate. However, they may be too similar to the cases, in that their exposure status is related to that of  the cases (Rothman & Greenland, 1998). When identifying  hospital-based cases, hospital controls have several advantages. They are generally easy to contact and are less likely to be lost to follow-up.  They are in a similar position to the cases, have more time and may be more willing to help, thus reducing recall bias. However, care must be taken to avoid selection bias. Many hospitalised patients' exposures will not be representative of  the source population's exposures. This bias can be minimised by restricting controls to those with a diagnosis not thought related to the exposure. Also, choosing controls with different  diagnoses will tend to balance out the effect  of  any introduced by a specific  disease (Rothman & Greenland, 1998). The considerations when choosing controls with other diseases are similar to those for  hospitalised- controls. This type of  control is often  chosen in registry-based studies, and in particular with cancers. 2.2.2 Bias The purpose of  the analysis of  a case-control study is to quantify  the associated risk each factor  under study has with the disease (Breslow & Day, 1980). The observed associations may however be affected  by bias, confounding  and random variation. These problems hinder internal validity and the first  aim is to minimise these effects  so the true associations can be estimated. There are two main types of  bias: selection bias and information  bias, which are discussed next. Confounding  is an issue much related to bias and is discussed in the next section. Random variation is an artefact  of  any study and its effects  can be minimised by increasing the study size. Selection bias can be introduced when the cases or controls do not form  a representative sample from  the source population (Rothman & Greenland, 1998). A common source of  this systematic error is non-response bias. When a considerable proportion of  the control group chooses not to participate in the study, this proportion may be different  in characteristics to those who do choose to participate (Gordis, 2000). The Healthy Worker Effect  is also a common selection bias, particularly to occupational cohort studies (Checkoway, 1989). If  there is a difference  in the proportion of  workers between study groups, then the groups may additionally differ  due to the Healthy Worker Effect.  Workers are known to be generally healthier than the rest of  the population, especially those who remain in employment. To minimise selection bias, the case and control groups should be chosen appropriately (Checkoway, 1989). Attempts should be made to increase response rates, e.g. by providing incentives for  compliance, not making compliance too time consuming, etc. If  there is non-response in a study, it should be investigated to see if  there are significant  differences  between the characteristics of  responders and non-responders. Information  bias consists of  misclassification  of  subjects in two ways: differential  or non-differential misclassification.  Non-differential  information  bias occurs when the cases and controls are equally likely to be misclassified  according to their exposure or disease status. This will tend to bias the effect  estimate towards the null (Checkoway, 1989). Differential  information  bias is of  greater concern. Here, the likelihood of  misclassification  differs between the cases and controls. The bias of  the effect  estimate can then occur in either direction from  the truth (Checkoway, 1989). Recall bias is an example of  differential  information  bias common in case-control studies. Case subjects will generally be willing to answer study questions to the best of  their knowledge and take time thinking through their exposure histories. Control subjects are generally less likely to do so, particularly if  they are population controls, as they have less interest in providing the most accurate information  to the study to improve the research in that area. In this way the controls will be subject to more misclassification  bias than the cases. Information  bias is minimised by ensuring as accurate data recall amongst cases and controls as possible (Rothman & Greenland, 1998). It is also important to investigate the magnitude of  the effect  of this bias in the study, such as by questionnaire validation. 2.2.3 Confounding A confounder  is a factor  associated with exposure and with the disease, but it is not a step in the causal pathway from  exposure to disease (Checkoway, 1989). Common confounders  are age, gender, ethnicity and smoking habits. Distortion caused by confounding  factors  can lead to an overestimation or underestimation of the true effect  of  the exposure under study (Rothman & Greenland, 1998). Failure to control for  confounders can lead to a biased estimate. Mistakenly controlling non-confounders  can reduce the precision of  an estimate. Misclassification  of  confounders  can also reduce the ability to control for  confounding  (Checkoway, 1989). Confounders  can be controlled in the study design, in the analysis, or both (Checkoway, 1989). The source population, from  which the cases and controls were sampled, can be restricted, thus reducing any possible confounding.  However, this will reduce the sample size of  cases and controls. Confounders  can be controlled via matching the cases and controls on the main confounders.  This will help to optimise the efficiency  of  the analysis and improve the precision of  the effect  estimates. Further details about matching are discussed in section 2.2.4. The analysis can also simultaneously control for  potential confounders  by including the potential confounder  variables in the logistic regression model. Confounders  are identified  by previous studies, biological knowledge of  the disease, and known fea- tures of  the study design. In order to assess potential confounders  accurately, reliable information  is required on them from  the cases and control subjects. 2.2.4 Matching Matching is a method used to attempt to control for  the most important confounders.  Individual matching involves pairing one or more controls to each case with respect to levels of  the matching factors.  Frequency matching involves selecting a set of  controls to each group of  cases within a stratum of  matching factor  values (Rothman & Greenland, 1998). Matching attempts to make the distribution of  the matching factors  for  the control group more similar to that of  the case group. In this way, the method improves statistical efficiency  and increases the precision of  the effect  estimates (Checkoway, 1989). However matching is costly; information  is lost, in that subjects are excluded from  the analysis that did not match. Also, the effect  estimate of  the matched factors  can no longer be estimated as their relationship with disease has been distorted. After  matching, the factors  matched upon may still be confounders  or may introduce selection bias that also needs controlling for  in the analysis (Rothman and Greenland, 1998). Care should be taken not to overmatch so the cases are too similar to the controls apart from  on the exposure under study. Matching on a variable associated with exposure but not disease harms statistical efficiency.  Matching on an intermediate between exposure and disease may harm the study validity and result in an effect  estimate biased towards the null. A third type of  overmatching using convenient controls may harm cost efficiency  or introduce bias (Rothman and Greenland, 1998). 2.2.5 Analysis of  Matched Case-Control Studies Logistic regression is the usual technique used to analyse case-control studies and allow for  confounders. However, when the data is matched, then prior information  is known about the distribution of  patients across strata, and conditional logistic regression must be used to account for  this. Breslow and Day (1980) show that when using logistic regression for  a simple matched study design where each case is matched to one control and there is one covariate, the effect  estimate is biased by 100%. For a subject's covariate vector x, the logistic regression model for  the probability distribution of the binary dependent variable Y,  is: ea+0'x P(y  = lb) = where a is the intercept parameter and the fl's  are the covariate parameters. So if  Y  is the indicator variable representing a case with value 1 and a control with value 0, then the above model estimates the probability that a subject with covariate vector x is a case. The parameter coefficients  are estimated via maximum likelihood estimation. Now a stratum-specific  logistic regression model can be specified  for  a matched case-control study with K  strata. Let nut denote the number of  cases and no^ denote the number of  controls in stratum k, k = 1 ,2 , . . . , K.  Thus the conditional logistic regression model is: pCtk+P'x p(y = n*) = l + e a k + 0 , x (2.1) where ak denotes the stratification  variable for  the fcth  stratum. The conditional likelihood for  the fcth  stratum reflects  the probability of  the observed data relative to the probability of  all possible configurations  of  the data amongst the cases and controls within that stratum. The number of  possible combinations of  the cases among all nk = n n + nofc  subjects, is denoted by Cfc  and given by the expression: Cfc nifc!(n fc +nlf c)! The subscript j denotes any one of  the Ck  assignments and i indexes the observed data and ij indexes the observed data for  the j th assignment. Subjects 1 to n n correspond to the cases and n\k + 1 to rifc  correspond to the controls. So, the conditional likelihood for  stratum k  can be written as: k [ P > E - ^ m ^ i n * : , , , \ y i j = i)nr/=„„+1 \ V i i = o n This conditional likelihood can be simplified  by applying Bayes theorem and substituting in the conditional logistic model (2.1): T-T™!* P{yi=\\xi)P(xi ) T-rn/c Pfai=0|xi)P(:Ci)  lli=l P(yi  = 1) lli=n lk + l P(yt=0 ) v-̂ c* m n l t /''(j/.,-! n k P(Vi.=0|x3i.)P(s3..) 2^j = lilli,=l P(yi,=l) llij=nlf c + l P(y is=0) J h(P) > /  a\ '•'•1=1  i-(y i = i)  » « = » i i , t i P(y;=0) lk(P)  = p7— fn nik  ""j' . P(v  ij=l) llij=nlf c + l P(yij  =o) n?=i p(yi = ii xi)p{Xi) nr=n lfc +i nvi = 01*0^*0 E ^ i t n ^ i p{vi, - i i ^ ) ^ ( ^ ) nr;=„ lfc +i ^ = o i ^ ) ^ ^ . ) } nr=i p(xi)n?±\ p ^ = = nn f c PC-r TT"fc 1 1=1 ^V t̂; l i i = 1 l+e„k+fi'i: i lli = nlf c + l 1 + e«fc+/3'a \pCfc  rrrrn P/V . \ n n i f c e J TTra>: 1 1 Z ĵ = l \ l l i J = l ^K-Ljij)  lUj  = l 1+e"k+f>'*ji j llij=nlf c+l J rrnf c pct-'i n n i f c pQfc+/ 3'3;i nn f c 1 , _ 11»=1 ^ W llt = l e llt=Wlfc  + l 1+ec.k+P'* j '  ~ m n t p(t  -sI rrnif c pOk+f' 1).; n n t 1 . i 2_J = 1 1.11̂ =1 ^K-Ljljl LUj=l e lli^mt+l at+P'vjij J J"|«ifc eak+0'xi h(P)  ^ J—jTiifĉ  eak+p'x ji j r rUk p'xi HP)  = ^ ^ w (2.2) The likelihood function  is then the product of  the tk(P)  in (2-2) over all K  strata: K K rrnifc  p/3'xi m = n m = n (2-3) ib=i fc=i  2_j=i llij=i e The conditional likelihood estimators for  the f3  parameters are those values that maximise the like- lihood in equation (2.3). Statistical software  packages, such as SAS, are able to estimate these parameters for  matched conditional logistic regression models. 2.3 Assessing Exposure To perform  a case-control study, exposure measurements to the suspected cause of  disease are required for  each subject. Firstly, the biological meaning of  exposure is described and how it may affect  cancer development in section 2.3.1. There are many different  variables that attempt to model this biological exposure as explained in section 2.3.2. Clear and unambiguous evidence for  exposure to potential carcinogens is rare. There are then many approaches to collect the surrogate exposure information  which are described in the following  sections: workplace records, assessing the current workplace, self-administered  questionnaires, personal interviews and expert estimation, or via a JEM. 2.3.1 Cancer Biology Cancer is essentially the term for  a group of  diseases characterised by a malignant growth of  cells. The first  stage of  cancer is initiation. This is when a critical gene is irreversibly mutated. Gene mutations may be caused by ageing, exposure to chemicals, radiation, hormones or other factors  within the body and the environment. A promoter is then required to encourage the cells to reproduce and grow. The promoter acts after  the initiator and acts regularly over a certain period of  time. After  this a tumour may be visible and invasion and metastases to new body sites can occur. The time between cancer occurrence and detection is called the latent period (Rothman & Greenland, 1998). It is uncertain what causes cancer initiation and promotion, and when in a given subject (Siemiatycki, 1991). Also, different  factors  can make a subject more susceptible to the effects  of  carcinogens, e.g. age, gender, genes, health status, and diet. The initiator may be a carcinogen over the biologically effective  dose, or a genetic pre-disposition. Therefore  people exposed to an initiating carcinogen at a higher dose over a longer time than others will be more likely to develop cancer. Also those exposed to a promoting carcinogen regularly at a higher dose over a longer time than others will be more likely to develop cancer. Duration of exposure is an important factor  when analysing the effects  of  a carcinogen. For a subject with cancer, it is difficult  to know what contributed to the initiation and promotion of  cancer and when. As cancer is a chronic disease, there is also a large time period during which these exposures could potentially have happened. Thus, much of  the subject's lifetime  exposure needs to be estimated (Siemiatycki, 1991). An exception is the latency period, but this period varies between subjects. An approach called exposure lagging can be used to allow for  the latency period (Rothman & Greenland, 1998). As exposures close to cancer diagnosis are unlikely to have contributed to cancer development, only exposures preceding a certain cut-off  time before  cancer diagnosis are estimated. As the latency period is unknown, analysis can be performed  for  a range of  cut-off  values. 2.3.2 Measuring Exposure fdeally  the exposure to each possible carcinogen, in terms of  frequency  and dose, over each subject's lifetime can be measured. The dose is the actual amount of  substance that reaches the biological target (Checkoway, 1989). Given the same exposure to a carcinogen, the dose may vary between people as it depends upon many factors  such as inhalation rate, health status, genetics, age, gender, etc. If  the dose and exposure are related, then the exposure can be measured as a surrogate. Exposure measurement of  carcinogens is comprised of  concentration, duration and frequency  (Check- oway, 1989). The concentration of  a carcinogen is important, as lower concentrations are less likely to con- tribute to cancer development. The frequency  of  exposure to promoting carcinogens is also important, as occasional exposure is less likely to encourage cancer growth. Exposure to a promoting carcinogen over a longer period of  time is also more likely to result in cancer growth. Note that these measurements may vary over time. Subjects may be exposed to carcinogens at a differing  concentration and frequency  as their specific job tasks change. Also, legislation or technical improvements may change the concentration or frequency  of exposures over time. A time measure of  exposure can be constructed, such as cumulative exposure, which aggregates the concentrations over time, ft  is also useful  to collect data on the nature of  the exposures, e.g. the route of  contact, and any behaviour that protects against the exposure. Accurate measurements of  exposure, especially past exposures, are difficult  to attain. Occupational exposures account for  a considerable proportion of  lifetime  exposures and the exposures can be adequately measured. The main methods of  estimating occupational exposures are by workplace records, assessing the current workplace, self-administered  questionnaires, interviewing the subject, and job exposure matrices (JEMs). 2.3.3 Workplace Records fdeally  consistent exposure information  to all possible carcinogens is available for  each study subject at all their past workplaces. Many workplaces maintain exposure records, but collection methods vary between workplaces (Checkoway, 1989). If  recent workplace exposures can be estimated, it will still be difficult  to find  data on early exposures. 2.3.4 Assessing Current Workplace Given all employers gave consent, current exposures can be consistently measured at all subjects' workplaces. This would be a timely and costly undertaking. Also, earlier workplaces in a subject's job history may no longer exist, and if  they do, practices and thus exposures may have altered considerably since the subject was employed there (Siemiatycki, 1991). 2.3.5 Self-Administered  Questionnaires Study subjects can be asked to complete questionnaires about their exposure history. Subjects will have difficulty  remembering exposures many years ago, and may not have understood their exposures well enough to describe them accurately. Questionnaire responses from  cases and controls are also liable to recall bias (Siemiatycki, 1991). Cases may be more willing to accurately recall past exposures than controls, as they are more interested in furthering  knowledge into their particular cancer. Also, sometimes information  can only be obtained from  proxy respondents (e.g. spouses, family  members), who will have even less recall ability. 2.3.6 Personal Interviews The subjects themselves may have little knowledge of  their lifetime  occupational exposures, so some other approach is required. One method is to interview subjects and ask them to recall any known exposures or describe their occupations and workplaces in detail (Siemiatycki, 1991). Interviewing subjects may improve their recall ability. The interviewer can also ask additional questions to find  more accurate details about important exposures. However, interview bias may be introduced, as interviewers may have pre-conceptions that influence  the subject's responses. Many interviewers are usually required, which introduces extra variability between the interviewers' results. The exposure data can be improved upon by using the knowledge of  technical experts (e.g. chemists, engineers, hygienists) and also workplace records. The technical experts may be able to estimate exposures backward in time given current estimates, although the exposure estimates will be lacking in precision. 2.3.7 Job Exposure Matrix (JEM) Another approach used to assess exposures is to ascertain the occupational histories of  the subjects, either via a self-administered  questionnaire or interview, and then use a Job Exposure Matrix (JEM) to code the data. A JEM is typically a matrix containing all possible occupations as one dimension, and all possible occupational agents as the other dimension. One can then look up a particular agent for  each subject's job and find,  depending on the JEM, either an indication, yes or no, whether the subject should have been exposed to the particular agent in that job, or more precisely, an estimate of  the probability that they were exposed. Often  the exposure is only measured above some pre-defined  concentration or frequency. A problem with JEMs is that they are very costly and timely to complete accurately. Often  a study will use a pre-developed JEM, which may not be relevant to the particular group of  subjects in the study (Siemiatycki, 1991). A downfall  to the JEM is that it can only distinguish between exposures up to the level of  the job code (Siemiatycki, 1991). Individuals with the same job are assumed to have the same exposure and no further  allowance is made for  within job variability. Exposures in a certain job vary between workers depending on the particular work habits of  the worker, the workers specific  tasks, and the company the worker is employed with. However, exposures will be much more similar within the same industry and occupation, than between different  industries and occupations. Occupational exposures vary across time as employment practices and safety  regulations change and technology improves. A time dimension can be added, but JEMs commonly only represent one economy type at one particular moment in time. Chapter 3 Risk Factors for  Bladder Cancer Bladder cancer affects  the inner lining of  the bladder and develops slowly. As it grows, it may spread to other organs near the bladder. Carcinogenic chemicals are absorbed by the blood and filtered  into the urine, which accumulates in the bladder. There were 4,841 new cases (3,636 male, 1,205 female)  of  bladder cancer diagnosed during 2000 in Canada. There were 1,082 male deaths and 437 female  deaths due to bladder cancer. It was the fourth  most common cancer in men in terms of  incidence (5% of  cancers), and the ninth most common in terms of  mortality (3% of  cancer mortality). It was the thirteenth most common cancer in terms of  incidence (2% of  cancers) and mortality (1% of  cancer mortality) for  women (National Cancer fnstitute  of  Canada, 2004). Internationally, the highest incidences of  bladder cancer occur in Western Europe and North America (Schottenfeld  and Fraumeni, 1996). Non-occupational bladder cancer risk factors  are introduced in section 3.1. Section 3.2 describes the occupations with consistent elevated risks of  bladder cancer and section 3.3 describes the specific  occupational chemicals associated with bladder cancer. 3.1 Non-Occupational Bladder Cancer Risk Factors The risk of  developing bladder is increased in males, those of  Caucasian ethnicity and risk increases with age. About three times as many males develop bladder cancer as females  in Canada. The disease is more prevalent in Caucasians than other ethnic groups, with the incidence rate for  male Caucasians being at least double the incidence rate for  any other male ethnic group. Risk increases with age, with about two-thirds of bladder cancer cases occurring among persons aged 65 years and older (Schottenfeld  and Fraumeni, 1996). Lifestyle  factors  that have associations with increased risk of  bladder cancer include smoking, diets high in saturated fat,  coffee  drinking, artificial  sweeteners, drinking water quality and use of  hair dyes. Cigarette and tobacco smoking increase the risk of  bladder cancer. Smokers have around two to three times the risk of  non-smokers. Risk also increases for  increased intensity and duration of  smoking (Schottenfeld and Fraumeni, 1996). Cessation of  cigarette smoking has been associated with a 30% to 60% reduction in bladder cancer risk in many studies (IARC, 1986). Increased bladder cancer risks have been associated with diets high in saturated fat.  A possible association has been suggested between coffee  drinking and bladder cancer risk in case-control studies, but the findings  are inconsistent across studies. Artificial  sweeteners were suggested as potential human bladder carcinogens based on animal experiments, but epidemiological studies on humans have not substantiated the relation. Most studies that have evaluated alcohol consumption as a risk factor  for  bladder cancer have not supported a positive association. Epidemiological studies seem to support the association between chlorination by-product levels in drinking water sources and bladder cancer risk (Schottenfeld  and Fraumeni, 1996). Use of  hair dyes may be associated with bladder cancer risk as animal experiments indicate that some compounds in hair dyes are mutagens and people who dye their hair appear to excrete compounds in their urine. Results from  several epidemiological studies, however, have not supported this association (Hartge et al, 1982). Family or personal history of  bladder cancer increases the risk of  developing bladder cancer. Repeated or chronic bladder infections,  or bladder stones, slightly increase the risk of  developing bladder cancer. Changes occur in the bladder as a result of  repeated or persistent infection.  The excessive use of  drugs containing phenacetin and the use of  Cyclophosphamide and Chlornaphazine have also been linked to risk of  bladder cancer in many case-control studies (Schottenfeld  and Fraumeni, 1996). 3.2 Occupations Associated with Bladder Cancer The associations between certain occupations and bladder cancer risk are unclear. Studying large populations and recording personal occupational exposures accurately is difficult.  Some epidemiological studies have consistently found  associations with bladder cancer in certain occupations. However, these are often  based on small samples and observed relative risks have typically been less than two. Table 3.1 shows the main occupations associated with increased risk of  bladder cancer. This table was adapted from  Schottenfeld and Fraumeni (1996) to include the IARC monograph occupations classified  as having an association with bladder cancer and also lists some of  the agents suspected of  increasing the risk of  bladder cancer within each occupation. The IARC monographs are a series of  independent assessments of  animal and human studies into the carcinogenic risks posed to humans by a variety of  agents, mixtures and exposure circumstances. For many agents there only exists evidence of  carcinogenicity from  animal studies, for  some there are inconsistent epidemiological studies and for  many there is no evidence. Since its inception in 1969 to 2004, the IARC monographs have reviewed more than 885 agents. The IARC classifies  each agent according to its potential for  human carcinogenicity into one of  four  categories: 1 - definitely  carcinogenic, 2A - probably carcinogenic, 2B - possibly carcinogenic, 3 - not classifiable,  or 4 - probably not carcinogenic. Chemicals classified  as IARC group 3 are not necessarily non-carcinogenic, for  this classification  is where the available studies are of  insufficient  quality, consistency or statistical power to permit a conclusion. Further studies are required to aid the agent's IARC classification. Bladder cancer has well-established relationships with certain occupational exposures in human stud- ies. These include aromatic amine manufacturing  as 2-naphthylamine, benzidine, and 4-aminobiphenyl are considered definitely  carcinogenic to humans by IARC, the manufacture  of  certain dyes as auramine and magenta are considered definitely  carcinogenic to humans, rubber industries due to aromatic amines partic- ularly 2-naphthylamine, painters as paints contain aromatic amines as well as other carcinogenic agents, and aluminium and coke production and coal gasification  due to coal tar pitches considered definitely  carcinogenic to humans by IARC (IARC, 1987). There are also some occupations with a consistent excess risk of  bladder cancer for  which causative agents are only hypothesised. Dry cleaning solvent-exposed workers are potentially exposed to many chemi- cals and risks associated with bladder cancer have been seen in many human studies. Boot and shoe workers have an increased risk of  bladder cancer and leather workers have also been associated with risk of  bladder cancer, but the causal agents are uncertain between leather dust, dyes, benzene and solvents. Hairdressers and barbers have consistently seen an association with bladder cancer possibly due to hair dyes. Other occupations exposed to petroleum products including polycyclic aromatic hydrocarbons (PAHs) have seen excess bladder cancer incidence such as petroleum refinery  workers and truck drivers. Other occupations for  which evidence is limited are machinists, metal workers, chemical workers, textile workers, carpenters, construction workers, miners, mechanics, gas station attendants, medical workers, photographic workers, pulp and paper workers, and welders (Schottenfeld  and Fraumeni, 1996). 3.3 Occupational Bladder Carcinogens Studies into the association of  occupation and bladder cancer often  only suggest possible agents that may be contributing to the increased risk within the occupation. However, epidemiological studies aimed at examining the relation between the particular agent exposures and bladder cancer are uncommon. Much of  the evidence on chemical carcinogens is again based on animal studies. Table 3.2 summarises the agents considered related to bladder cancer risk in humans and includes their IARC classification  (if  any), whether the evidence is from  animal or human studies, and the results (if  any) from  a study by Siemiatycki, 1991. The agents the IARC classified  as definitely  carcinogenic to humans and have asssociations with bladder cancer are the aromatic amines 2-napthlamine, 4-aminobiphenyl and benzidine, magenta, arsenic, coal tar pitches, mineral oils, and drugs Cyclophosphamide and Chlornaphazin. There are a group of  chemicals for  which there is little evidence of  bladder carcinogenicity in humans, but evidence in animal studies. Bladder tumours have been seen in animals exposed to 1,3- dichloropropene, 2-(2-formylhydrazino)-4-(5-nitro-2-furyl)thiazole,  2-nitroanisole, 4-chloro-ortho- phenylene- diamine, benz(a)anthracene, CI basic red 9, citrus red no. 2, disperse blue 1, n-[4-(5-nitro-2-furyl)-2- thiazolyl]acetamide, niridazole, nitrilotriacetic acid, n-nitrosodi-n-butylamine, oil orange SS, ortho- aminoa- zotoluene, para-cresidine, para-dimethylaminobenzene, ponceau 3R, and sodium ortho-phenylphenate. The human bladder carcinogenicity of  these chemicals requires further  research. Table 3.3 summarises the results of  agents showing an association with blader cancer risk from  a study by Siemiatycki in 1991. Siemiatycki carried out an extensive study of  occupational carcinogens for  all major cancer sites. The case-control study consisted of  males aged 35 to 70, resident in the Montreal metropolitan area and diagnosed with a histological confirmed  cancer between September 1979 and June 1985. Analysis was performed  with bladder cancer patients (n = 484) and with cancer controls (n = 1,879) and with population controls (n = 533) separately. In addition, the analysis was repeated for  French Canadians only (around 60% of  the cases and controls). The occupational history of  study subjects was obtained through personal interviews and occupational exposures assessed by experts in occupational hygiene, epidemiology, engineering and chemistry. Experts assessed each subject's exposure by the concentration, frequency  and confidence  it occurred during the period the subject was employed. Each criterion was assessed according to a three-point scale as shown in table 3.4. Odds-ratios and 90% confidence  intervals were then calculated for  'any exposure' and 'substantial exposure' to each agent individually. The non-occupational confounders  accounted for  in the analysis were age (less than 55, 55 or older), family  income (tertiles), cigarette index (none, 1-800, 800+ cigarette-years), coffee  index (0-50, 51+ cup-years), and the respondent to the questionnaires and interviews (proxy, self).  The subject was considered to have 'any exposure' if  there was reasonable confidence (level 2 or 3) the exposure existed prior to 5 years before  diagnosis. The subject was considered to have 'substantial exposure' if  they had 'any exposure', concentration x frequency  > 3, and at least 5 years of exposure accumulated up to 5 years before  diagnosis. So, the exposure was probable or definite,  was above a background level, and occurred for  at least 5% of  the working time (i.e. no exposure criterion was coded as level 1) for  at least 5 years. The odds-ratios shown in table 3.3 for  those chemicals tested by Siemiatycki are the most significant out of  all 4 possible configurations  (cancer or population controls, all ethnicities or French Canadians) with at least 4 exposed cases. Siemiatycki found  significant  increases (at a 10% level) of  risk of  bladder cancer for  'substantial exposure' to cadmium compounds, carbon tetrachloride, diesel engine emissions, engine emissions, formaldehyde,  ammonia, asphalt (bitumen), chlorine, fabric  dust, laboratory products, natural gas combustion products, photographic products, polyester fibres,  and polyethylene. Significant  results at a 5% level with odds-ratios above 2 were found  for  'substantial exposure' to carbon tetrachloride (French, cancer controls), diesel engine emissions (French, population controls) and natural gas combustion products (all, cancer controls), and for  'any exposure' to acrylic fibres  (all, population controls), ionizing radiation (French, cancer controls), calcium carbonate (French, cancer controls), and titanium dioxide and titanium compounds (all, cancer controls). Chapter 4 Data Resources The BC Cancer Agency (BCCA) has collected occupational and lifestyle  data on male cancer patients resident in British Columbia (BC) and diagnosed between 1983 and 1990. The study area of  British Columbia, Canada is described in section 4.1 to give an overview of  the study subjects and their occupational exposures. The BCCA study is described in more detail in section 4.2. The National Institute for  Occupational Safety  and Health (NIOSH) has developed a Job Exposure Matrix (JEM) that gives exposure estimates to many agents in various US jobs. Using the translation system developed by BCCA, the Canadian job codes from  the BCCA questionnaire results can be translated to US job codes. These translations can then be used to estimate the lifetime  exposure to each agent for  each patient, and in turn these exposures can be used in the analysis of  the case-control study. The NIOSH JEM is described in section 4.3 and the translations introduced in section 4.4. 4.1 The Study Area - British Columbia, Canada The province of  British Columbia (BC) on the West Coast of  Canada has seen rapid popularisation over the last century. In 1931 the population of  BC was estimated to be a mere 0.7 million. By 1986, it had grown to 3 million. At the beginning of  the 20th century more than half  of  the population was under the age of  thirty and men outnumbered women nearly two to one. The percentage of  males and females  living in BC has been roughly equal since the 1960s. The population has also been ageing since the post World War If  baby boom. Most of  the population of  BC during the early to mid 1900s descended from  European origin. More recently the population has become more multi-cultural, including a significant  increase in the proportion of  Asian immigrants. In the 1986 census of  Canada, 60% of  British Columbians reported a single ethnic origin; nearly half  of  these had British ancestry, nearly 30% had other European ancestry, and 13% were Asian (Schrier and Ip, 1994). Over the last century BC's economy has been highly dependent on resource-based industries such as logging, the railway, mining, fishing  and agriculture. Manufacturing  activities were based on the processing of  natural resources, such as canning salmon, or producing lumber and paper. Resource-based industries continued to employ the largest share of  the labour population until the early 1990s. 4.2 BCCA Data The Health and The Workplace study of  the BC Cancer Agency (BCCA) provided the patient data for this project. The Health and The Workplace study was initiated in 1983 to provide a population-based occupational study of  male cancer patients resident in BC. Only male cancer patients were included in this study, as during the last century men were much more likely to have occupational exposure to carcinogens than women. Women also tended to spend longer periods of  time working in households, where exposures were too variable to be studied. The study consisted of  a self-administered  questionnaire mailed to all males aged 20 years and older diagnosed with cancer ascertained by BCCA during the period 1983 to 1990. The questionnaire requested detailed job descriptions for  up to 10 jobs held for  at least a year. It also requested information  on socio- demographic factors  and lifestyle  factors  such as drinking and smoking habits. Questions regarding lifetime consumption of  alcohol (wine, spirits, and beer) were initially omitted and then added during the fist  year of the study. Questionnaires were sent to all new cases for  the first  2 years. Subsequently, once 1000 completed questionnaires were returned for  a given cancer site, data collection was ceased for  that site. More information about the study can be found  in Band et al. (1999). The BCCA sent in total 25,726 questionnaires to eligible male cancer cases, of  which 15,463 were returned, giving a response rate of  60.1%. Patients aged over 80 were the only age group with a response rate below 50%. Also, liver, stomach, pancreas and unknown primary cancer sites had response rates below 50%. Response bias was addressed in the article by Band et al. (1999) and was thought to be minimal. Non-responders were not significantly  different  to the responders in education or smoking status. However, responders were more likely to be employed in the Managerial and Administrative occupational group than non-responders, and this was the only occupational group that was significantly  different  between the two groups (p < 0.001). These patients have low exposure to agents according to the NIOSH JEM, so perhaps the general cancer population has slightly greater occupational exposure than those responders included in the study. The questionnaire was originally sent to non-cancer regional controls, but the response rate was low. This would have introduced selection bias, with the non-cancer controls that responded being more interested in health research, and therefore  healthier. So, internal cancer controls of  different  sites to the case site were used as controls. This also minimises the recall bias, as the controls should be equally willing to provide complete responses in the questionnaire as the cases. This changes the interpretation of  the results however, and instead of  identifying  possible occupational carcinogens for  a particular cancer site per-se, carcinogens identified  are those more likely to contribute to cancer in that site rather than in other sites of  the body. The primary tumour site and diagnosis date information  was available for  all 15,463 patients through the BCCA Registry. Histological confirmation  of  diagnosis was obtained for  all patients. The primary tumour site was coded by the 9th revision of  the International Classification  of  Diseases (WHO) and grouped into 3-digit categories for  analysis. Misclassification  of  case or control status is very unlikely. The job descriptions given in the questionnaire were coded manually into an industry and an occupa- tion code according to the 1980 Canadian Standard Industrial Classification  (CSIC) and the 1980 Standard Occupational Classification  (CSOC), respectively (Statistics Canada, 1981). The CSIC consists of  853 pos- sible codes and the CSOC consists of  503 possible codes. The questionnaire also asked for  the location, start year, end year and duration of  each job, and included a tick box to indicate if  the job was part-time or seasonal. If  there was no indication in the tick box that the job was part-time or seasonal, the job was assumed to be full-time.  If  the duration of  the job was not given, this was taken as the end year minus the start year. Patients could provide information  for  a maximum of  10 jobs. So, it was assumed that if  a patient reported 10 jobs they had no further  jobs. Only 661 patients, or 4.7% of  those with jobs, recorded 10 jobs. This would not have too much effect  on exposure estimates, as any additional jobs may be of  short duration or occur close to the diagnosis date. The BCCA questionnaire was found  to be highly valid. For 81 patients who indicated working in one of  two large companies in BC, personnel records from  these companies were searched to check the starting year and duration of  employment of  these patients. The interclass correlation between the company records and questionnaires was 0.996 (95% confidence  interval, 0.993 to 0.997) for  starting year (excluding 2 patients with missing start date information)  and 0.971 (95% confidence  interval, 0.954 to 0.981) for duration of  employment (excluding 4 patients with missing duration of  employment and 3 patients that were not reported as being employed by the company records). A list of  all employees employed for  at least 3 years was also available from  one of  the companies. From the list of  all questionnaire respondents, there were no further  patients employed in that company than those that had reported so in the questionnaire. The questionnaire was also found  to be very reliable by comparing the responses of  87 patients who filled  out the questionnaire on two occasions. Here the kappa statistic was 0.94 for  smoking (ever/never) and the interclass correlation for  age started smoking and number of  cigarettes smoked per day were 0.92 and 0.81, respectively. The interclass correlation for  the number of  years worked in the most common occupations recorded (construction, farming,  clerical, and sales) were 0.92, 0.93, and 0.96, respectively; missing information  ranged from  2% (construction) to 14% (clerical and sales). When all occupational information  was recorded (80% of  the pairs), the interclass correlation was 0.92 for  work start year and 0.89 r for  total years worked (Band et al., 1999). There were other factors  thought related to cancer development that the questionnaire did not inquire about, such as coffee  drinking, diet, drinking water, use of  hair dyes, history of  cancer in the family,  genetic pre-dispositions, and overall health status. It would be difficult  to ask questions about such factors  without making the questionnaire very long. This should also not make much difference  to the analysis as information was sought on the most important risk factors. The estimation of  occupational exposures given all the occupational information  from  the patients' questionnaires is detailed in section 5. Firstly, the questionnaire data is edited and prepared for  analysis as described in section 4.2.1 and the inclusion criteria for  analysis is stipulated in section 4.2.2. 4.2.1 Data Editing The questionnaire data was edited for  errors and inconsistencies before  analysis. The protocol for  coding the initial questionnaire responses included entering the start year at age 12 if  the patient was raised or worked on a farm  from  birth. The questionnaire data still included jobs before  age 12, however, and 290 patients had jobs before  they were 12 and 230 patients had jobs before  they were 11. The patients with childhood jobs were reasonably randomly distributed across the cancer sites. Many of  these patients reported working or living on farms  during their childhood. Data is included for  those childhood jobs given so as to provide as accurate a summary of  personal lifetime  exposure as possible. Five patients had their first  job coded as starting before  they were born, so the birth dates were assumed correct as they came from  two sources (the questionnaire and the BCCA patient record), and the job dates were adjusted so the job started when the patient was born. 4.2.2 Analysis Inclusion Criteria The criteria for  the BCCA questionnaire mailing were BC males over 20 years old diagnosed with cancer between 1983-1990. The inclusion criteria for  the analysis were that the questionnaire was completed and the primary cancer site was known. There was only one patient who did not complete the questionnaire. Patients who completed only part of  the questionnaire are discussed in section 5.2.1. When the primary cancer site was unknown in a patient, the tumour will have a true site, but it was just undetectable at the time. Patients with unknown primary cancer sites were excluded, as they could not serve as controls for  other cancer sites. Some of  the primary unknown cancer sites could truly be the same site as the cancer case site. This would result in additional misclassification  errors amongst the cases and controls. Therefore,  708 patients with an unknown primary cancer site were excluded. This resulted in a total of  709 patients failing  to meet the inclusion criteria, leaving 14,754 patients for  analysis. Age criteria were also considered; such as excluding older patients due to questionnaire recall inability, a low questionnaire response rate and it being unlikely their cancer was due to occupational exposure. The latent period of  cancer is not known precisely, so some cancers in old age could be due to occupational exposure up to retirement, and also many patients worked after  retirement. There are few  old patients, and matching of  cases to controls on age later in the analysis will decrease their proportion further.  The exclusion of  younger patients was considered, as their cancers are also unlikely to be due to occupational exposures. Many younger patients were brought up on farms  and thus exposed to many agents. All patients aged 20 and above were included so the study results could be generalised to all BC male cancer patients aged 20 and over. 4.3 NIOSH JEM The JEM developed by the National Institute of  Occupational Safety  and Health (NIOSH) in the US was chosen for  the estimation of  exposure probabilities for  the BCCA data. A JEM based in North America was desired so the occupational exposures approximately represented those in BC and Canada. The NIOSH JEM constructed from  the National Occupational Exposure Survey (NOES) covered a broad range of  agents and estimated exposure probabilities from  actual measurements taken in a representative sample of  US workplaces. Some jobs were excluded from  the JEM, however, and the JEM does not measure the changes in exposure over time. Prom 1981 to 1983, NIOSH conducted the National Occupational Exposure Survey (NOES) to de- velop estimates of  the number of  workers potentially exposed to 12,945 chemical, physical, and biological agents in selected industries. Of  the 12,945 agents, 9,557 (74%) have corresponding Chemical Abstracts Service (CAS) codes, 4,952 (38%) have corresponding Registry of  Toxic Effects  of  Chemical Substances (RTECS) codes, allowing the substances to be compared across different  studies. Jobs were classified  ac- cording to the US 1980 Census of  Population Industrial Classification  (USCENfND),  and the US 1980 Census of  Population Occupational Classification  (USCENOCC). The US 1980 Census of  Population consists of  231 Industrial Classifications,  and 503 Occupational Classifications.  The NOES survey involved visits to 4,490 establishments in 121 industry groups (52%) employing approximately 1,800,000 workers in 377 occupational categories (75%). The field  guidelines and sampling methodology are discussed next, and further  details can be found  in NIOSH (1988) and NIOSH (1989) respectively. 4.3.1 Field Guidelines Specifically  trained surveyors collected exposure data via walk-through inspections of  each facility.  Exposure to an agent was only recorded if  the agent had been observed in sufficient  proximity to an employee so that one or more physical phases of  the agent were likely to enter or contact the body of  the employee, fn  addition, an employee was classified  as exposed to an agent if  the exposure occurred for  at least 30 minutes per week (on an annual average) or once per week for  90% of  the weeks of  work year. Thus, the JEM does not measure the level of  exposure, but exposure above a certain concentration and frequency.  This JEM exposure level can be thought of  'considerable exposure' and throughout this thesis shall be referred  simply as 'exposure'. The presence of  engineering controls over potential exposure was also recorded. The amount of employees exposed for  more than 4 hours a day or at least 90% of  the working year was recorded and defined as full-time  exposure. The exposures were classified  into trade name or actual agents. Approximately 70% of  the data collected were from  trade name products, and ingredients were determined for  85% of  these. 4.3.2 Sampling Methodology The target establishments were those within an industry on a list of  target USCENIND codes, located in the United States, and reporting 8 or more employees at the time of  the survey. Businesses with less than 8 employees were considered too numerous and transient to survey accurately. To construct a sample of  the target establishments a two-stage systematic selection procedure was employed involving stratification  by number of  employees, SfC  and geographical location. The first  stage of  the sampling procedure identified  establishments from  604 geographical combi- nations of  contiguous counties within metropolitan or urbanised areas. These were stratified  by employee concentration by USCENfND  code and geography into 98 strata. The second stage involved systematically selecting the 4,894 facilities  to be surveyed from  the strata by selecting independently across different  sizes of  facilities,  where the number of  employees defined  the size. A total of  4,490 facilities  co-operated with the study and were ultimately surveyed for  the NOES JEM. A downfall  of  the JEM is that the list of  target USCENIND codes excluded 110 (48%) industries. Many of  the employees in these industries were thought by NIOSH to have little agent exposure, so were not surveyed, e.g. finance,  insurance, and real estate. Some industries were thought to be so large and hetero- geneous that they warranted surveys of  their own, e.g. mining. While others, such as private households, were not surveyed as they were thought to be difficult  to survey accurately. Agricultural production, rail- road transportation, federal,  state and municipal government industries were also excluded from  the NOES survey. The final  NIOSH NOES JEM gives, for  each job (industry and occupation code) and agent, the ratio of  the expected number of  employees considerably exposed nationally, to the amount employed in that job nationally. The estimation of  the number of  employees exposed nationally, given the survey results involved weighting each survey facility  according to the probability of  including a facility  like it in the sample. The weightings were determined by ratio estimation, with ratio factors  determined using outside sources such as the Bureau of  the Census publication County Business Patterns (CBP), or the Dun Master Inventory (DMI). The amount employed in each job nationally was estimated via Duns Marketing Index (Dun and Bradstreet, 1980). The NIOSH JEM is essentially a 3-dimensional array with the USCENIND codes on one axis, the USCENOCC codes on another axis, and the agent codes on the final  axis. The elements of  the array are the exposure probabilities, however, only exposure probabilities greater than zero are recorded. Therefore, when a job-agent exposure estimate is missing from  the NIOSH JEM, it is difficult  to distinguish between the situations: the job-agent was surveyed with no exposure, there were no employees in that industry and occupation in the US, or the industry was not studied by NIOSH. 4.4 BCCA Canadian to US Job Translations To translate the Canadian job codes to US equivalents, the BCCA translations were used (Svirchev, 1993). Experts in occupational coding designed a system to translate the 853 CSIC codes to 231 USCENIND equivalents, and to translate the 503 CSOC codes to 499 USCENOCC equivalents. The occupational categories are generally similar for  the CSOC and USCENOCC, apart from  for fabrication,  processing, assembly, and machine operating occupations. Here, the CSOC classifies  the occupa- tions according to the product produced, whereas the USCENOCC classifies  according to the equipment used (Svirchev, 1993). The CSIC also has more specific  categories than the broader USCENIND categories. For example, the USCENIND defines  all hospitals in one category: 831 Hospitals. Whereas the CSIC includes eight categories distinguishing between the type of  hospital: 8619 Other Specialty Hospitals, 8617 Children's (Paediatric) Hospitals, 8616 Nursing Stations and Outpost Hospitals, 8615 Addiction Hospitals, 8614 Men- tal (Psychiatric) Hospitals, 8613 Extended Care Hospitals, 8611 General Hospitals, and 8612 Rehabilitation Hospitals. Industry titles corresponding to each CSIC code were matched to US industry title equivalents. The matching industry titles were verified  and those equivalents that did not correspond well or appeared infrequently  were excluded. This resulted in, a group of  USCENIND codes relating to each CSIC code, and similarly a group of  USCENOCC codes relating to each CSOC code. Many translations are not one-to-one relationships. The relationship between the Canadian and US codes is often  many-to-many. For example CSOC 2181 Mathematicians, Statisticians, Actuaries, translates to three USCENOCC codes: 066 Actuaries, 067 Statisticians, and 068 Mathematical Scientists n.e.c.. However, USCENOCC 068 Mathematical Scientists n.e.c. translates back to two CSOC codes: 2181 Mathematicians, Statisticians, Actuaries, and 2189 Occupations in Mathematics, Statistics, Systems Analysis, and related fields  n.e.c.. Here, n.e.c. denotes Not Elsewhere Classified.  All translations for  a particular job were considered equal and no indication was given of  which translation was more likely or closer to the 'truth'. Translations of  the patients' jobs are described further  in section 5.2.4. Chapter 5 Exposure Assessment In order to analyse the agents for  associations with bladder cancer incidence, a measure of  exposure to many separate agents for  each patient is required. For each US job, the NIOSH JEM gives the probability of  a person employed in that job being considerably exposed to many agents. Considerable exposure is exposure that occurs for  at least 30 minutes per working week or at least once per week for  90% of  the weeks of  the working year (see section 4.3.1). The BCCA questionnaire data includes the duration and type of  jobs held by each patient in the study. As all jobs held by the study subjects are coded according to the Canadian job codes, and the JEM is coded by US job codes, each Canadian job needs translating into US equivalents first. The JEM probabilities in conjunction with the duration of  each job provide a measure of  cumulative exposure. The cumulative exposure to a given agent for  a patient is estimated as the aggregation across all jobs of  the product of  that job's exposure probability estimate and the duration of  that job. This gives an expected number of  work-years with considerable exposure to each agent. The cumulative exposure estimate will give a higher weighting to the agent exposure probabilities in a patient's main job. Firstly the cumulative exposure definition  is explained in more detail in section 5.1. The process of actually calculating the cumulative exposure index is then described in section 5.2. 5.1 Cumulative Exposure For each agent, the cumulative exposure for  a patient is defined  as the aggregation over the patient's jobs of the product of  the probability of  considerable exposure and the job duration. Let i denote the ith patient (i  — 1 , . . . , 15463), j denote the j th job (j  = 1 , . . . , 10) and k  denote the fcth  agent (k  = 1 , . . . , 12945). So, the cumulative exposure, Eik, to agent k  for  patient i, is estimated as: 10 Eik ^ ^ ^ijk^ij 3= 1 where tij is the duration (in job-years) of  job j, and e ^ is the exposure probability estimate for  job j. A job-year is defined  as one year in a full-time  job. If  a job is part-time, it is considered half  as much work time as a full-time  job and thus the duration in years is divided by two. So t^ is calculated as: dij if  PT=0 tij  — < dij/2 if  PT=1 where dij is the duration of  job j , which is divided by two if  the job was indicated to be seasonal or part-time (PT  = 1 ) . If  the duration of  the job is not given, then the job duration is approximated by the end year minus the start year. The exposure probability estimates, e f̂c,  now need calculating. Before  using the JEM to look up the exposure probabilities, each Canadian job needs to be translated to US equivalents. Each Canadian job consists of  a CSIC industry (Xij) and CSOC occupation (yij) code. As discussed in section 4.4, the Canadian and US job codes do not have a one-to-one relationship. The relationship is often  many-to-many: for  each CSfC,  there can be many USCENIND equivalents and for  each USCENOCC, there may be many CSOC equivalents. Using the BCCA translation rules, let giND denote the function  that translates Xij into Sij different US industry codes, and gocc denote the function  that translates yij into Ty US occupation codes. Every possible permutation of  the translated industry and occupation codes is considered equal for  each Canadian job. For example, if  one Canadian job translates to 2 US industry codes and 3 US occupation codes, then there are 2 x 3 = 6 possible permutations of  US industry-occupation combinations. The set of  all possible combinations of  g/ND(%ij),  gocciVij)  is then the translation of  Canadian job j for  person i to US equivalents. The term 'job-translation' will be used to refer  to each of  these possible US industry-occupation combinations in this thesis. Each of  the possible job-translations will not always be of  equal value. Some of  the US job-translations will be closer to the true meaning of  the Canadian job than others. Some of  these job-translations may not even exist in practice in the US. Estimating differing  weights for  each job-translation possibility, however, is very difficult.  The amount employed in each industry and occupation combination in the US could be estimated. Given the amount employed in a US job, however, the proportion that corresponds with each of  many Canadian job equivalents could not be estimated, as the relationship between US and Canadian job codes is often  many-to-many. Also the proportions employed in each group would change over time. Therefore,  each job-translation is weighted equally, so that the JEM probabilities are averaged over all translations. So, ê fc  is calculated as: _ lCf=l  Ylt=1  fjEM,k{giND,s(Xjj)goCC,t(yij)} eijk  — C T J-  ij where x^ and y^  are the CSIC and CSOC codes respectively, for  patient i's  j th job. The function  that gives the sth translation of  CSIC code Xij is denoted by grND,s(xij), and the function  that gives the tth translation of  CSOC code y^ is given by gocc,t{Vij)-  The NIOSH JEM matrix function  that gives the exposure probability estimate to agent k  for  each US industry and occupation code combination given, is denoted by f J EM,k- 5.2 Calculating Cumulative Exposure Figure 5.1 depicts the flowchart  displaying the approach used to calculate the cumulative exposure estimates for  each eligible patient. Initially 14,754 patients met the inclusion criteria with completed questionnaire data and a known cancer diagnosis. Some further  patients were excluded from  the study as described below. The analysis was designed so any major cancer site could be chosen as the basis for  the case series and potential occupational carcinogens could be analysed for  that site. To enable analysis of  possible carcinogens for  any cancer site, exposure estimates were calculated for  all eligible cancer cases and controls. 5.2.1 Providing Adequate Job Information To calculate the cumulative exposure estimates for  each patient, much information  was required from  the questionnaire. Each patient needed to adequately describe each occupation they had and report the industry, so it could be coded into a CSIC and CSOC code. Also, for  each job they needed to provide the start year and end year, and indicate whether the job was part-time or seasonal. The start and end year was required to see if  the job occurred within 5 years of  diagnosis. If  any of  this information  was missing or unclear then the exposure estimate could not be calculated. Only 705 (4.8%) patients did not complete all the necessary data, so these patients were excluded from  the remainder of  the study. Patients were excluded if  any of  the job information  (industry, occupation, duration or start and end year) was missing or unclear. The job end date, start date and codes are the most important pieces of  information  for  calculating the cumulative exposure. Table 5.1 shows the distribution of  patients and the extent to which they completed the job codes and job end and start dates. The exclusions only form  a small proportion (4.8%) of  the patients. However, they should form  a random sample from  the patients so they do not affect  the later case-control analysis. The consequences of excluding these 705 patients from  the study did not make a considerable difference  to the types of  matched cases and controls in the later bladder cancer analysis as described in section 6.3.2. 5.2.2 Canadian Jobs and Latency The job history data was adjusted to allow for  a 5-year latency period. All jobs starting less than 5 years before  the patient's year of  diagnosis were deleted. All jobs with end dates less than 5 years before  diagnosis were reduced so they ended 5 years prior to diagnosis. The durations of  the jobs were then adjusted accordingly, that is, exposure in the 5 year period preceding diagnosis was not considered. Originally 26 patients had no jobs recorded. After  the deletion of  jobs within 5 years of  diagnosis, 100 patients had no jobs. There are now 14,049 patients eligible for  analysis, for  which exposure estimates needed calculating. The 100 patients that reported no jobs are estimated to be unexposed to all agents. The remaining patients reported 63,638 jobs that started more than 5 years before  their diagnosis; this is an average and standard deviation of  4.6 and 2.5 jobs per patient respectively. These patients contributed 483,138 work-years or an average and standard deviation of  34.6 and 12.2 work-years per patient reporting jobs. 5.2.3 Coding the Canadian Jobs According to the CSOC/CSIC Some Canadian occupations were not included in the CSOC classifications,  and thus could not be translated to US equivalents. The two occupations excluded from  the CSOC coding were occupations in the armed forces  and students. There were 2,927 jobs (4.6%) reported in the armed forces  (commissioned officers  and other ranks). From the dates given, many of  these jobs were during World War, II, and some were during World War L Some patients were also employed in the armed forces  for  their entire working life,  as 31 patients reported no other jobs than those in the armed forces.  These exposures occurred in different  countries, in different wars and at different  times and are thus very difficult  to estimate, and therefore  were assumed to be zero in this study. Also, some men may not have considered this work a job and may not have reported it. The 2,927 jobs in the armed forces  belong to 2,667 patients. Employment in the armed forces contributes 16,685.5 work-years (3.5% of  all work-years reported). This is an average and standard deviation of  6.3 and 6.1 work-years per patient respectively. The distribution of  work-years in the armed forces  per patient is right-skewed with a median of  5 work-years. Assuming no exposure for  armed forces  occupations means the cumulative exposures for  these 2,667 patients may be slightly underestimated. Table 5.2 shows the distribution of  cancer sites for  the work-years in occupations in the armed forces  compared to the work- years in other occupations. The number of  patients reporting any employment in the armed forces,  and the work-years employed in the armed forces  seem to form  the same distribution across cancer sites as the other occupations. Additionally, 2 patients were coded as full-time  students for  3 and 4 years and are assumed to have no exposure as teachers were by NIOSH. Therefore,  of  the 14,049 patients included for  analysis, 100 reported no jobs before  5 years prior to diagnosis, and 31 reported only armed forces  jobs. So, 131 are estimated as not exposed to all agents, and exposures need calculating for  the remaining 13,918 patients. 5.2.4 Translating Canadian Jobs to US Equivalents Using the BCCA translations described in section 4.4, each Canadian industry given was translated into an average of  1.2 US industry translations and each occupation was translated into an average of  2.7 US occupation translations. This resulted in 214,189 possible US job-translations for  the 60,709 Canadian jobs (an average and standard deviation of  3.5 and 4.8 US job-translations per Canadian job respectively). This large number of  job-translations per Canadian job was partly due to considering each permutation of translated industry and occupation code. The number of  job-translations may have been reduced if  the BCCA translations were performed  on the Canadian job industry and occupation pairs rather than the two independently. Table 5.3 shows a summary of  the proportion of  jobs, work-years, and patients employed in each CSIC major group (2-digit code). The construction industry employed the largest proportion, 10.9% of  the 60,709 Canadian jobs, had the largest proportion, 29.3%, of  patients ever employed within it, and contributed the largest proportion, 10%, of  work-years. Agriculture and manufacturing  industries were also large employers of  the 13,918 patients. To assess the translation accuracy, table 5.4 shows the number of  job-translations and work-years contributed by each USCENStC translation grouping. The US industry equivalent for  Canadian job CSIC code Xij is considered to be the average of  the Sij USCENSIC equivalent translations. The work-years contributed by CStC code Xij, is the job's work-years, Uj. Therefore,  each USCENSIC translation equiva- lently contributes %/Si j of  work-years. The US industry translations seem adequate, as the proportion of work-years in each major industrial grouping remains approximately equal before  and after  translation. Table 5.5 shows the proportions of  jobs and work-years in each major CSOC group (2-digit code) and table 5.6 shows the proportion of  translations in approximately equivalent USCENSOC groupings. Most patients were employed in occupational groups managerial and administrative, sales, farming,  product fabricating,  assembling, repairing and construction. Of  the 60,709 jobs, 39% were in these occupations, and 41% of  the work-years were in these occupations. Again, the US translations look adequate, as the proportion of  work-years in each major occupational grouping is approximately equal before  and after  translation, although the US major occupational group definitions  are not as consistent with the Canadian ones as the major industrial groupings were. 5.2.5 US Industries Studied by NIOSH Many US industries were not included in the JEM, as discussed in section 4.3.2. This resulted in 100,444 (47%) of  the job-translations not on the JEM because the US industry was not studied by NIOSH. Agent exposures will be underestimated in some jobs, but many of  these jobs should be truly non-exposed. 5.2.6 US Job-Translations on the JEM For 25% of  US job-translations the industry was studied by NIOSH, but the industry and occupation com- bination was not found  on the JEM for  any agents. It is difficult  to distinguish between the possibilities that 1) NIOSH studied the job and found  no exposure to all agents, 2) there were no employees in that industry and occupation in the US, and 3) NIOSH did not study the job, as they believed it would not have considerable exposure to any agents. Zero exposure could be assumed for  cases 1 and 3. However, for case 2 the job-translation is not valid, so it should be excluded from  the analysis and thus the average. As these different  possibilities were not detectable, and cases 1 and 3 are more likely, it was assumed that the exposure was zero for  all agents. Again exposure may be underestimated for  some jobs. For a Canadian job, with any US translations of  case 3, which are not valid, then the exposure estimate will be underestimated by averaging over too many translations. Of  the 214,189 job-translations, 60,356 (28%) could be found  on the NIOSH JEM. These 29,306 Canadian jobs belong to 10,420 patients, leaving 3,629 (26%) patients estimated as unexposed to all agents. Table 5.7 shows the distribution of  jobs and work-years in each major CSIC grouping for  all Canadian jobs and those with any exposure estimated by the JEM. The construction industry had the greatest proportion, 96.4% of  jobs with exposures on the JEM. The industries of  fishing  and trapping and finance,  insurance and real estate had no jobs on the JEM, as they were not studied by NIOSH. The agriculture, mining and government services industries also had a very small proportion of  jobs exposed. Table 5.7 also shows the work-years contributed in each industry and the equivalent amount of  work-years accounted for  on the JEM. A lesser proportion of  work-years were accounted for  on the JEM, due to many translations having zero exposure. Overall, 32.9% of  work-years were accounted for  on the JEM, or equivalently, were considered exposed on the JEM, whilst 48.3% of  jobs were exposed. Table 5.8 shows the same variables for the Canadian major occupational groupings. No jobs in teaching, religion or fishing  and trapping occupations were considered exposed. A small proportion of  those jobs in social science, farming,  sales, services, and mining occupations were considered exposed. A large proportion of  those jobs in materials processing, machining, and construction were considered exposed. As the JEM is assumed representative of  jobs located in North America, the locations of  the Canadian jobs with exposures estimated from  the JEM should be examined. Table 5.9 shows the locations of  the 29,306 Canadian jobs with any JEM exposures estimates, and the 153,618.6 work-years accounted for  on the JEM. Of  the 29,306 jobs at least 66.8% were in BC, and at least 71.4% of  the work-years accounted for  were in BC. At least 87.4% of  the jobs and 88.2% of  the work-years were in Canada. A maximum of  8.4% of  the jobs and 7.9% of  the work-years were outside Canada, and 4.2% of  the jobs and 3.9% of  the work-years had unknown locations. Therefore,  a very small proportion of  the exposures were estimated for  jobs outside Canada, and many may still have been located in North America. The few  work-years employed outside North America may have different  levels of  exposure, but it should make little difference  to the results. 5.2.7 Applying the JEM Firstly a subset of  the NIOSH JEM was created to enable easier electronic data handling. The JEM subset was restricted to the probabilities of  exposure in males only. The exposure probabilities were calculated as the ratio of  the NIOSH estimate of  those exposed in that job nationally, to the amount employed in that job nationally according to Dun's Marketing Index (Dun and Bradstreet, 1980). Sometimes this was slightly larger than 1 when NIOSH observed more employees than actually recorded as employed in Dun's Marketing Index. A maximum of  1 was set for  these proportions so they represented true probabilities. In addition, the JEM was restricted to only those US job-translations found  in the study. Therefore,  the NIOSH male JEM subset consisted of  405,183 industry-occupation-agent combinations, with 12,688 agents. The JEM subset was then applied to the 60,356 US job-translations. This resulted in over 9 million person-job-translation-agents. 5.2.8 Calculating Cumulative Exposure The person-job-translation-agents data were then compiled into the required cumulative agent estimates. Firstly, for  each person-job-agent, the probability estimates for  each US job translation were averaged. This resulted in a probability of  considerable exposure to each agent for  each Canadian job. Next, each probability estimate associated with each job was multiplied by its duration and divided by 2 if  the job was part-time. These estimates were then aggregated across jobs for  each patient and agent to give the final  cumulative exposure estimates to each agent. 5.2.9 Final Cumulative Exposure Estimates The process resulted in over 4 million person-agent exposure estimates, Eik, greater than zero. All remaining person-agent exposures were estimated to be zero. 10,420 patients (74%) had cumulative exposures greater than zero for  some agents. The 4 million person-agent estimates consisted of  11,882 different  agents. There- fore  none of  the patients in the study were estimated as being considerably exposed to 1,091 of  the NfOSH agents. The exposure estimates to each agent for  each patient tended to be quite small. The average cu- mulative exposure for  an exposed patient to each of  the 11,882 agents was calculated. The average and standard deviation of  these average cumulative exposures were 0.36 and 0.61 respectively. Similarly, the average and standard deviation of  the maximum cumulative exposure for  an exposed patient of  all 11,882 agents was 5.13 and 8.58 respectively, with the overall maximum exposure being 67.60 for  a patient exposed to continuous noise. The histograms of  cumulative exposure across all patients for  each agent are generally very right-skewed with the majority of  patients having a cumulative exposure of  zero or near zero. Taking the natural logarithm of  the positive cumulative exposures tends to make them normally distributed. Chapter 6 Statistical Approach Chapter 5 described how the cumulative exposure to each agent for  each patient was estimated. Before testing for  associations between cumulative exposure and bladder cancer incidence, possible confounders must be considered. The most important confounders  of  age at diagnosis and year of  diagnosis are first used as matching variables. They and additional factors  are taken into account in the conditional logistic regression analysis. The matching as discussed in section 6.1 identifies  the case and control patient groups. Possible confounders  considered are discussed in section 6.2 and the characteristics of  the cases and controls are outlined in section 6.2.1. A conditional logistic regression base model is developed to account for  the most important confounders  and is described in section 6.3. The consequences of  the subjects excluded due to missing occupational data previously in section 5.2.1 on this base model are investigated in section 6.3.2. Finally each agent is tested independently whilst simultaneously adjusting for  the important con- founders  as described in section 6.4. In addition, the agents are grouped into components that may act synergistically on bladder cancer development via principal component analysis as outlined in section 6.5. 6.1 Matching The analysis of  potential occupational carcinogens involves matching bladder cases to cancer controls on exact age at diagnosis and year of  diagnosis. Age is a well-known important risk factor  for  all cancers; bladder cancer risk increases with age. Age is also associated with occupational exposures; on average older patients will have had greater exposure to carcinogens than younger patients and they may have also been out of  the workforce  for  longer. In addition age is associated with most risk factors,  such as smoking habits, alcohol drinking habits, level of  education, etc. Matching on age to make the control comparison group more similar to the case group in age distribution was considered the best method to deal with these problems. There are also differences  between patients diagnosed in different  years due to questionnaire collection ceasing in later years for  some cancer sites. The questionnaire questions also differed  in early years. Again, matching on year of  diagnosis was used to allow for  these. Frequency matching was used to maximise the number of cases and controls used in the analysis. Patients with lung cancer were excluded from  the control series for  the bladder cancer analysis, as lung cancer is too strongly associated with cigarette smoking. After  removing the 2,808 lung cancers and those 705 patients with missing occupational data (see section 5.2.1) there are 1,066 possible bladder cases and 10,175 eligible controls. Matching on age and year of  diagnosis leads to 1,062 bladder cancers and 8,057 controls. These matched subjects were all diagnosed between 1983 and 1987, as all bladder cancers were diagnosed during this period. Also matching restricted the age range to 21 to 95. Further characteristics of the cases and controls are discussed in section 6.2.1. 6.2 Possible Confounder  Variables When analysing each occupational exposure for  its association with bladder cancer, there are many other factors  that could potentially act as confounding  variables and thus need controlling for  in the analysis in addition to those already matched upon. Information  was sought on potential confounders  in the BCCA questionnaire. This included information  on who completed the questionnaire, ethnic origin, marital status, years of  formal  education, smoking habits and alcohol consumption habits of  the patient. Caucasian men are known to be at greater risk of  developing bladder cancer than men of  other ethnicity (see section 3.1). Therefore  a simple ethnicity variable (Caucasian/Non-Caucasian) was considered in the analysis. Smoking is also a known risk factor  for  bladder cancer. Therefore,  many variables attempting to model lifetime  exposure to smoking, in particular cigarette smoking, were considered in the analysis. The continuous variables were categorised, as there may not be a linear relationship between the variable and bladder cancer, with sensible cut-offs  chosen. Most variables also have an unknown category for  when the patient did not provide an answer. Variables considered were ever smoker versus never smoker (cigarettes, pipes, or cigars), cigarettes smoked per day (0, 1-19, 20-29, 30+), years smoked cigarettes (0, 1-29, 30-44, 45+), cigarette pack-years defined  as the packs smoked per day multiplied by the years smoked, where 1 pack contains 20 cigarettes (0, 1-24, 25-49, 50+), and whether the patient quit smoking before  diagnosis and if  so, the number of  years they have quit for  (non-smoker, current smoker, 1-4, 5-9, 10+). The responses to some of  these smoking variables were quite varied, particularly pack-years. These variables are not as reliable or valid as possible as people may be unwilling or unable to provide true answers. Also, smoking habits differ  across a patient's lifetime  and this was not reflected  in the questionnaire. Although alcohol consumption is not thought associated with bladder cancer, it is associated with many other cancer sites. The occurrence of  malignant tumours of  the oral cavity, pharynx, larynx, oesophagus and liver is causally related to the consumption of  alcoholic beverages (IARC, 1987). Knowledge of  whether the patient drinks alcohol or not, was the only alcohol variable considered. An alcohol score variable was recorded as the aggregation across beer, wine and spirits of  the units drank per week multiplied by the years drank. Much of  this information  was unknown, especially for  the controls, leading to much rnisclassification and some differential  rnisclassification.  The reported units per week are also liable to much error and fluctuated  greatly across a patient's lifetime. An important variable in the analysis of  questionnaire data is who completed the questionnaire. Proxy responders are known to not complete the questionnaire as well as the patient themselves and also tend to leave more questions unanswered. Therefore  the simple variable, person completing the questionnaire (patient, proxy) was considered in the analysis. Educational level is related to occupation, age and income and associated with life-style  confounders like smoking, diet, and access to healthcare. The education variable (< 8 years, 8-11 years, high school graduate, post secondary education) was considered in the analysis. Marital status is also related to age, occupation and life-style  confounders  and so the marital status variable (single, married or common-law, widowed, separated or divorced) was also considered in the analysis. Naturally age and year of  diagnosis are adjusted for  in the analysis by performing  conditional logistic regression conditional upon the matched distributions of  these variables. 6.2.1 Characteristics of  Cases and Controls Table 6.1 shows the characteristics of  the cases and controls on the main possible confounding  variables. The bladder cases are still slightly older than the controls after  matching. Whereas the cases are almost uniformly  distributed across year of  diagnosis, the controls are more concentrated across earlier years, which is to be expected due to many cancer sites ceasing questionnaire collection before  the end of  the study. The average amount of  work-years contributed by each case and control is relatively similar. Also, 0.3% of  both cases and controls had no jobs. The bladder patients are more likely to be of  Caucasian ethnicity and to have smoked. If  they smoked cigarettes, bladder cancer patients were more likely to still be smokers at diagnosis. They also smoked slightly more cigarettes per day and for  more years than the controls, tf  the cases had quit by diagnosis, they had quit for  a shorter period of  time than the controls. The patient himself  was more likely to have completed the questionnaire if  he was a bladder case than if  he was a control. Cases and controls are relatively similar on marital status and education and whether they drink alcohol. 6.3 Developing the Base Model A base model is required to model the probability of  a subject being a case whilst taking into account the most important confounders  and the matched variables. As described in section 2.2.5, conditional logistic regression is the most appropriate model for  this situation. A parsimonious model that still explains the data is necessary so the model is more stable and more easily generalised. Backwards stepwise selection was the regression modelling technique implemented here. Maximum likelihood estimation was used to estimate the model parameters using the SAS procedure PHREG. The Wald statistic was used to test the hypothesis that a model parameter was zero and this was rejected when the resulting p-value was less than 0.20. The Wald statistic for  a parameter is the square of  the parameter estimate divided by its standard error. This is asymptotically distributed as a chi-square distribution (Hosmer and Lemeshow, 2000). However, the disadvantage of  the Wald statistic is that for  large parameter estimates, the estimated standard error is inflated,  resulting in failure  to reject the null hypothesis when the null hypothesis is false  (Menard, 2002). For categorical variables, the p-value testing the global null hypothesis that the coefficient  for  each dummy variable was zero was considered. Stepwise selection strategies are common in regression modelling. It is possible for  forward  selection methods to exclude variables such as those involved in suppressor effects.  A suppressor effect  is when one variable may appear to have.a statistically significant  effect  only when another variable is controlled. Backwards selection methods may not miss these variables, as they are all included in the initial model. However, the method is sensitive to the choice of  initial model. Research has shown that the choice of  alpha level of  0.05 is too stringent, often  excluding important variables from  the model. Choosing a value for  alpha in the range from  0.15 to 0.20 is highly recommended (Hosmer and Lemeshow, 2000). All the possible confounding  variables (ethnic origin, marital status, education, who completed the questionnaire, smoking status, cigarettes per day, years smoked cigarettes, cigarette pack-year, years quit smoking, and alcohol status) were entered into the conditional logistic regression model where year of  di- agnosis and age at diagnosis were the strata variables. The variables were deleted from  the model with the largest p-value in the following  order: education, marital status, cigarette pack-year, cigarettes per day, smoking status, then years quit smoking. 6.3.1 The Base Model The method described above resulted in a base model including the variables; who completed the question- naire, years of  smoking cigarettes, ethnic group and alcohol status. Table 6.2 shows the resulting odds-ratios for  these variables in the base model. A feature  of  the questionnaire responses was that questionnaires completed by proxies were signifi- cantly less likely to be bladder cases than controls after  taking into account age, year of  diagnosis, ethnicity, and smoking and drinking habits. This difference  is due to the prognosis for  bladder cancer being much better than for  other cancers in the control group. Thus bladder cancer patients are more willing or able to answer the questionnaire than other cancer patients. Patients reporting their ethnicity as non-Caucasian were less likely, although not significantly  at a 5% level, to develop bladder cancer than other control cancers when taking all other important confounders into account. This is as expected as Caucasians are at greater risk of  bladder cancer than other ethnic groups. Alcohol drinking was associated with a decreased risk of  bladder cancer here (but ever drinking alone is not significant,  p = 0.29), as alcohol drinking is a risk factor  for  other cancers serving as controls. There was a definite  dose-response relationship between smoking and risk of  bladder cancer. The risk was significantly  increased with each increased category of  cigarette smoking duration. Other models were tested and the previous base model was best in terms of  log likelihood. Table 6.3 shows the log likelihood for  the base model and some related models. The least significant  variable in the base model was ethnicity. Adding ethnicity to model 1 excluding it was significant  at the 20% level. Replacing the cigarette year variable with any other smoking variable did not improve the log likelihood. No remaining variables were significant  (all had p-values > 0.3) when added to base model. When the more complex alcohol score variable was added to the model, or replaced the simple alcohol status variable, it was not significant  at the 20% level. Many interaction terms were added to the base model such as cigarette years and alcohol status, and smoking status and alcohol status, but none were significant  at the 20% level. 6.3.2 Consequences of  the Missing Data Exclusions on the Base Model tn section 5.2.1 705 subjects were excluded from  analysis due to missing occupational information.  If  these had not been deleted, matching on age and year of  diagnosis would result in 1,125 cases and 8,492 controls. Their characteristics on the possible confounding  variables are compared with the cases and controls after exclusions in table 6.4. The distributions across all variables for  cases and controls matched from  all subjects and after  exclusions are very similar. The only differences  are that the cases and controls after  exclusions are slightly younger and a marginally greater proportion reported jobs (99.7%, rather than 99.3%). A decrease in the average age is expected, as the older patients may be more susceptible to recall difficulties.  The proportion of  questionnaires completed by the patient increases after  exclusions as much of  the missing occupational data was from  proxy questionnaires. The proportion of  unknown responses in the variables is also slightly decreased after  exclusions, as patients who did not complete their occupational histories often also did not complete the lifestyle  factor  questions. Table 6.5 shows the distribution of  cancer sites across the controls before  and after  exclusions. The distribution of  cancer sites comprising the control group remains very similar after  excluding subjects with missing occupational information. Using the same method as in the previous section, the cases and controls without exclusions resulted in the same variables in the base model. Table 6.6 shows base model comparison of  the significant  confounding variables. The odds-ratios differ  only slightly before  and after  exclusions. The risk for  non-Caucasian patients is just significantly  decreased with all subjects included, than with only those with complete occupational data. The alcohol score variable is actually significant  (p-value = 0.17)  when added to the base model before exclusions. However, the alcohol score is not a reliable variable as discussed in section 6.2. Also, the Pearson correlation between alcohol score and alcohol status is quite high at 0.63. Replacing the alcohol status variable by the alcohol score variable does not improve the log-likelihood. Hence, the base model resulting from  the cases and controls was considered approximately the same regardless of  missing occupational data exclusions. 6.4 Testing the Agents Individually Now the 9,119 cases and controls are identified,  the cumulative exposure estimates from  section 5 for  these patients can be used to analyse each agent's relation with bladder cancer risk whilst taking into account the important confounders  via the base model. Firstly, those agents the cases and controls are exposed to are summarised in section 6.4.1. Each agent is tested separately using conditional logistic regression and the base model, but there are many possible ways to analyse the cumulative exposure variable. First section 6.4.2 introduces an indicator ever/never exposure variable. As it is unlikely that the continuous cumulative exposure variable has a linear relationship with bladder cancer risk, the cumulative exposure is split into tertiles according to the exposed controls and a dose-response analysis is performed  as described in section 6.4.3. Another consideration is that when testing many agents, there are bound to be significant  results by chance alone. This can be taken into account with multiple testing techniques, where various methods are described in section 6.4.4. 6.4.1 Agents with Exposed Cases There needs to be sufficient  bladder cases exposed to an agent to enable analysis of  the exposure-disease relationship and to have confidence  in the results. Here exposed means that the cumulative exposure estimate is greater than zero. Of  all 11,882 agents that any patients included in the study were exposed to (see section 5.2.9), only 8,986 agents had at least one bladder cancer case exposed. Table 6.7 shows the distribution of bladder cases exposed to the 8,986 agents. On average 40 bladder cases were exposed to each agent with a standard deviation of  91 and a median of  5. Of  all 8,986 agents, each patient was exposed to an average and standard deviation number of  agents of  440 and 390 respectively. The median number of  agents exposed to was 340. The 5,699 agents with at least 3 bladder cases exposed are considered for  analysis. The 3,450 with at least 9 bladder cases exposed are considered for  the dose-response analysis. 6.4.2 Ever/never Any exposure versus no exposure is a simple indicator variable to analyse and interpret. However, in this study a cumulative exposure above zero does not mean the subject was ever exposed. All exposures in the NIOSH JEM were given a probability of  exposure, and the majority of  occupation-agent combinations had a low exposure probability estimate or were based on small numbers. The ever/never of  the cumulative exposure indicates whether the patient ever had a probability above zero of  being exposed (across all job translations) to the agent. Alternatively, the ever/never variable indicates whether NIOSH studied any of the patient's US job-translations and found  any employees exposed to that agent in their sample. The ever/never analysis is restricted to only those 5,699 chemicals with at least three bladder cancer cases ever exposed to that chemical to ensure sufficient  numbers for  analysis. 6.4.3 Dose-Response Each agent can be tested in the conditional logistic regression base model using the continuous cumulative exposure variable. However, this assumes the association between cumulative exposure and bladder cancer risk is linear. As this relationship is unlikely to be linear, the cumulative exposure is categorised instead. A dose-response relationship can be investigated by categorising the cumulative exposure according to biological risk levels, e.g. low, medium and high. However, there are no biological cut-off  values that will apply to all agents separately or as a group. Instead the groups are devised based on the cumulative exposure distributions. When the groups are divided according to the controls' cumulative exposures, the null hypothesis is that if  there were no association with bladder cancer risk then the cases should separate equally into the groups and no differences  between the cases and controls could be detected. So, the unexposed constitute one group and the exposed are divided into tertiles according to the cumulative exposures of  the controls. Thus four  groups are created: unexposed (reference  group), low exposure (lower 33% of  exposed controls), medium exposure (mid 33% of  exposed controls), and high exposure (top 33% of  exposed controls). If  there is a truly increasing dose-response relationship between the agent and bladder cancer then the low, medium and high exposure groups should have an odds-ratio significantly  greater than 1 and with the risk increasing across the groups. An agent is considered a carcinogen here if  the true dose-response relationship with cumulative exposure is increasing or has a threshold so the risk remains relatively flat  until the threshold where it increases significantly.  Additionally, an ordinal test was performed  to test for  a linearly increasing risk across the four  exposure groups. This involved assigning labels of  0, 1, 2 and 3 to the reference,  low, medium and high exposure groups respectively and tested the hypothesis that the slope amongst them in the conditional logistic regression model was zero. Assigning group medians as the ordinal score was considered, but as the scale of  cumulative exposure is not linear with risk then the simple 0, 1, 2, 3 scoring was preferred. The dose-response analysis is restricted to only those chemicals with at least nine bladder cancer cases ever being exposed to that chemical to ensure sufficient  numbers for  the low, medium and high exposure groups for  analysis. 6.4.4 Multiple Comparisons The p-value resulting from  testing the association of  one chemical exposure with the incidence of  bladder cancer, is the type I error, the probability of  a false  positive. However, when multiple chemicals are tested and multiple p-values result, many positive results are expected by chance alone. If  multiple chemicals are tested, but interest lies in looking at the results of  only one, then this is not a concern. However, if  a list of possible bladder cancer carcinogens is required, as is the case here, then the multiple testing must be taken into account. The most conservative way to allow for  multiple comparisons is to make a Bonferroni  style adjustment to control the Family-Wise Error Rate (FWER). The FWER is the probability of  at least one false  positive from  all chemicals tested. The Bonferroni  adjustment involves multiplying each p-value by the number of chemicals tested and comparing this to the desired FWER, usually 5%. Effectively  chemical exposures are declared significant  when their p-values are extremely small. This method lacks power and there will many true bladder carcinogens that do not get detected. Hochberg (1988) provides strong control (under all configurations  of  the true and false  hypotheses) of the FWER, but with greater power. If  m hypotheses H\, Hi,  • • •, Hm are tested with corresponding p-values Pi, P2, • • •, P m , then the p-values are ordered < P(2) < • • • < P(m) and H^  denotes the null hypothesis corresponding to P^y Hochberg's step-up procedure controls the FWER at a rate of  a as follows: let k  be the largest i for  which P ^ = m_ |f 1_ i; then reject all H^  i = 1,2,... ,k. Control of  the FWER is a conservative requirement that is often  not necessary. Alternatively, the False Discovery Rate (FDR) can be controlled allowing more power than the FWER controlling procedures. The FDR is the expected rate of  false  positives among the rejected hypotheses. Whereas the Hochberg procedure guarantees that the probability of  at least one false  positive is less than a, the Benjamini and Hochberg (1995) controls the rate of  false  discoveries at an expected value of  a%. Thus, in reality the true rate of  false  discoveries could be much larger (or smaller) than a. Although the Benjamini and Hochberg procedure has greater power than the Hochberg procedure, the overall type I error could be much higher than a. Note that when the FWER is controlled at rate a, the expected rate of  false  positives (FDR) is less than a%. Benjamini and Hochberg control the FDR at a rate of  a when the hypotheses are independent as follows: let k  be the largest i for  which = then reject all H^  i — 1, 2 , . . . , k. Controlling for  the FDR essentially declares the same or, more usually, a greater number of  hypoth- esises significant  than controlling the FWER. The FDR is intuitively appealing here as it looks at the error rate in the list of  chemicals selected, and has greater power than other procedures, so the Benjamini and Hochberg procedure is favoured  in this thesis. Although the hypotheses to be tested here are not all inde- pendent, the Hochberg and Benjamini and Hochberg procedures should provide a guideline to the control of the multiplicity problem. 6.5 Testing the Agents in Groups: Principal Components Analy- sis NIOSH provided exposure estimates for  many agents, many of  which are related to each other or have the same exposure probabilities across all jobs. Firstly, an attempt was made to construct natural groups from the agents. This organisation task has not been completed by NfOSH  for  their agents. The CAS (Chemical Abstracts Service) Registry is the largest substance identification  system in existence. The registry contains records for  more than 23 million organic and inorganic substances each assigned a unique CAS registry number, yet there is no defined  grouping structure to the CAS registry numbers. Grouping the NIOSH agents would be difficult  because agents with similar names could have different  compositions or functions, or ones with similar functions  or make-up could have different  names. To distinguish between all possible different  carcinogenic effects  of  chemicals, the groups would have to be quite small so that each agent is almost considered separately anyway. The process would require experts, be very complex, costly, and time consuming. Given that grouping based on the agent names would be complex; grouping based on the cumulative exposure distributions across subjects can be undertaken. One approach to group the agents in this way is via principal component analysis. The ideas of  principal component analysis are discussed in section 6.5.1. The components extracted can then be used in the conditional logistic regression model and the analysis approach taken is described in section 6.5.2. 6.5.1 Principal Component Analysis The aim of  principal component analysis (PCA) is to reduce the dimensionality of  a data set which consists of  a large number of  interrelated variables, while retaining as much as possible of  the variation present in the data set. This is achieved by transforming  to a new set of  variables, the principal components (PCs), which are uncorrelated, and which are ordered so that the first  few  retain most of  the variation present in all  the original variables (Jolliffe,  1986). The terms component and factor  are often  used interchangeably in PCA. The first  PC is found  by seeking a linear combination of  the original variables that extracts the maximum variation from  the data. This variation is removed and the second PC is sought explaining the maximum variation remaining in the data, and so on. Rotation methods serve to make the resulting PCs more interpretable. Varimax rotation is an orthogonal rotation method that minimises the number of variables that have high component loadings on each PC so that each PC has variables with either large or small loadings. The component loadings produced are the correlation coefficients  between the variables and PCs. Variables highly correlated with a PC are the defining  constituents of  that PC. A common rule is that a component loading is considered "weak" is less than 0.4 and "strong" if  greater than 0.6. The percent of  variation explained by a PC is the average of  the squared component loadings across all variables. The total variation explained by a PC is its eigenvalue. The important PCs to extract are those that combined account for  most of  the variation in the data. The Kaiser rule (Kaiser, 1960) is the most commonly used method to decide which PCs to extract. This criterion recommends extracting those PCs with eigenvalues of  at least one. The variable scores are standardised across subjects and combined according to the linear combi- nation of  variables described by the component to calculate the component score for  each subject. These components scores can then be used in place of  the original variable scores in other analysis such as logistic regression. The multicollinearity problems no longer exists for  multiple regression as the components are independent of  each other and the dimensionality of  the regression has been reduced considerably. It is important to include all variables relevant to uncovering the latent structure in PCA and exclude irrelevant variables. PCA does not require multivariate normality apart from  for  significance  testing. Including more variables into the PCA is not a good idea when there is a possibility of  suboptimal factor solutions ("bloated factors").  Too many similar variables will mask the true underlying factors.  To avoid suboptimization, PCA should start with a small set of  the most defensible  variables that represent the range of  each component. For algebraic reasons it is essential that there are more subjects than variables. There should be at least twice as many subjects as variables (Kline, 1994). 6.5.2 Grouping of  Agents and Analysis Approach There were many more possible agents (11,132) than subjects (9,119) to include in a PCA to uncover the latent structure of  the data and many of  the agents were not thought to be possible bladder carcinogens. The agents with the greatest potential of  being bladder carcinogens were instead identified  via the individual testing described in section 6.4 and principal components analysis performed  on the continuous cumulative exposures of  this subset. The cumulative exposure across these selected agents may be correlated as some patients were exposed to more than one agent at a time with some agents always occurring together in certain jobs. PCA with varimax rotation was performed.  Those PCs with eigenvalues greater than one were extracted. Agents were assigned to the component with which they had the greatest component loading and a component loading was considered "weak" if  less than 0.4 and "strong" if  greater than 0.6. Component scores were also calculated for  each patient, and these each have a zero mean and unit standard deviation across patients. Component cumulative exposures were created using a weighted average of  the cumulative exposures of  those agents assigned to each component using the component loading as the weighting. The dose-response and ordinal analysis could then be repeated comparing components rather than the agents individually. Also, an ever/never style analysis was performed  where a dichotomous variable for  each component indicated whether the patient was ever exposed to any of  the agents assigned to that component, versus the patient was never exposed to the assigned agents. Similarly, a dichotomous variable for  each component indicated whether the patient was ever exposed to all of  the agents assigned to that component, versus the patient was never exposed to at least one of  the assigned agents. The analyses of  ordinal dose-response, ever exposed to any, and ever exposed to all, were also each combined in a multivariate conditional logistic regression. The correlations amongst the newly created variables were checked for  no significant  multicollinearity. Chapter 7 Results The results are reported in the order described in chapter 6. Firstly section 7.1 reports individual agent results. The agents selected that exhibit a significant  association with bladder cancer risk are described in section 7.2. Section 7.3 then reports the results of  the principal component analysis on the selected agents. The selected agents and their properties are discussed further  in section 7.4. Finally, section 7.5 also provides a comparison of  the results from  this study for  those IARC classified  carcinogens with bladder cancer associations and those possible bladder carcinogens identified  by Siemiatycki (1991). 7.1 Individual Agent Results Section 7.1.1 reports the ever versus never exposed results for  the 5,699 agents with at least 3 bladder cases exposed. Section 7.1.2 reports the dose response results for  those 3,450 agents with at least 9 bladder cases exposed. Both the ever/never and dose-response results for  those 3,450 agents with at least 9 bladder cases exposed are listed in the appendix in table A.l. Note that to save space, only the NIOSH agent codes are given in the table, so the associated NIOSH agent names can be found  at the following  website: www.cdc.gov/noes/srch-noes.html. 7.1.1 Ever/never Table 7.1 summarises the results for  the 5,699 agents with at least 3 bladder cases exposed with the ever versus never exposure as the exposure variable tested. A significantly  (at the 5% level) increased odds-ratio was seen for  646 agents, of  which 163 have odds-ratios above 2. Table 7.2 lists the 7 agents that remain significant  at a 5% level after  adjusting for  multiplicity using the Benjamini and Hochberg procedure. The top 2 agents (2, 5- pyrrolidinedione, l-(2-((2-((2-((2-aminoethyl)amino)ethyl)amino)ethyl) amino)ethyl)-, monopolyisobutenyl derivs, and natural gas, liquified)  with the smallest p-values are also significant  at the 5% level for  the Hochberg multiple testing procedure. 7.1.2 Dose-Response The cumulative exposure estimates for  the 3,450 agents with at least 9 bladder cases exposed are divided into tertiles according to the controls. To give an idea of  the values of  these cut-off  values, the average and standard deviation of  the cumulative exposure estimate for  the 33rd percentile was 0.05 and 0.14 respectively, and the average and standard deviation for  the 67th percentile was 0.25 and 0.48 respectively. Table 7.3 shows the distribution of  p-values for  the dose-response variables. The p-values and odds- ratios associated with the ordinal test of  a linear dose-response trend (by assigning scores of  0, 1, 2 and 3 to the non-exposed, low, medium, and high cumulative exposures respectively) axe included in the table. Of the 3,450 agents, 350 had a significant  (5% level) linear increasing dose-response relationship; 22 and 2 of which were significant  at the 5% level after  adjusting for  multiplicity using the Hochberg and Benjamini and Hochberg multiple testing procedures respectively. The results for  the top 22 significant  agents are shown later in table 7.5. The top 2 significant  agents are 1, 2-ethanediamine, reaction products with chlorinated isobutylene homopolymer, and natural gas, liquified. Table 7.3 also shows the distribution of  p-values for  the odds-ratio for  the low, medium and high exposure groups. There were 377 agents with a significantly  (5% level) increased risk for  the low exposure group, 290 agents with a significantly  (5% level) increased risk for  the medium exposure group, and 215 agents with a significantly  (5% level) increased risk for  the high exposure group. None of  the p-values were significant  (5% level) after  adjusting for  multiple testing using the Hochberg or Hochberg and Benjamini procedures when looking at the results for  each exposure group separately. Table 7.4 compares the results from  the ever versus never exposure and dose-response analysis. Significantly  (5% level) increased odds-ratios for  both the ever versus never exposure variable and ordinal variable were seen for  307 agents. Also, 107 agents had significantly  increased odds-ratios for  both the ever/never and ordinal variable at a 1% level. 7.2 Selecting Significant  Associations It is useful  to select a small subset of  agents that exhibit sufficient  evidence indicating possible bladder carcinogenic properties that warrant further  research. Many of  the agents tested exhibited some positive relationship with bladder cancer risk. To provide a shorter selected list of  agents with less chance of  false positives, those agents with a significantly  increased ever exposure risk or a significantly  increasing linear dose- response relationship after  separately adjusting for  multiplicity via the Hochberg and Benjamini procedure were selected. In order to have adequate numbers for  dose-response analysis, the selection process was restricted to those 3,450 agents with at least 9 cases exposed. This enabled greater power in the ever/never analysis meaning that now 25 and 4 agents were significant  at the 5% level after  adjusting for  multiplicity using the Hochberg and Benjamini and Hochberg multiple testing procedures respectively. This selection procedure resulted in 30 selected agents, as listed in table 7.5. All of  the agents had an ever/never p-value less than 0.2% and ordinal p-value less than 0.8%. Of  the agents selected 20 had ever/never odds-ratios greater than 1.3. Only 9 agents did not have all three dose-response levels with odds- ratios significantly  greater than one at a 20% level, whereas 22 agents did not have all three dose-response levels with odds-ratios significantly  greater than one at a 5% level. 7.2.1 Linear Exposure As mentioned previously in section 5.2.9, the distribution of  the positive cumulative exposures for  most agents is highly right-skewed. Figure 7.1 shows the histograms of  the positive cumulative exposures for  each of  the 30 selected agents. However, a logarithm transformation  results in normally distributed positive cumu- lative exposures for  most agents. A linear conditional logistic regression fit  through the original cumulative exposures would be highly dependent on extreme observations. A linear regression through the transformed cumulative exposures would be a much more robust fit.  However, a logarithmic transformation  leaves the question of  what to do about the zeros, the non-exposed patients. The results of  a linear regression fit  would vary depending on the score assigned to the non-exposed patients. The Box-Cox transformation  (Box and Cox, 1964) is used instead as it tends to a logarithmic transformation  as A tends to zero. The box-cox transformation  is as follows: if  A / 0 x(X)  = < log(X)  if  A = 0 A value of  A of  1/100 or 0.01 was chosen as sufficiently  small to transform  the distributions of  the original positive cumulative exposures to be approximately normal. Figure 7.2 shows the histograms of the transformed  positive cumulative exposures for  the 30 selected agents. Table 7.6 shows the results of fitting  a straight line through these transformed  cumulative exposures for  the top 30 agents. As expected, all 30 agents have a very significant  increasing linear trend, all with a p-value less than 0.0014. If  all 3,450 agents were tested, then the top 4 agents (sulfonic  acids, petroleum, magnesium salts; natural gas liquefied; phosphorodithioic acid, mixed O, 0-bis(sec-Bu and 1,3-dimethylbutyl) esters, zinc salts; 1, 2-ethanediamine, reaction products with chlorinated isobutylene homopolymer) would remain significant  after  adjusting for multiplicity using the Hochberg procedure. 7.3 Grouped Agent Results: Principal Components Analysis The exposures to the chosen 30 significant  agents may not be independent. For example, some agents may always occur together, so if  a patient was exposed to one agent then they were also exposed to the partnering ones. It would then be difficult  to distinguish which agent is truly associated with the bladder cancer risk. Also, the analysis of  the chosen 30 agents has been only univariate thus far.  A multivariate analysis could be performed  to allow the effects  of  an agent's exposure to be jointly adjusted for  all the effects  of  the other agent exposures. The patients' cumulative exposures to many of  the selected agents are highly correlated with each other (12 of  the agents have a Pearson correlation coefficient  greater than 0.9 with at least one other agent), and hence multicollinearity is a potential problem. Therefore,  principal components analysis was performed  to examine the relationships between the agents. Performing  principal components analysis on the 30 agents resulted in 10 components with an eigen- value greater than one. Table 7.7 lists the component loadings for  each agent and identifies  those loadings greater than 0.4. The largest component accounts for  26.3% of  the total variance, and the first  3 components combined account for  more than 50% of  the total variance. The component scores were used to create a cumulative exposure variable for  each component by using the component scores of  those agents associated with a component as the weights and computing a weighted average of  the agents' cumulative exposure. The following  sections provide the results of  a dose-response analysis, an 'any exposure' analysis and an 'all exposure' analysis. 7.3.1 Component Groups - Dose-Response Table 7.8 shows the results of  the dose-response analysis on the component groups. As expected all compo- nents show a significantly  linearly increasing dose-response relationship. Table 7.9 shows that components 2, 4 and 9 remain significant  (5% level) after  backwards selection when all 10 component ordinal variables are entered into a multivariate conditional logistic regression model. 7.3.2 Component Groups - Any Exposure Table 7.10 shows the results of  the 'any exposure' analysis on the component groups where 'any exposure' is defined  as cumulative exposure greater than zero for  any members of  the component. As expected all com- ponents show a significantly  increased risk if  a patient was ever exposed to any members of  the component. Table 7.11 shows that components 3, 6 and 10 are significant  (5% level) from  backwards selection when all 10 component 'any exposure' variables are entered into a multivariate conditional logistic regression model. 7.3.3 Component Groups - All Exposure Table 7.12 shows the results of  the 'all exposure' analysis on the component groups where 'all exposure' is defined  as cumulative exposure greater than zero to all of  the members of  the component. As expected all. the components show a significantly  increasing risk when a patient is exposed to-all members of  the component. The odds-ratio for  the 25 patients ever exposed to all 5 agents in component 2 is large at 3.11. Table 7.13 shows that components 1 and 2 remain significant  (5% level) after  backwards selection when all 10 component ordinal variables are entered into a multivariate conditional logistic regression model. 7.4 Properties of  the Selected Agents The cumulative exposures of  the agents were derived from  a JEM that distinguished exposure probabilities according to job type. So it was suspected that the selected 30 agents would be grouped in some way according to jobs. If  workers in a job are exposed to a particular agent that is a bladder carcinogen, but they are always exposed to other agents alongside the carcinogen, then it would be difficult  to distinguish between the agents. This effect  can be seen to some extent in the selected agents. Figure 7.3 shows a breakdown of the total cumulative exposure to each agent contributed by all study patients according to US job (industry and occupation pair). As an example, 85% of  the total cumulative exposure to X2307 (alkenes, C15-18 alpha-, reaction products with sulfurized  dodecylphenol calcium salt, sulfurized)  experienced by the 9,119 patients was due to employment in a timber cutting or logging occupation in the logging industry. All agents comprising the first  principal component have a substantial proportion of  their cumulative exposure due to employment in this job. In fact,  timber cutting or logging occupations in the logging industry accounted for the largest proportion (32%) of  the total cumulative exposure to all 30 agents. Interestingly, all cumulative exposure to Y1006 (natural gas, liquified)  was due to employment in gasoline service station related occupations. Furthermore, a proportion of  the participants of  the NIOSH NOES study employed in this occupational group were exposed to 28 of  the agents selected, all but 73075 (SN, tin - MF unknown) and 90590 (clay, nec). It must be noted that these jobs are the US classifications, and there may be more than one possible Canadian job translation equivalent. However, the gasoline service station related occupations only translate to one Canadian equivalent; Gasoline Service Stations - Service Station Attendants. As expected, many of  the agents seem to form  principal component groupings according to the distribution of  the US job equivalents they occurred in. The cumulative exposures across all 9,119 pa- tients for  X2305 (2,5-pyrrolidinedione, l-(2-((2-((2-((2-aminoethyl)amino)ethyl)amino)ethyl)amino)ethyl)-, monopolyisobutenyl dervis., reaction pr) and X2308 (sulfonic  acids, petroleum, magnesium salts) comprising the fourth  principal component have a correlation of  0.999. Their similarity can also be seen across their job distributions. Table 7.14 lists the agents in order of  the principal components as per table 7.7. The US job that contributes most to the cumulative exposure of  that agent and what percentage it contributes (as seen in figure  7.3) is given. Additionally, it provides the JEM proportion of  employees exposed to that agent in the listed job. For example, just over 50% of  people in the NIOSH NOES study employed in timber cutting and logging occupations were exposed to each of  the agents comprising the first  principal component. However, a very small proportion (1-3%) of  people employed in gasoline service station related occupations were exposed to each of  the agents comprising the second principal component. Table 7.14 also lists the number of  bladder cancer cases exposed to each agent and the CAS (Chemical Abstracts Service) number for  each NIOSH agent if  applicable. The CAS number was used to identify  an IARC classification,  which is also provided if  available. An IARC classification  could not be found  for  most agents mainly due to the complexity of  the agents involved. There is very little information  available on many of  these complex chemicals. 7.4.1 Discussion of  the Selected Agents First  Principal  Component.  Most (52%) of  the cumulative exposure to agents in this component was due to employment in timber cutting and logging occupations. Just over half  of  the NIOSH NOES study participants employed in timber cutting and logging occupations were exposed to each of  these agents. X2298 (phenol, dodecyl-, sulfurized,  carbonates, calcium salts, overbased) has not been classified  by IARC. It is an ingredient (1-5%) in "Energol CLO 50M" (diesel engine oil). X2293 (sulfonic  acids, petroleum, calcium salts, overbased) is not classified  by IARC, but is 100% of  the ingredients of  "Syndustrial P Compressor Oil (All Grades)". X2295 (phosphorodithioic acid, O, 0-bis(2-ethylhexyl) ester, zinc salt) is not classified  by IARC, but comprises 0.5-1% of  "CHAMPION SUPER GRADE 5W20 (4229)" (petroleum based lubricating oil): X2689 (1, 2-ethanediamine, reaction products with chlorinated isobuylene homopolymer) is not classified by IARC, but comprises 40-50% of  "VANLUBE 869" (industrial antioxidant). Here 60713 (products of combustion - gasoline (leaded)) is considered the same IARC classification,  possibly carcinogenic, as gasoline engine exhaust (IARC monographs Vol.: 46 (1989) (p. 41)). X5263 (products of  combustion - jet fuel and gasoline, unleaded) has not been classified  by IARC. However, as gasoline engine exhaust is classified  as possibly carcinogenic and jet fuel  is classified  as not classifiable,  then X5263 may have carcinogenic properties. The agents comprising the first  principal component seem to be petroleum or mineral oil based, and occur frequently  in the logging industry. Second  Principal  Component.  Most (52%) of  the cumulative exposure to agents in this component was due to employment in gasoline service station related occupations. Very few  (less than 4%) of  the NIOSH NOES study participants employed in gasoline service station related occupations were exposed to these agents. Few bladder cancer cases were exposed to these agents, with the maximum being 35 cases exposed to X1401 (2-butenedioic acid (E)-, polymer with 1,3-butadiene and ethenylbenzene). This agent has not been classified  by IARC, but ethenylbenzene is classified  as possibly carcinogenic, and 1,3-butadiene is classified  as probably carcinogenic. Thus, agent X1401 may have possible carcinogenic properties. 83048 (nonylphenoxyethanol) is not classified  by TARC, however it is an alkylphenol ethoxylate. Alkylphenol ethoxylates are used in industrial detergents (such as those used for  wool washing and metal finishing),  in some industrial processes, and in some liquid clothes detergents (Warhurst, 1995). Alkylphenols are an environmental concern as they do not break down in the environment and accumulate in rivers, fish and birds, and they have oestrogenic properties. The effects  of  nonylphenol on cultured human breast cells (Soto et al., 1991) also led to health concerns in humans. Subsequently, many European countries have brought in controls on alkylphenols, and Switzerland has banned the use of  all alkylphenol ethoxylates. X4267 (ether, tert - butyl methyl) is classified  as not classifiable  by TARC. Methyl tert-butyl ether is a volatile synthetic chemical that has been used widely since the 1980s in proportions up to 15% as a component of  gasolines for  its octane-enhancing and air pollution-reducing properties. In service stations where fuels  containing > 10% methyl tert-butyl ether are delivered, the average concentration to which attendants are exposed is about 0.5 ppm (2 mg/m3) (IARC monographs Vol.: 73 (1999) (p. 339)). No epidemiological studies have directly addressed the relationship between methyl tert-butyl ether exposure and human cancer risk. However, inhalation of  methyl tert-butyl ether resulted in increased incidence of renal tubular tumours in male rats. Perhaps the relationship seen between the cumulative exposure to agents in the second component and bladder cancer risk in this study is due to the possible carcinogenic effects  of  X4267 (ether, tert - butyl methyl). Alternatively, some other chemical exposures involved in the gasoline service station attendant job could make these employees at greater risk of  bladder cancer. Third  Principal  Component.  Much (27%) of  the total cumulative exposure to this component comes from  employment as a plumber pipefitter  and streamfitter  apprentice in the construction industry. Agent 90320 (asphalt) is classified  by IARC as not classifiable  (IARC monographs Supplement 7 (1987) (p. 133)). There have been no epidemiological studies looking directly at the association with asphalt exposure and human cancer risk. However, a cohort study of  US roofers  indicated an increased risk for  cancer of  the lung and suggests increased risks for  bladder cancer. The asphalt group is on the priority list of  agents to consider in future  IARC monographs due to several ongoing epidemiological studies. For example, Randem, et al. (2003) recently found  increased lung cancer incidence rates in a cohort of  male Norwegian asphalt workers. Siemiatycki (1991) also found  a significantly  increased risk of  bladder cancer for  substantial exposure to asphalt. The relationship seen between the cumulative exposure to agents in the third component and bladder cancer risk in this study could be due to the possible carcinogenic effects  of  90320 (asphalt). Fourth  Principal  Component.  Much (25%) of  the total cumulative exposure to this component is due to employment as a carpenter in the ship and boat building industry. A further  23% of  the total cumulative exposure is due to employment as a miscellaneous electronic equipment repairer in the pulp paper and paperboard mill industry, and 18% of  the total cumulative exposure to this component is due to employment as a heavy truck driver in the trucking service industry. Fifth  Principal  Component.  X1075 (phosphorodithioic acid, 0-(2-ethylhexyl) O-isobutyl ester, zinc salt) is not classified  by IARC, but is <10% of  the ingredients of  "multi-purpose lubricant (dri-side)". Agent 36955 (hexane) is not classified  by IARC, but is chemical made from  crude oil and often  used to produce solvents. Sixth  Principal  Component.  M1150 (cyclohexylamine, n - ethyl -) is not classified  by IARC, although cyclohexylamine is not classifiable  (IARC monographs Supplement 7: (1987) (p. 178)). The classified  IARC group is for  cyclamates, which are artificial  sweeteners. The IARC states that the evidence that the risk of bladder cancer is increased among users of  artificial  sweeteners is inconsistent. Exposure to M1150 could have associations with increased risk of  bladder cancer. M0984 (ethanol, 2-(2-(2-butoxyethoxy) ethoxy)-) has not been classified  by IARC, but it is a triethylene glycol ether. Some monoethylene glycol ethers are nominated for  IARC review, so it is possible that M0984 could have some carcinogenic effects. Seventh  Principal  Component.  Agent 90590 (clay, nec) has not been classified  by IARC, however Siemiatycki (1991) found  a significantly  increased risk of  bladder cancer for  ever exposure to clay dust. Perhaps inhalation of  the dust of  agent 90590 has an effect  on bladder cancer development here. Agent T1475 (solvent refined  heavy paraffinic  distillate (petroleum)) has been classified  by IARC. It is either classified  as definitely  carcinogenic or not classifiable  depending upon whether it is untreated and mildly treated mineral oil or a highly-refined  mineral oil respectively (IARC monographs Vol.: 33 (1984) (p. 87)). Eighth  Principal  Component.  Cumulative exposure to T1909 (nonylphenol ethylene oxide adduct) was mostly due (64%) to employment as a lathe and turning machine set-up operator in the ship and boat building industry. It has not been classified  by IARC, but it is also an alkylphenol ethoxylate as agent 83048 is from  the second component. Ninth  Principal  Component.  Agent 92500 (oil, hydraulic) is a mineral oil and depending upon whether it is untreated and mildly treated or highly-refined,  then it is classified  as definitely  carcinogenic or not classifiable  respectively (IARC monographs Vol.: 33 (1984) (p. 87)). Agent P0620 (impact noise) is probably an example of  an agent that always occurs alongside the possible carcinogen (here, hydraulic oil). Tenth  Principal  Component.  Most (51%) of  the total cumulative exposure to X1894 (2-propenoic acid, 2-me-, C12 ester, poly w/ C16 2me2propenoate, iso-CIO 2me2propenoate, me 2me2propenoate, C18 2me2propenoate, C14 2me2propenoate) is due to employment as a power plant operator in a hospital. All of  the NfOSH  NOES subjects employed in this job were exposed to agent X1894. A further  23% of  the total cumulative exposure is due to employment as a knitting, looping, taping, and weaving machine operator in the apparel and accessories (except knit) industry. 7.5 IARC and Siemiatycki Results Comparison It would be useful  to see how the results compare for  those agents that are already considered bladder carcinogens. Table 7.15 shows table 3.2 from  section 3.3 updated with corresponding agent results where possible. The chemicals listed are those IARC has classified  that include bladder as one the cancer sites affected  by the chemical. Most chemicals were translated into a NIOSH equivalent via the CAS number. When there were multiple CAS numbers for  the chemical, or multiple NfOSH  equivalents, then all are provided. Often  there were no NIOSH equivalent chemicals. Often  if  there was a NIOSH equivalent chemical then it had very few  cases exposed possibly because the use of  the chemical had been restricted. The results for  the IARC classified  definitely  carcinogenic chemicals do not seem to support that classification,  although the odds-ratios are not significant  and the numbers are small. It could be the case that these chemicals are more carcinogenic for  other cancer sites so are not showing a result for  bladder. However, the results for  the fARC  classified  possibly carcinogenic chemicals are much more consistent. Lead is the only possibly carcinogenic classified  agent to show consistent results. Table 7.16 shows the Siemiatycki potential bladder carcinogens from  table 3.3 updated with corre- sponding agent results where possible. Finding equivalent NIOSH results was more difficult  here as often Siemiatycki would group chemicals in broad categories. There were then no such NfOSH  categories to com- pare with. The results in this study were consistent with Siemiatycki's for  titanium dioxide, engine emissions, diesel engine emissions, calcium carbonate, formaldehyde,  and asphalt. Other agents showed results similar to Siemiatycki's, but were not statistically significant,  often  due to small numbers. Chapter 8 Discussion Identifying  occupational carcinogens is a difficult  task. Cancer is a chronic disease, so a long time period may elapse between occupational carcinogen exposure and cancer symptoms. In order to observe enough study subjects that develop cancer, estimation of  the occupational exposures and other confounders  is often imprecise or may involve many assumptions. Despite some concerns over the validity of  epidemiological studies, it is clear that formal  attempts at examining the human experience are required when extrapolating the results of  animal experiments is difficult. Whilst many assumptions were made in this study, and the methodology and JEM applied not perfectly  precise, we have been as rigorous as possible to provide as valid results as we could given the difficult  problem and data available. In testing so many chemical exposures we were bound to find  many positive associations by chance alone, however allowance for  multiplicity was made in the calculations. Section 8.1 provides a summary of  the study's findings,  fn  section 8.2 issues of  bias and confounding  are discussed and the validity of  the study methodology is discussed in section 8.3. Further research directions are outlined in section 8.4. 8.1 Summary of  Findings Many of  the patients in our study were potentially exposed to many of  the NfOSH  agents investigated. There were 5,699 agents for  which at least 3 bladder cancer cases had potentially been exposed, of  which 3,450 had at least 9 bladder cancer cases exposed. Positive associations with bladder cancer were seen in many agents. A significantly  (5% level) increased odds-ratio was seen for  ever exposure to 646 agents. A significantly  (5% level) increasing linear dose-response was seen in 350 agents. A subset of  30 agents was selected as exhibiting sufficient  results to indicate possible bladder carcino- genie properties requiring further  research. The agent exposures were correlated, mainly amongst jobs, and 10 independent groups of  agents were identified.  Most of  the selected agents seemed to have some petroleum or mineral oil base. IARC has classified  occupational exposures in petroleum refining  (IARC monographs Vol. 45; 1989) and diesel engine exhaust (IARC monographs Vol. 46; 1989) as probably carcinogenic, marine diesel fuel  (IARC monographs Vol. 45; 1989), gasoline engine exhaust (IARC monographs Vol. 46; 1989), heavy residual fuel  oils (IARC monographs Vol. 45; 1989), and gasoline (IARC monographs Vol. 45; 1989) as possibly carcinogenic. Excess bladder cancer risk has also been observed frequently  among truck and motor vehicle drivers (Silverman, 1989). This cancer risk is thought to be partly due to exposure to the polycyclic aromatic hydrocarbons (PAHs) contained in exhaust emissions. A significant  (5% level) odds-ratio of  1.92 for  ever exposure to the PAH benz(a)anthracene was found  in this study. Only a small number of  bladder cancer cases were ever exposed to the other PAHs. Most of  the cumulative exposures to the 30 agents occurred due to employment in timber cutting and logging occupations, ship and boat building and construction industries, and occupations involving motor vehicles (gasoline service station attendant, mechanic, truck driver, motor vehicle production). Significantly increased bladder cancer risks were seen for  these occupations and industries when an analysis was performed on the same BCCA occupational data as in this study, but when looking at ever or usual employment in an occupation or industry (Band et al, 2004). We hope to have provided some further  insight into what particular chemicals within these occupations may contribute to the increased bladder cancer risk. 8.2 Bias and Confounding Issues of  bias and confounding  are discussed further  in the following  sections. 8.2.1 Comparability of  Source Populations for  Cases and Controls The bladder cases were obtained from  a well-defined  source population of  males resident in BC aged over 20 when diagnosed with cancer between 1983 and 1990 and ascertained by BCCA. BCCA receives information from  every newly diagnosed cancer case in the province. So the bladder cases are approximately an exhaustive group of  eligible subjects diagnosed with bladder cancer within the study period. The controls came from the same source population and are almost an exhaustive group of  those patients diagnosed with cancer within the study period, apart from  the primary unknown sites and the lung cancer patients are excluded from  analysis. Thus the controls represent those in the source population that would have been cases if  their "primary cancer were diagnosed as bladder cancer. 8.2.2 Selection Bias Non-response bias was minimal as the response rate was quite high (64.7% of  bladder cancers responded and 64.1% of  possible control cancer sites responded) and there were no major differences  between non- responders and responders apart from  responders were more likely to have managerial or administrative as their usual occupation. The cases and controls may not represent the source population fully  with respect to occupational exposure, as the source population may be slightly more exposed to potential occupational carcinogens than the cases and controls. This difference  should not be large and the results should still extend to the entire source population. Patients with missing occupational data were excluded from  analysis, but they were a small subgroup and did not differ  from  the remaining patients substantially. The source population contains all males aged over 20; so those that worked for  many years in few jobs, those that changed jobs frequently,  those with few  work-years, and those non-workers. Healthy worker bias is unlikely here as both the cases and controls had the same proportion (0.3%) of  non-workers, both had reported a similar amount of  jobs (the cases reported an average and standard deviation of  4.6 and 2.5 jobs and the controls reported an average and standard deviation of  4.9 and 2.6 jobs), and both had a similar duration of  work-years for  each patient. 8.2.3 Information  Bias Little recall bias was expected as all responders from  the source population used in the analysis responded to the questionnaire as if  they were a case. The information  gained on the cases and controls would have contained some error, but very little, if  any of  this would have been differential. Misclassification  of  case and control status is very unlikely as diagnoses were classified  using the ICD-9 codes and all were histologically confirmed.  The patients with a primary unknown cancer site were also removed from  the analysis to avoid this misclassification.  The questionnaire was considered valid and reliable. However, the questionnaire data will contain much misclassification.  Data is often  unknown for some patients, some patients could have accidentally answered incorrectly or have been unwilling to answer truthfully.  However the degree of  the misclassifications  should not differ  between cases and controls. A variable that may have suffered  from  differential  misclassification  was the alcohol score variable, so this variable was excluded from  the analysis. A greater proportion of  controls received the early questionnaire that omitted questions regarding alcohol drinking habits so the controls were likely to have greater misclassification on the alcohol score variable than the cases. Misclassification  of  occupational exposure is expected due to the difficulty  in approximating the true exposure. Again, this misclassification  should not be differential.  Patients with missing occupational data were excluded from  analysis, however the occupational data given by the remaining patients could be with error. The larger errors occur in approximating the occupational exposure from  the occupational histories. The JEM used to estimate exposures in different  US jobs has limitations, for  example, it was only applicable to one period in time, the jobs were located in the US, there was no allowance for  exposure variability within a job, and many industries were excluded from  the JEM. The calculation method will have incorporated error also; all job translations were considered equally adequate, the average was considered an appropriate method to combine exposures across job-translations, a part-time job was considered half  as any work-hours as a full-time  job, exposures missing from  the JEM were considered zero, and the exposures were considered to have equal weight across a persons lifetime  (e.g. here childhood and early exposures are as important as late exposures or exposures after  retirement, when it could be the case that early exposures are much more important than late exposures). 8.2.4 Confounding  by Non-Occupational Variables Non-occupational confounders  were taken into account in the matching and conditional logistic regression base model. The confounders  were consistent with current knowledge of  bladder cancer and the questionnaire design. Information  was not sought for  some other possible bladder cancer non-occupation confounders,  but they may not have had much effect  on the base model beyond the main confounders  identified. 8.2.5 Confounding  by Occupational Exposure Occupational exposure to known bladder cancer carcinogens may confound  the effect  of  other occupational exposures. Different  occupational exposures may cluster together within a job and it would be hard to distin- guish which agents actually contributed to the exposure-disease association. These problems are addressed in part by the principal component analysis, which considers groups of  agents that group together according to the cumulative exposure estimates. 8.3 Evaluation of  Methods The methodology used in this study had many limitations. The greatest possibly being the use and applica- bility of  the NIOSH JEM. As discussed in chapters 4 and 5, some important industries were excluded from the NIOSH NOES study. The NIOSH JEM only covered a short period in time, whilst we were estimating exposures over patients' lifetimes,  during which time many workplace conditions have changed. Assump- \ tions were made that the NIOSH JEM probabilities represented the lifetime  probabilities of  exposure for  the BCCA patients. Assumptions that the US jobs were comparable to Canadian jobs seem fair  and the great majority of  the BCCA patients' jobs occurred within Canada. The results of  this study indicate another limitation of  the JEM, in that it only differentiates  up to the level of  the job. No allowances are made for exposure variability or different  working habits within a job. Also, in this study we were unable to distinguish between different  concentrations and frequencies  of  exposure, as was the case in Siemiatycki (1991) study. However, using a JEM with probability assessments enabled us to put a much more precise estimate on the actual exposure of  a patient rather than that obtained through interviewing the patient. There were also many possible ways of  analysing the results. Two additional methods, which were not considered as relevant as the ever/never variable and dose-response analysis, are discussed in the next two sections. 8.3.1 Ever/never 0.5 As mentioned before  in section 6.4.2, the ever/never variable used in the analysis does not indicate definite exposure to an agent. In an attempt to find  a variable that more truly measures ever versus never exposure, a possible cut-off  value for  the cumulative exposure estimates was hypothesised. A cut-off  value of  0.5, so ever exposure represented those patients expected to be exposed for  at least half  a work-year (or a part-time work-year) versus those expected to be exposed for  less, was a natural cut-off  value. However, very few patients had exposures that high. Of  the 2,772,021 exposed bladder cases and controls and agent pairs, 622,560 (22%) had a cumulative exposure greater than 0.5. Of  all 8,986 agents, only 2,237 (25%) had at least three bladder cases with cumulative exposure above 0.5. The value of  0.5 was chosen as an attempt to capture those patients with likely exposure in a job (i.e. probability of  exposure greater than 0.5) and those with a large amount of  expected exposed years. However, if  a patient's true chance of  exposure in their Canadian job was at least 50%, they did not necessarily have a cumulative exposure greater than 0.5. This is because the probability of  exposure calculated for  the Canadian job was the average over all those for  the US translation combinations and some of  these may have had zero exposures. Therefore  the results for  the analysis on cumulative exposures above 0.5 versus those below 0.5 were not provided in this thesis, and instead the results from  the simple ever/never variable were presented. 8.3.2 Siemiatycki Comparison The study undertaken by Siemiatycki (1991) was described in section 3.3. Exposure variables comparable to his 'any exposure' and 'substantial exposure' were considered. His definition  of  'any exposure' was exposure that was at least probable, at least at a background level and occurred at least 1% of  the time. This was not easily comparable to the calculated cumulative exposure as NIOSH measured exposures occurring at least 1.4% of  the time at any concentration and the confidence  was measured by the JEM probability. Siemiatycki's definition  of  'substantial exposure' was exposure that was at least probable, above a background level and occurred at least 5% of  the time for  at least 5 years. The 'any exposure' and 'substantial exposure' classified exposures were different  in terms of  the frequency  of  exposure, and these differences  were not detectable from  the NIOSH JEM. This thesis could have involved recording the NIOSH JEM data in categories similar to that of Siemiatycki, e.g. by labelling confidence  of  an exposure as those JEM probabilities less than 0.5 as 'possible', those 0.5 - 1 as 'probable', and those equal to 1 as 'definite'.  However, this throws away information,  and the majority of  JEM estimates would be labelled 'possible' and those 'definite'  would be unreliable, as they would be based on small samples. Also, combining the confidence  codes of  each US job-translation possibilities for each Canadian job would be difficult. Siemiatycki's 'any exposure' and.'substantial exposure' definitions  constitute cut-off  values consistent across all agents tested. A consistent cut-off  value for  cumulative exposure across all agents in the study was considered inappropriate as the agents have different  exposure distributions. To apply to many agents, a very low cut-off  value with little practical meaning would be required. As there is no biological method to define  a cut-off  value for  each agent's cumulative exposure, cut-offs  based on the individual distribution of  exposure to each agent were considered. For example, this could involve labelling the top 10% exposures for  each agent as being 'exposed' and the remaining 90% as not. The exposure distributions of  Siemiatycki's study subjects should be very similar to those experi- enced by the BCCA subjects as they are both male Canadian subjects exposed over roughly the same period in history. A suggestion was that the proportions used for  Siemiatycki's definitions  of  'any exposure' and 'substantial exposure' could be transferred  to each agent in the current study. For example, 'substantial ex- posure' could refer  to the top 10% of  exposures and 'any exposure' to the top 30%. However, the proportions vary across chemicals and it would be very difficult  to find  equivalent proportions for  each of  the thousands of  chemicals studied by NIOSH, so an average was considered. Depending on which chemicals were averaged over and what Siemiatycki design configuration  was used (population or cancer controls, French Canadian population or all ethnicities); the proportion of  'any exposure' and 'substantial exposure' ranged from  1.5 - 2.9% and 4.8 - 8.2% respectively. However, applying any of  the cut-off  proportions in the ranges calculated to all agents in the study will result in some agents having patients classified  as exposed when they have zero cumulative exposure. The clearest distinction in the cumulative exposure estimates could be made between those with a value of zero (no exposure) and those with a value above zero (exposed). Given these two groups were different,  they were separated in the analysis of  the ever/never variable and the dose-response analysis. 8.4 Future Directions An immediate step to take is to assess the impact of  any measurement error in the cumulative exposure variables. The cumulative exposure assessments were composed of  the duration of  employment, obtained from  questionnaire responses, which are known to often  contain error, and the JEM probabilities of  exposure, which more importantly were often  based on small numbers and certainly contained a margin of  error. The positive cumulative exposures were normally distributed after  taking a Box-Cox transformation.  So, a multiplicative measurement error model could be fit  to the data and Bayesian methods used to assess the implications of  different  levels of  measurement error (Gustafson,  Le and Vallee, 2002). As most of  the uncertainty in the results is due to the JEM used, a more comprehensive one could be applied to the data in the future  if  one became available that contained more information  on the exposures, allowed for  changes in time, gave some measure of  error or variability, and studied exposures in more industries. Alternatively, industry specific  JEMs created in Canada, or ideally within the BCCA, could be applied in conjunction with the current NIOSH JEM to improve some of  the estimates or provide estimates for  the industries that were not studied. Other improvements beyond the JEM include conducting pairwise (industry and occupation code combined) translations of  the Canadian jobs that could include weightings as to which translations are more likely. Further information  could be sought via questionnaires on the patients, such as family  history of cancer, body weight, stress levels, fitness  levels, diet, etc. Finally, the 30 agents identified  in this study warrant further  research. Perhaps animal studies could be undertaken on exposure to the chemicals, or specific  cohort studies conducted to examine the relationships seen. Chapter 9 Bibliography 1. Ashby J, Tennant RW. Chemical structure, Salmonella mutagenicity and extent of  carcinogenicity as indicators of  genotoxic carcinogenesis among 222 chemicals tested in rodents by US NCI/NTP. Mutation  Research. 1988; 204:17-115. 2. Band PR, Le ND, MacArthur AC, Fang R, Gallagher RP. Identification  of  occupational cancer risks in British Columbia: A Population-based Case-Control Study of  Bladder Cancer. Unpublished. 3. Band PR, Spinelli JJ, Threlfall  WJ, Fang R, Le ND, and Gallagher RP. Identification  of  occupational cancer rpisks in British Columbia, Part I: Methodology, descriptive results, and analysis of  cancer risks by cigarette smoking categories of  15,463 incident cancer cases. Journal  of  occupational and environmental  medicine.  1999; 41:224-232. 4. Benjamini Y, Hochberg Y. Controlling the false  discovery rate: a practical and powerful  approach to multiple testing. Journal  of  the Royal Statistical  Society.  Series  B. 1995; Vol. 57: 289-300. 5. Box GEP, Cox DR. An analysis of  transformation.  Journal  of  the Royal Statistical  Society.  1964; Vol. 26: 211-243. 6. Breslow NE, Day NE. Statistical  Methods  in Cancer  Research: Volume  1 - The  analysis of  case-control studies.  International Agency for  Research on Cancer; 1980. 7. Canadian Cancer Statistics 2004. Toronto, Canada: National Cancer Institute of  Canada; 2004. 8. Census 1980. Bureau of  the Census alphabetical index to industries and occupations, 1980. Washing- ton, DC: U.S. Department of  Commerce, Bureau of  the Census; 1980. 9. Checkoway H, Pearce N, Crawford-Brown  DJ. Research Methods in Occupational Epidemiology. Ox- ford:  Oxford  University Press; 1989. 10. Doll R, Peto R. The  causes of  cancer. Quantitative  estimates of  avoidable  risks  of  cancer in the United States  today.  Oxford:  Oxford  University Press; 1981. 11. Dos Santos Silva 1. Cancer  Epidemiology:  Principles  and  Methods.  International Agency for  Research on Cancer; 1999. 12. Dun's  Marketing  Index.  Dun and Bradstreet, Inc.; 1980. 13. Gordis L. Epidemiology:  Second  Edition.  W.B. Saunders Company; 2000. 14. Gustafson  P, Le ND, Vallee M. A Bayesian approach to case-control studies with errors in covariables. Biostatistics.  2002; Vol 3: 229-243. 15. Hartge P, Hoover R, Altman R, et al. Use of  hair dyes and risk of  bladder cancer. Cancer  Research. 1982; 42:4784-4787. 16. Hochberg Y. A sharper Bonferroni  procedure for  multiple tests of  significance,  Biometrika.  1988; Vol. 75: 800-802. 17. Hosmer DW, Lemeshow S. Applied  Logistic Regression: Second  Edition.  Wiley; 2000. 18. IARC (fnternational  Agency for  Research on Cancer). Evaluation  of  the Carcinogenic  Risk of  Chemicals to Humans:  Diesel and  Gasoline Engine Exhausts and  Some Nitroarenes,  Vol.  46- IARC, Lyon; 1989. 19. IARC (International Agency for  Research on Cancer). Evaluation  of  the Carcinogenic  Risk of  Chemicals to Humans:  Occupational Exposures in Petroleum  Refining;  Crude  Oil and  Major  Petroleum  Fuels,  Vol. 45. IARC, Lyon; 1989. 20. IARC (International Agency for  Research on Cancer). Evaluation  of  the Carcinogenic  Risk of  Chemicals to Humans:  Some Chemicals  that Cause Tumours  of  the Kidney  or Urinary  Bladder  in Rodents  and Some Other Substances,  Vol.  73. IARC, Lyon; 1999. 21. IARC (International Agency for  Research on Cancer). Evaluation  of  the Carcinogenic  Risk of  Chemicals to Humans:  Tobacco  Smoking,  Vol.  38. IARC, Lyon; 1986. 22. IARC (International Agency for  Research on Cancer). IARC  Monographs  on the Evaluation  of  Car- cinogenic Risks to Humans,  Suppl.  7. Overall  evaluations of  carcinogenicity:  an updating  of  IARC Monographs,  Vol  1 to 42. IARC, Lyon; 1987. 23. JollifFe  IT. Principal  Component  Analysis. Springer-Verlag; 1986. 24. Kaiser, HF. The application of  electronic computers to factor  analysis. Educational  and  Psychological Measurement.  1960: 20, 141-151. 25. Kline P. An Easy Guide  to Factor  Analysis. Routledge; 1994. 26. Menard S. Applied  logistic  regression analysis. Sage publications; 2002. 27. Monson RR. Occupational Epidemiology.  CRC Press, Inc.; 1980. 28. National Cancer Institute of  Canada: Canadian Cancer Statistics 2004, Toronto, Canada, 2004. 29. Pott P. Chirurgical  Observations. London: Hawes, Clarke and Collins; 1775. 30. Randem BG, Langard S, Dale I, Kongerud J, Martinsen JI, Andersen A. Cancer incidence among male Norwegian asphalt workers. American Journal  of  Industrial  Medicine.  2003; 43(l):88-95. 31. Rothman JR, Greenland S. Modern  Epidemiology:  Second  Edition.  Lippincott Williams & Wilkins; 1998. 32. Seta JA, Sundin DS, Pedersen DH. National  occupational exposure survey field  guidelines.  Vol.  I. Cincinnati, OH: U.S. Department of  Health and Human Services, Centers for  Disease Control, National Institute for  Occupational Safety  and Health, DHHS (NIOSH); 1988. Publication No. 88-106. 33. Schottenfeld  D, Fraumeni JF. Cancer  Epidemiology  and  Prevention: Second  Edition.  Oxford  University Press; 1996. 34. Schrier D, Ip F. British Columbia's  Changing  Ethnic Mosaic.  Statistics Canada, Ottawa; 1994. 35. Sieber WK. National  occupational exposure survey sampling methodology.  Vol.  II.  Cincinnati, OH: U.S. Department of  Health and Human Services, Centers for  Disease Control, National Institute for Occupational Safety  and Health, DHHS (NIOSH); 1989. Publication No. 89-102. 36. Siemiatycki J. Risk Factors  for  Cancer  in the Workplace.  CRC Press; 1991. 37. Silverman DT, Levin LI, Hoover RN, et al. Occupational risks of  bladder cancer in the United States: I. White Men. Journal  of  the National  Cancer  Institute.  1989; 81:1472-1480. 38. Soto AM, Justicia H, Wray JW, and Sonnenschein, C. p-Nonylphenol, an estrogenic xenobiotic released from  'modified'  polystyrene. Environmental  Health  Perspectives.  1991; 92:167-173. 39. Standard  Occupational Classification  1980. Ottawa, Ontario, Canada: Statistics Canada; 1981. 40. Standard  Industrial  Classification  1980. Ottawa, Ontario, Canada: Statistics Canada; 1981. 41. Svirchev LM, Kan D, Lai AM, Ng V, Moody JM, Band PR. Translations of  the United States industrial and occupational classifications  into the Canadian standard industrial and occupational classifications. British Columbia  Cancer  Agency Monograph  No.  4• 1993. 42. Tomatis L, Aitio A, Day NE, Heseltine E, Kaldor J, Miller AB, Parkin M, Riboli E. Cancer:  Causes, Occurrence and  Control.  IARC Publications No. 100, International Agency for  Research on Cancer, Lyon; 1990. 43. Warhurst AM. An environmental  assessment of  alkylphenol  ethoxylates  and  alkylphenols.  Warhurst; 1995. 44. World Health Organization. International  Classification  of  Diseases: Ninth  Revision. Geneva: WHO; 1977. Chapter 10 Tables and Figures Table 2.1: Two by Two Contingency Table for  Cohort Studies Diseased Not Diseased Total Sample Exposed Si A1+B1 Not Exposed Aa Bo A0 + B0 Table 2.2: Two by Two Contingency Table for  Case-Control Studies Diseased Not Diseased Exposed <Zl bi Not Exposed o 0 bo Total Sample ai + ao h+bo Occupation or Industry IARC" Suspected agents Aluminium production 1 Pitch violates, coal-tar pitch volatiles, aromatic amines Aromatic amine manufactur- N/A 2-naphthylamine, benzidine, 4-aminobiphenyl. Possibly: ing workers MDA (4,4-methylene-dianiline), MBOCA (4,4-methylene- bis(2-chloroaniline), 4-cholor-o-toluidine (4-COT). Boot and shoe manufacture 1 Leather dust, dyes, benzene and other solvents and repair Leather workers 3 Leather dust, dyes, solvents Coal gasification 1 Coal tar, coal-tar fumes,  individual PAHs Coke production 1 Coal-tar fumes,  polynuclear aromatic hydrocarbons (PAHs) Drivers of  trucks and other N/A Motor exhaust (polycyclic aromatic hydrocarbons, nitro- motor vehicles PAHs) Dry cleaning solvent-exposed 2B Benzene, naphtha, gasoline, stoddard solvent (mineral or workers white spirits), carbon tetrachloride, trichloroethylene, tetra-' chloroethylene, chlorofluorocarbon  solvents, chlorinated sol- vents, amyl acetate, bleaching agents, acetic acid, aqueous am- monia, oxalic acid, hydrogen peroxide and dilute hydrogen flu- oride solutions Dyestuffs  workers and dye N/A 3 aromatic amines (2-naphthylamine, benzidine, 1- users naphthylamine), o-toluidine, 4,4-methylene bis(2- methylaniline). Auramine manufacture 1 2-naphthylamine, auramine, other chemicals Magenta manufacture 1 Magenta, ortho-toluidine, 4,4-methylene bis(2-methylaniline), ortho-nitrotoluene Hairdresser or barber 2A Some compounds in hair dyes, aromatic amines, aminophe- nols, hydrogen peroxide, aminoanthraquinones, azo dyes, lead acetate, volatile solvents, propellants, aerosols, formaldehyde, methacrylates Painters 1 Paints (benzidine, polychlorinated biphenyls, formaldehyde, asbestos) and solvents (benzene, dioxane, methylene chloride). Petroleum refining 2A Aliphatic hydrocarbons, aromatic hydrocarbons, hydrogen sul- fide,  polycyclic aromatic compounds Printing processes 2B Carbon black, titanium dioxide, azo, anthraquinone and tri- arylmethane dyes, and phthalocyanines Rubber industry 1 Aromatic amines, solvents, 2-naphthylamine, phenyl-b- naphthylamine (PBNA). Textile manufacturing 2B Textile-related dusts, dyes, optical brighteners, organic sol- vents and fixatives,  benzidine, formaldehyde,  flame  retardants (including organophosphorus and organobromine compounds) a IARC classification  where 1 is definitely  carcinogenic, 2A is probably carcinogenic, 2B is possibly carcinogenic, and 3 is not classifiable Siemiatycki ORs IARC Chemical Name Any Substantial Class" Evidence 1,3-Dichloropropene 2B Animal 2-(2-Formylhydrazino)-4(5N2F)T6 2B Animal 2-Naphthylamine 1 Human 2-Nitroanisole 2B Animal 3,3'-Dichlorobenzidine 2B Human 3,3'-Dimethoxybenzidine 2B Human 4,4'-Methylenebis(2-chloroaniline) 2A Animal Adriamycin 2A Animal 4-Aminobiphenyl (xenylamine) 1 Human 4-chloro-ortho- phenylenediamine 2B Animal, Arsenic 1 Human Auramine 2B Human Benz(a)anthracene 2A Animal Benzidine 1 Human Benzidine based dyes 2A Animal Carbon black 2.2* 1.8 2B Human Chlordane 2B Human Chloroform  (in drinking water) t t 2B Human CI Basic Red 9 2B Animal Citrus Red No. 2 2B Animal Coal tar pitches 0.9 2.3 1 Human Cyclophosphamide 1 Human Diesel engine emissions 1.4 2.3** 2A Human Disperse Blue 1 2B Animal Engine emissions 1.2* 1.3* 2A Human Gasoline 1.1 0.9 2B Human Lead 2B Human Magenta 1 Human Mineral oils6 1.2 2.2 1 Human N- [4- (5-Ni tro-2-Furyl) 2TZ] Ab 2B Animal Niridazole 2B Animal Nitrilotriacetic Acid 2B Animal N,N-Bis(2-CE)-2-NL6 (Chlornaphazine) 1 Human N-Nitrosodi-n-butylamine 2B Animal Oil Orange SS 2B Animal ortho-Aminoazotoluene 2B Animal para-Chloro-ortho-Toluidine 2A Human para-Cresidine 2B Animal para-Dimethylaminobenzene 2B Animal Phenacetin 2A Human Ponceau 3R 2B Animal Sodium ortho-phenylphenate 2B Animal Tetrachloroethylene t t 2A Human Trichloroethylene 0.6 0.7 2A Human a IARC classification  where 1 is definitely  carcinogenic, 2A is probably carcinogenic, 2B is possibly carcinogenic, and 3 is not classifiable b 2-(2-FormyIhydrazino)-4(5N2F)T = 2-(2-Formylhydrazino)-4-(5-Nitro-2-Furyl) Thiazole, N-[4-(5- Nitro-2-Furyl)2TZ]A = N-[4-(5-Nitro-2-Furyl)-2-Thiazolyl] Acetamide. Mineral oils = Minerail oils, untreated or mildly treated. N,N-Bis(2-CE)-2-NL = N,N-Bis(2-Chloroethyl)-2-naphthylamine f  Less than 4 cases exposed * Significant  at p=0.10, one-sided, with at least 4 exposed cases ** Significant  at p=0.05, one-sided, with at least 5 exposed cases Chemical Name Siemiatycki ORs IARC Any Substantial Class' 1 Evidence Acrylic fibres 3.9** 3.3 Aliphatic aldehydes 1.4* 1.6 Ammonia 1.2 2.1* Asphalt (bitumen) 0.9 2.2* 3 Animal Cadmium compounds 1.6 4.9* 1 Human Calcium carbonate 1.9** 1.6 Carbon black 2.2* 1.8 2B Human Carbon tetrachloride 1.6 2.5** 2B Human Chlorine 1 2.7* Clay dust 2.2* 1.8 Creosote 2.6* 2.6 Diesel engine emissions 1.4 2.3** 2A Human Engine emissions 1.2* 1.3* 2A Human Fabric dust 1 3.7* Formaldehyde 1.2 1.7* 2A Human Hydrogen cyanide 3.4* 0 Ionizing radiation 44** 0 1 Human Laboratory products 1.5 5.5* Lead chromate 1.8* 2.2 2B Human Lead compounds 1.3* 1.1 2B Human Natural gas comb, products 1.6* 3.8** Photographic products 2.5 2.9* Polyester fibers 1.4 2.5* Polyethylene 2.5* 13 3 Animal Titanium compounds ^ 7** 2.2 Titanium dioxide ^ y** 4.5 3 Animal a IARC classification  where 1 is definitely  carcinogenic, 2A is probably carcinogenic, 2B is possibly carcinogenic, and 3 is not classifiable * Significant  at p=0.10, one-sided, with at least 4 exposed cases ** Significant  at p=0.05, one-sided, with at least 5 exposed cases Table 3.4: Siemiatycki Exposure Coding Code Confidence Concentration Frequency 1 2 3 Possible Exposure Probable Exposure Definite  Exposure Low: background level Medium: intermediate situations High: agent in concentrated form Low: 1-5% of  working time Medium:5-30% of  working time High: >30% of  working time Figure 5.1: Calculating Cumulative Exposure Analysis Design Table 5.1: Distribution of  Missing Job Code, and Start and End Year Data Start and End Years Job Codes Total All Missing Some Missing All Complete All Missing 18 7 111 136 Some Missing 0 27 389 416 All Complete 17 136 14,049 14,202 Total 35 170 14,549 14,754 Table 5.2: Cancer Site Distribution of  Jobs in the Armed Forces Armed Forces Employment Total Work-Years Primary tumour site IDC-9a Ever % Never % Armed Forces % Other % Oral cavity and pharynx 140-149 78 2.9 482 4.2 441.0 2.6 17,224.5 3.7 Esophagus 150 31 1.2 145 1.3 145.0 0.9 6,122.5 1.3 Stomach 151 111 4.2 459 4.0 663.5 4.0 19,646.0 4.2 Colon 153 224 8.4 881 7.7 1,299.0 7.8 39,553.5 8.5 Rectum 154 202 7.6 849 7.5 1,169.0 7.0 37,396.0 8.0 Liver 155 4 0.1 40 0.4 17.0 0.1 1,313.5 0.3 Pancreas 157 27 1.0 110 1.0 203.0 1.2 4,619.0 1.0 Larynx 161 68 2.5 244 2.1 398.0 2.4 10,482.0 2.2 Lung 162 618 23.2 2,195 19.3 4,061.0 24.3 96,771.5 20.7 Soft  tissue sarcoma 171 18 0.7 136 1.2 103.0 0.6 3,782.0 0.8 Melanoma skin 172 89 3.3 557 4.9 595.0 3.6 17,006.5 3.6 Non-melanoma skin 173 288 10.8 924 8.1 1,996.5 12.0 41,110.0 8.8 Prostate 185 294 11.0 1,161 10.2 1,761.5 10.6 55,460.0 11.9 Testis 186 13 0.5 213 1.9 69.0 0.4 2,797.5 0.6 Bladder 188 218 8.2 848 7.5 1,468.5 8.8 37,678.5 8.1 Kidney 189 92 3.4 461 4.1 524.0 3.1 18,360.5 3.9 Brain 191 39 1.5 288 2.5 309.0 1.9 8,119.5 1.7 Hodgkin's disease 201 4 0.1 104 0.9 15.0 0.1 1,781.0 0.4 Non-Hodgkin's lymphoma 202 128 4.8 626 5.5 821.5 4.9 23,159.0 5.0 Multiple myeloma 203 24 0.9 104 0.9 140.0 0.8 4,558.0 1.0 Leukemia 204-208 33 1.2 205 1.8 153.0 0.9 6,994.0 1.5 Other sites - 64 2.4 350 3.1 333.0 2.0 12,517.5 2.7 Total 2,667 100.0 11,382 100.0 16,685.5 100.0 466,452.5 100.0 a IDC-9, International Classification  of  Diseases, 9th Revision Industrial Group CDN Code" Jobs % Work- Years % Ever Industry % Agriculture 01-02 v 4,845 8.0 46,067.5 9.9 3,557 25.6 Fishing, Trapping 03 652 1.1 5,317.0 1.1 513 3.7 Logging, Forestry 04-05 2,945 4.9 17,216.0 3.7 1,759 12.6 Mining, Quarrying, Oil Well 06-09 2,802 4.6 14,093.5 3.0 1,545 11.1 Manufacturing Food, beverage, tobacco 10-12 1,694 2.8 12,449.5 2.7 1,100 7.9 Rubber, plastic, leather, 15-24 419 0.7 3,023.0 0.6 276 2.0 textile, clothing Wood, furniture,  paper, 25-28 5,616 9.3 43,103.0 9.2 3,377 24.3 printing Other 29-39 5,969 9.8 40,885.0 8.8 3,988 28.7 Construction 40-44 6,591 10.9 46,339.0 9.9 4,084 29.3 Transportation 45-47 5,690 9.4 46,492.0 10.0 3,249 23.3 Communication, Utility 48-49 1,707 2.8 16,247.0 3.5 1,035 7.4 Wholesale 50-59 3,345 5.5 24,543.0 5.3 2,108 15.1 Retail 60-69 5,270 8.7 39,676.5 8.5 3,327 23.9 Finance, Insurance, Real Estate 70-76 2,101 3.5 16,657.5 3.6 1,306 9.4 Services Business 77 1,297 2.1 10,399.5 2.2 827 5.9 Government 81-84 3,900 6.4 35,600.5 7.6 2,633 18.9 Education, Health 85-86 2,455 4.0 23,524.0 5.0 1,528 11.0 Other 91-99 3,411 5.6 24,812.0 5.3 2,404 17.3 Total 60,709 100.0 466,445.5 100.0 a Canadian 1980 Standard Industrial Classification Industrial Group US Code" Job-translations % Work-Years* % Agriculture 010-020 19,827 9.3 46,138.1 9.9 Fishing, Hunting, Trapping 031 800 0.4 5,317.0 1.1 Forestry 030 419 0.2 1,192.0 0.3 Mining 040-050 10,053 4.7 14,093.5 3.0 Manufacturing Food, beverage, tobacco 100-130 6,360 3.0 12,370.5 2.7 Rubber, plastic, leather, 132-150, 2,213 1.0 3,115.9 0.7 textile, clothing 210-220 Wood, furniture,  paper, 160-170, 21,999 10.3 59,109.2 12.7 printing 230-240 Other 180-200, 25,968 12.1 40,656.8 8.7 250-390 Construction 060 17,943 8.4 45,884.8 9.8 Transportation 400-430 14,574 6.8 50,119.9 10.7 Communications, Utilities 440-470 6,165 2.9 13,233.3 2.8 Wholesale 500-570 13,526 6.3 20,865.0 4.5 Retail 580-690 24,212 11.3 37,350.0 8.0 Finance, Insurance, Real Estate 700-712 3,984 1.9 16,667.5 3.6 Services Business, Repair 721-760 12,099 5.6 18,578.4 4.0 Public Administration 900-932 10,876 5.1 33,993.3 7.3 Professional 812-892 3,104 1.4 9,567.5 2.1 Personal Services 761-791 3,920 1.8 3,814.0 0.8 Entertainment, Recreation 800-802 16,147 7.5 34,378.9 7.4 Total 214,189 100.0 466,445.5 100.0 a US 1980 Census of  the Population Industrial Classification b Work-years contributed Occupational Group CDN Code" Jobs % Work- Years % Ever Occupation % Managerial and Administrative 11 5,923 9.8 54,044 11.6 3,099 22.3 Natural Sciences, Engineering, 21 2,005 3.3 14,034 3.0 959 6.9 Mathematics Social Sciences 23 298 0.5 2,759 0.6 191 1.4 Religion 25 179 0.3 1,966 0.4 81 0.6 Teaching 27 965 1.6 8,511 1.8 589 4.2 Medicine and Health 31 735 1.2 9,208 2.0 454 3.3 Artistic, Literary, Recreational 33 686 1.1 5,534 1.2 423 3.0 Clerical 41 3,770 6.2 25,223 5.4 2,257 16.2 Sales 51 5,470 9.0 42,231 9.1 3,076 22.1 Services 61 3,672 6.0 27,392 5.9 2,344 16.8 Farming, Horticultural, Animal 71 4,929 8.1 46,544 10.0 3,583 25.7 Husbandry Fishing, Trapping 73 523 0.9 4,631 1.0 442 3.2 Forestry, Logging 75 2,229 3.7 13,028 2.8 1,454 10.4 Mining, Quarrying 77 1,568 2.6 7,213 1.5 927 6.7 Materials Processing 81-82 4,629 7.6 32,819 7.0 3,187 22.9 Machining 83 2,314 3.8 16,709 3.6 1,326 9.5 Product Fabricating, Assem- 85 6,020 9.9 43,966 9.4 3,224 23.2 bling, Repairing Construction 87 6,560 10.8 48,779 10.5 3,580 25.7 Transport Equipment Operat- 91 4,998 8.2 37,180 8.0 2,937 21.1 ing Material Handling 93 1,715 2.8 12,019 2.6 1,316 9.5 Other Crafts  and Equipment 95 1,207 2.0 10,902 2.3 734 5.3 Operating Not Elsewhere Classified 99 314 0.5 1,761 0.4 290 2.1 Total 60,709 100.0 466,446 100.0 ° Canadian 1980 Standard Occupational Classification Occupational Group US Code a Job- translations % Work-Years 6 % Executive, Administrative, Man- 003-037 11,991 5.6 54,260 11.6 agerial Natural Sciences, Engineering, 043-083, 9,446 4.4 21,355 4.6 Mathematics 213-235 Social Sciences 166-175, 485 0.2 2,479 0.5 178-179 Religion 176-177 179 0.1 1,966 0.4 Teaching 113-163 12,196 5.7 8,686 1.9 Medicine and Health 084-106, 572 0.3 7,092 1.5 203-208 Artistic, Literary, Recreational 164-165, 1,528 0.7 5,029 1.1 183-199 Administrative Support, Clerical 303-389 11,375 5.3 26,286 5.6 Sales 243-285 25,168 11.8 37,630 8.1 Services 403-469 10,733 5.0 27,276 5.8 Farming 473-489 18,337 8.6 45,275 9.7 Fishing, Trapping 497-499 523 0.2 4,631 1.0 Forestry, Logging 494-496 2,188 1.0 10,232 2.2 Extractive occupations 613-617 2,182 1.0 2,609 0.6 Precision production 633-699 11,542 5.4 24,696 5.3 Machine Operators 703-779 16,495 7.7 30,883 6.6 Fabricators, Assemblers, Me- 503-549, 31,144 14.5 43,120 9.2 chanics, Repairers 783-799 Construction 553-599 16,531 7.7 39,790 8.5 Transport Equipment Operating 803-834 9,749 4.6 36,252 7.8 Material Moving 843-859 9,006 4.2 17,203 3.7 Handlers, Equipment Cleaners, 863-889 12,819 6.0 19,697 4.2 Helpers, Laborers Total 214,189 100.0 466,446 100.0 a US 1980 Census of  the Population Occupational Classification 6 Work-years contributed Industry Group CDN Code" Jobs Work-Years J E M All % J E M 0 All % Agriculture 01-02 56 4,845 1.2 328.7 46,067.5 0.7 Fishing, Trapping 03 0 652 0.0 0.0 5,317.0 0.0 Logging, Forestry 04-05 2,167 2,945 73.6 11,128.9 17,216.0 64.6 Mining, Quarrying, Oil Well 06-09 151 2,802 5.4 528.2 14,093.5 3.7 Manufacturing Food, beverage, tobacco 10-12 1,311 1,694 77.4 7,205.2 12,449.5 57.9 Rubber, plastic, leather, 15-24 321 419 76.6 1,427.6 3,023.0 47.2 textile, clothing Wood, furniture,  paper, 25-28 5,065 5,616 90.2 29,067.3 43,103.0 67.4 printing Other 29-39 4,972 5,969 83.3 23,774.5 40,885.0 58.1 Construction 40-44 6,351 6,591 96.4 32,702.7 46,339.0 70.6 Transportation 45-47 3,904 5,690 68.6 24,258.6 46,492.0 52.2 Communication, Utility 48-49 1,119 1,707 65.6 6,049.9 16,247.0 37.2 Wholesale 50-59 962 3,345 28.8 2,522.0 24,543.0 10.3 Retail 60-69 1,541 5,270 29.2 6,373.7 39,676.5 16.1 Finance, Insurance, Real Estate 70-76 0 2,101 0.0 0.0 16,657.5 0.0 Services Business 77 204 1,297 15.7 564.1 10,399.5 5.4 Government 81-84 24 3,900 0.6 109.2 35,600.5 0.3 Education, Health 85-86 547 2,455 22.3 4,170.3 23,524.0 17.7 Other 91-99 611 3,411 17.9 3,407.9 24,812.0 13.7 Total 29,306 60,709 48.3 153,618.6 466,445.5 32.9 a Canadian 1980 Standard Industrial Classification 6 Work-years contributed by the JEM Occupation Group CDN Codea Jobs Work-Years JEM All % JEM6 All % Managerial and Administrative 11 1,237 5,923 20.9 8,765.2 54,043.5 16.2 Natural Sciences, Engineering, 21 580 2,005 28.9 2,676.0 14,034.0 19.1 Mathematics Social Sciences 23 11 298 3.7 21.4 2,758.5 0.8 Religion 25 0 179 0.0 0.0 1,965.5 0.0 Teaching 27 0 965 0.0 0.0 8,510.5 0.0 Medicine and Health 31. 307 735 41.8 2,901.7 9,207.5 31.5 Artistic, Literary, Recreational 33 288 686 42.0 1,602.7 5,534.0 29.0 Clerical 41 1,114 3,770 29.5 4,258.5 25,223.0 16.9 Sales 51 854 5,470 15.6 3,002.0 42,231.0 7.1 Services 61 609 3,672 16.6 2,702.3 27,391.5 9.9 Farming, Horticultural, Animal 71 94 4,929 1.9 233.1 46,543.5 0.5 Husbandry Fishing, Trapping 73 0 523 0.0 0.0 4,630.5 0.0 Forestry, Logging 75 1,726 2,229 77.4 9,784.8 13,027.5 75.1 Mining, Quarrying 77 141 1,568 9.0 354.8 7,213.0 4.9 Materials Processing 81-82 4,054 4,629 87.6 25,009.9 32,818.5 76.2 Machining 83 2,066 2,314 89.3 9,377.9 16,708.5 56.1 Product Fabricating, Assem- 85 4,753 6,020 79.0 20,351.6 43,965.5 46.3 bling, Repairing Construction 87 5,915 6,560 90.2 29,485.2 48,779.0 60.4 Transport Equipment Operat- 91 3,280 4,998 65.6 20,650.8 37,179.5 55.5 ing Material Handling 93 1,331 1,715 77.6 6,177.4 12,019.0 51.4 Other Crafts  and Equipment 95 760 1,207 63.0 5,460.4 10,901.5 50.1 Operating Not Elsewhere Classified 99 186 314 59.2 802.9 1,760.5 45.6 Total 29,306 60,709 48.3 153,618.6 466,445.5 32.9 a Canadian 1980 Standard Occupational Classification 6 Work-years contributed by the JEM Table 5.9: Location of  Canadian Jobs Found on the JEM Location Jobs % Work Years" % Alberta 1,723 5.9 7,036.0 4.6 British Columbia 19,562 66.8 109,677.8 71.4 Manitoba 819 2.8 3,528.0 2.3 New Brunswick 82 0.3 251.3 0.2 Newfoundland 29 0.1 91.2 0.1 Northwest Territories 34 0.1 74.1 0.0 Nova Scotia 60 0.2 211.3 0.1 Ontario 1,454 5.0 5,807.1 3.8 Prince Edward Island 14 0.0 25.5 0.0 Quebec 436 1.5 1,983.0 1.3 Saskatchewan 884 3.0 3,717.5 2.4 Yukon Territories 64 0.2 185.8 0.1 Canada 191 0.7 1,159.4 0.8 Canada + BC 252 0.9 1,805.9 1.2 Canada + elsewhere 124 0.4 949.2 0.6 Outside of  Canada 2,346 8.0 11,141.1 7.3 Unknown 1,232 4.2 5,974.5 3.9 Total 29,306 100.0 153,618.6 100.0 a Work-years contributed on the JEM Table 6.1: Characterisitics of  Cases and Controls Characteristic Cases (n  — 1062) Controls (n = 8057) Patients % Mean (± SD) Patients % Mean (± SD) Age at diagnosis, years 67.0 (11.4) 65.9 (10.9) Employment duration, work-years0. 36.7 (11.0) 35.9 (11.2) No jobs reported0 3 0.3 24 0.3 Year of  diagnosis 1983 222 20.9 2600 32.3 1984 215 20.2 1801 22.4 1985 221 20.8 1475 18.3 1986 216 20.3 1088 13.5 1987 188 17.7 1093 13.6 Ethnic origin Caucasian 1027 96.7 7665 95.1 Non-Caucasian 31 2.9 350 4.3 Unknown 4 0.4 42 0.5 Marital status Single 42 4.0 385 4.8 Married or common-law 889 83.7 6706 83.2 Widowed 66 6.2 492 6.1 Separated or divorced 57 5.4 403 5.0 Unknown 8 0.8 71 0.9 Education <8 years 118 11.1 894 11.1 8-11 years 480 45.2 3583 44.5 High school graduate 119 11.2 884 11.0 Post secondary education 298 28.1 2305 28.6 Unknown 47 4.4 391 4.9 Years 10.0 (2.3) 10.0 (2.3) Tobacco smoking Never smoker 117 11.0 1,444 17.9 Ever smoker Pipe and cigar only 34 3.2 323 4.0 Cigarette only 909 85.6 6,268 77.8 Unknown 2 0.2 22 0.3 Cigarette smoking only Current smoker 314 34.5 1,894 30.2 Former smoker 564 62.0 4,118 65.7 Unknown 31 3.4 256 4.1 Cigarette smoking only Cigarettes/day 21.3 (12.7) 20.9 (12.5) Years/smoked 36.5 (14.9) 33.5 (15.0) Pack-years 33.6 (29.0) 27.9 (28.3) Years quit (former  smokers) 16.8 (11.9) 18.3 (12.6) Alcohol consumption Never 113 10.6 842 10.5 Ever 811 76.4 5882 73.0 Unknown 138 13.0 1333 16.5 Person completing questionnaire Patient 888 83.6 6357 78.9 Other 150 14.1 1490 18.5 Unknown 24 2.3 210 2.6 a Prior to 5 years before  diagnosis Table 6.2: Odds Ratios (OR) for  Potentially Important11 Confounding  Variables Confounding  Variable No. of  Cases OR 95% Confidence  Interval Respondent to questionnaire Patient 888 1.00 - Proxy 150 0.65 0.53 - 0.78 Unknown 24 0.92 0.59 - 1.42 Ethnic origin Caucasian 1,027 1.00 - Non-Caucasian 31 0.71 0.48 - 1.05 Unknown 4 0.63 0.22 - 1.79 Alcohol consumption status Never drinker 113 1.00 - Ever drinker 811 0.88 0.70 - 1.11 Unknown 138 1.20 0.87- 1.67 Cigarette smoking duration, years 0 151 1.00 - 1-29 262 1.41 1.13 - 1.75 30-44 338 1.93 1.56 - 2.40 45+ 300 2.36 1.89 - 2.95 Unknown 11 1.16 0.60 - 2.23 a p-value < 20% Table 6.3: Log Likelihood for  Various Base Models Degrees of  Deviance from Model Variables Freedom -2LLa Base Model p-value Base Who  completed  questionnaire, ethnicity, alcohol  status, cigarette  years 10 5,433 ~ - 1 Who completed questionnaire, alcohol status, cigarette years 8 5,437 4.00 0.14 2 Who completed questionnaire, ethnicity, alcohol status, smoking status 8 5,468 NA NA 3 Who completed questionnaire, ethnicity, alcohol status, cigarette pack-years 10 5,448 NA NA 4 Who completed questionnaire, ethnicity, alcohol status, years quitsmoking 11 5,440 NA NA a LL = Log likelihood All Subjects Complete Occupational Data Characteristic Cases (n = 1125) No. (%) Controls (n  = 8492) No. (%) Cases (n  = 1062) No. (%) Controls (n  = 8057) No. (%) No jobs reported0 8 (0.7) 62 (0.7) 3 (0.3) 24 (0.3) Year of  diagnosis 1983 240 (21.3) 2,715 (32.0) 222 (20.9) 2,600 (32.3) 1984 229 (20.4) 1,907 (22.5) 215 (20.2) 1,801 (22.4) 1985 228 (20.3) 1,534 (18.1) 221 (20.8) 1,475 (18.3) 1986 231 (20.5) 1,181 (13.9) 216 (20.3) 1,088 (13.5) 1987 197 (17.5) 1,155 (13.6) 188 (17.7) 1,093 (13.6) Ethnic origin Caucasian 1,088 (96.7) 8,073 (95.1) 1,027 (96.7) 7,665 (95.1) Non-Caucasian 32 (2.8) 370 (4.4) 31 (2.9) 350 (4.3) Unknown 5 (0.4) 49 (0.6) 4 (0.4) 42 (0.5) Marital status Single 45 (4.0) 415 (4.9) 42 (4.0) 385 (4.8) Married or common-law 933 (82.9) 7,014 (82.6) 889 (83.7) 6,706 (83.2) Widowed 74 (6.6) 532 (6.3) 66 (6.2) 492 (6.1) Separated or divorced 65 (5.8) 448 (5.3) 57 (5.4) 403 (5.0) Unknown 8 (0.7) 83 (1.0) 8 (0.8) 71 (0.9) Education <8 years 123 (10.9) 972 (11.4) 118 (11.1) 894 (11.1) 8-11 years 508 (45.2) 3,781 (44.5) 480 (45.2) 3,583 (44.5) High school graduate 129 (11.5) 926 (10.9) 119 (11.2) 884 (11.0) Post secondary education 312 (27.7) 2,384 (28.1) 298 (28.1) 2,305 (28.6) Unknown 53 (4.7) 429 (5.1) 47 (4.4) 391 (4.9) Tobacco smoking Never smoker 123 (10.9) 1,503 (17.7) 117 (11.0) 1,444 (17.9) Ever smoker Pipe and cigar only 35 (3.1) 342 (4.0) 34 (3.2) 323 (4.0) Cigarette only 965 (85.8) 6,621 (78.0) 909 (85.6) 6,268 (77.8) Unknown 2 (0.2) 26 (0.3) 2 (0.2) 22 (0.3) Cigarette smoking only Current smoker 332 (34.4) 2,020 (30.5) 314 (34.5) 1,894 (30.2) Former smoker 600 (62.2) 4,326 (65.3) 564 (62.0) 4,118 (65.7) Unknown 33 (3.4) 275 (4.2) 31 (3.4) 256 (4.1) Alcohol consumption Never 119 (10.6) 881 (10.4) 113 (10.6) 842 (10.5) Ever 858 (76.3) 6,201 (73.0) 811 (76.4) 5,882 (73.0) Unknown 148 (13.2) 1,410 (16.6) 138 (13.0) 1333 (16.5) Person completing questionnaire Patient 934 (83.0) 6,644 (78.2) 888 (83.6) 6357 (78.9) Other 164 (14.6) 1,630 (19.2) 150 (14.1) 1490 (18.5) Unknown 27 (2.4) 218 (2.6) 24 (2.3) 210 (2.6) a Prior to 5 years before  diagnosis Table 6.4: Continued All Subjects Complete Occupational Data Cases Controls Cases Controls (n  = 1125) (n  = 8492) (n = 1062) (n = 8057) Characteristic Mean (SD) Mean (SD) Mean (SD) Mean (SD) Age at diagnosis, years 67.3 (11.4) 66.0 (10.9) 67.0 (11.4) 65.9 (10.9) Employment duration, work-years" 36.4 (11.3) 35.6 (11.6) 36.7 (11.0) 35.9 (11.2) Years of  Education 9.9 (2.3) 10.0 (2.3) 10.0 (2.3) 10.0 (2.3) Cigarette smoking only Cigarettes/day 21.2 (12.6) 20.9 (12.5) 21.3 (12.7) 20.9 (12.5) Years smoked 36.6 (14.9) 33.6 (15.0) 36.5 (14.9) 33.5 (15.0) Pack-years 33.5 (29.1) 28.1 (28.4) 33.6 (29.0) 27.9 (28.3) Alcohol score 416.7 (678.9) 422.7 (640.1) 411.7 (662.0) 415.5 (613.2) Former smokers only Years quit 16.1 (12.2) 17.6 (12.9) 16.8 (11.9) 18.3 (12.6) a Prior to 5 years before  diagnosis Table 6.5: Distribution of  Control Cancer Sites Before  and After  Exclusions Complete All Subjects Occupational Data Primary tumour site IDC-9a Patients % Patients % Oral cavity and pharynx 140-149 524 6.2 479 5.9 Esophagus 150 176 2.1 159 2.0 Stomach 151 353 4.2 330 4.1 Colon 153 1,101 13.0 1,044 13.0 Rectum 154 892 10.5 841 10.4 Liver 155 39 0.5 36 0.4 Pancreas 157 138 1.6 129 1.6 Larynx 161 304 3.6 284 3.5 Soft  tissue sarcoma 171 113 1.3 106 1.3 Melanoma skin 172 479 5.6 460 5.7 Non-melanoma skin 173 1,121 13.2 1,091 13.5 Prostate 185 1,479 17.4 1,415 17.6 Testis 186 91 1.1 86 1.1 Kidney 189 336 4.0 320 4.0 Brain 191 159 1.9 149 1.8 Hodgkin's disease 201 57 0.7 56 0.7 Non-Hodgkin's lymphoma 202 438 5.2 416 5.2 Multiple myeloma 203 123 1.4 116 1.4 Leukemia 204-208 211 2.5 201 2.5 Other sites - 358 4.2 339 4.2 Total 8,492 100.0 8,057 100.0 a IDC-9, International Classification  of  Diseases, 9th Revision Table 6.6: Odds Ratios (OR) for  Potentially Important" Confounding  Variables Before  and After  Exclusions All Subjects Complete Occupational Data Confounding  Variable Cases OR 95% CP Cases OR 95% CI6 Respondent to questionnaire Patient 934 1.00 - 888 1.00 - Proxy 164 0.64 0.53 - 0.77 150 0.65 0.53 - 0.78 Unknown 27 0.99 0.65 - 1.50 24 0.92 0.59 - 1.42 Ethnic origin Caucasian 1088 1.00 - 1027 1.00 - Non-Caucasian 32 0.68 0.47 - 1.00 31 0.71 0.48 - 1.05 Unknown 5 0.66 0.26 - 1.69 4 0.63 0.22 - 1.79 Alcohol consumption status Never drinker 119 1.00 - 113 1.00 - Ever drinker 858 0.88 0.70 - 1.10 811 0.88 0.70 - 1.11 Unknown 148 1.16 0.85 - 1.59 138 1.20 0.87- 1.67 Smoking duration, years 0 159 1.00 - 151 1.00 - 1-29 277 1.43 1.15 - 1.77 262 1.41 1.13 - 1.75 30-44 355 1.93 1.56 - 2.38 338 1.93 1.56 - 2.40 45+ 322 2.35 1.90 - 2.92 300 2.36 1.89 - 2.95 Unknown 12 1.08 0.58 - 2.02 11 1.16 0.60 - 2.23 a p-value < 20% b CI = Confidence  Interval Table 6.7: Distribution of  Bladder Cases Exposed Across the 8,986 Agents Cases Cumulative Cumulative Exposed Agents Frequency Percentage Percentage 201+ 539 539 6.0 6.0 101-200 512 1,051 5.7 11.7 21-100 1,282 2,333 14.3 26.0 10-20 955 3,288 10.6 36.6 9 162 3,450 1.8 38.4 8 275 3,725 3.1 41.5 7 340 4,065 3.8 45.2 6 319 4,384 3.5 48.8 5 552 4,936 6.1 54.9 4 332 5,268 3.7 58.6 3 431 5,699 4.8 63.4 2 1,029 6,728 11.5 74.9 1 2,258 8,986 25.1 100.0 Table 7.1: Distribution of  p-values for  Ever Exposure of  5,699 Agents Ever Exposed OR < 1 1 < OR < 2 OR > 2 p-value (p) Agents % Agents % Agents % p < 0.5% 1 0.0 128 2.2 24 0.4 0.5% < p < 1% 56 1.0 14 0.2 1% < p < 2.5% 4 0.1 128 2.2 32 0.6 2.5% < p < 5% 18 0.3 171 3.0 93 1.6 5% < p < 10% 63 1.1 241 4.2 63 1.1 10% < p < 20% 121 2.1 470 8.2 63 1.1 p > 20% 1,308 23.0 2,676 47.0 25 0.4 Total 1,515 26.6 3,870 67.9 314 5.5 Table 7.2: Agents Significant  After  Adjusting for  Multiplicity Using the Hochberg and Benjamini Procedure from  5,699 Ever Exposure Agent Name CAS Cases OR 95% CI 2,5-PYRROLIDINEDIONE, 12AE MPIB D a 67762-72-5 361 1.39 1.21 - 1.60 NATURAL GAS, LIQUIFIED 25 3.11 1.92 - 5.04 PHOSPHORODITHIOIC ACID, MOOB E ZS° 68784-31-6 3 35 1.38 1.20 - 1.60 1, 2-ETHANEDIAMINE, RP W C IB HP° 68891-84-9 2 5 2.89 1.79 - 4.67 ALKENES, C15-18 ALPHA-, RPW SDP CS S° 72275-86-6 301 1.38 1.19 - 1.60 ETHANOL, 2-(2-(2-BE)E)-a 143-22-6 176 1.48 1.23 - 1.77 PHENOL, DODECYL-, SULFURIZED, CCSOa 68784-26-9 390 1.34 1.16 - 1.53 ° See appendix table B.l for  agent name abbreviations Table 7.3: Distribution of  p-values Across 3,450 Agents Tested For Dose-Response Low Exposure Medium Exposure OR < 1 OR > 1 OR < 1 OR > 1 p-value (p) Agents % Agents % Agents % Agents % p < 0.5% 67 1.9 59 1.7 0.5% < p < 1% 47 1.4 42 1.2 1% < p < 2.5% 1 0.0 126 3.7 2 0.1 69 2.0 2.5% < p < 5% 2 0.1 137 4.0 8 0.2 120 3.5 5% < p < 10% 12 0.3 223 6.5 24 0.7 193 5.6 10% < p < 20% 73 2.1 353 10.2 83 2.4 287 8.3 p > 20% 1,108 32.1 1,301 37.7 1,005 29.1 1,558 45.2 Total 1,196 34.7 2,140 62.0 1,122 32.5 2,227 64.6 High Exposure Ordinal Trend Test OR < 1 OR > 1 OR < 1 OR > 1 p-value (p) Agents % Agents % Agents % Agents % p < 0.5% 36 1.0 86 2.5 0.5% < p < 1% 25 0.7 1 0.0 . 38 1.1 1% < p < 2.5% 3 0.1 68 2.0 5 0.1 106 3.1 2.5% < p < 5% 13 0.4 86 2.5 7 0.2 120 3.5 5% < p < 10% 42 1.2 177 5.1 29 0.8 226 6.6 10% < p < 20% 81 2.3 273 7.9 50 1.4 401 11.6 p > 20% 1,014 29.4 1,632 47.3 760 22.0 1621 47.0 Total 1,153 33.4 2,236 64.8 852 24.7 2,598 75.3 Table 7.4: Number of  Agents With Significant  Ever Exposure and Ordinal Trend Results Ordinal Trend Test Ever Exposure OR < 1 OR > 1 Total OR p-value p < 1% 1% < p < 5% p > 5% p < 1% 1% < p < 5% p > 5% < 1 <1% 1 0 0 0 0 0 1 1% - 5% 0 5 1 0 0 0 6 > 5% 0 7 614 0 0 119 740 > 1 <1% 0 0 0 107 74 20 201 1% - 5% 0 0 0 15 111 186 312 > 5% 0 0 224 2 41 1,923 2,190 Total 1 12 839 124 226 2,248 3,450 Table 7.5: Selected 30 Agents with Significant  Associations Dose-Response Ever Exposure Low Medium High Trend NIOSH Agent Name Cases OR 95% CI Pb OR pfc OR p6 OR Test X9078 1 - Propene, 2 - Methyl - , Sulfurized 397 1.27 1.11-1.46 0.34 1.11 0.01 1.30 0.00 1.39 0.0001 X2689 1, 2-Ethanediamine, RP W C IB HP a 25 2.89 1.79-4.67 0.16 1.93 0.01 2.95 0.00 4.00 <.0001 X2305 2,5-Pyrrolidinedione, 12AE MPIB D RP a 206 1.38 1.17-1.64 0.17 1.22 0.06 1.32 0.00 1.62 <.0001 X2303 2,5-Pyrrolidinedione, 12AE MPIB D a 361 1.39 1.21-1.60 0.00 1.42 0.00 1.42 0.01 1.33 <.0001 X1401 2-Butenedioic Acid (E)-, PW 1,3-B EBQ 35 2.18 1.47-3.22 0.20 1.62 0.00 2.68 0.02 2.25 0.0001 XI894 2-Propenoic Acid, 2M CEPWC2° 48 1.86 1.34-2.60 0.12 1.63 0.06 1.73 0.00 2.20 0.0002 X2307 Alkenes, C15-18 Alpha-, RPW SDP CS S° 301 1.38 1.19-1.60 0.00 1.40 0.00 1.39 0.01 1.36 0.0001 90320 Asphalt 499 1.29 1.13-1.47 0.00 1.40 0.24 1.13 0.00 1.34 0.0018 90590 Clay, NEC 375 1.29 1.13-1.48 0.01 1.33 0.03 1.27 0.02 1.28 0.0020 M1150 Cyclohexylamine, N - Ethyl - 28 2.29 1.48-3.54 0.02 2.31 0.00 3.59 0.95 1.03 0.0035 M0984 Ethanol, 2-(2-(2-BE)E)-" 176 1.48 1.23-1.77 0.00 1.57 0.06 1.34 0.01 1.52 0.0002 X4267 Ether, Tert - Buty Methyl 32 1.96 1.31-2.93 0.65 1.19 0.01 2.40 0.01 2.56 0.0003 36060 Heptane 457 1.30 1.14-1.49 0.00 1.35 0.05 1.22 0.00 1.33 0.0008 36955 Hexane 477 1.30 1.14-1.48 0.00 1.44 0.02 1.25 0.06 1.21 0.0071 P0620 Impact Noise 545 1.30 1.14-1.48 0.08 1.18 0.00 1.36 0.00 1.34 0.0001 Y1006 Natural Gas, Liquified 25 3.11 1.92-5.04 0.00 2.70 0.28 2.46 0.00 4.12 <.0001 T1909 Nonylphenol Ethylene OA" 80 1.63 1.26-2.11 0.52 1.17 0.01 1.76 0.00 2.01 <.0001 83048 Nonylphenoxyethanol 27 2.49 1.59-3.90 0.08 2.08 0.02 2.61 0.01 2.81 0.0001 S2599 OFW Steel 221 1.37 1.16-1.61 0.00 1.45 0.02 1.37 0.08 1.28 0.0022 92500 Oil, Hydraulic 51 1.74 1.26-2.39 0.56 0.80 0.00 2.30 0.00 2.15 <.0001 X2298 Phenol, Dodecyl-, Sulfurized,  CCSOa 390 1.34 1.16-1.53 0.01 1.33 0.00 1.41 0.03 1.27 0.0004 X2306 Phosphorodithioic Acid, MOOB E ZSa 335 1.38 1.20-1.60 0.00 1.40 0.00 1.41 0.01 1.34 0.0001 X2295 Phosphorodithioic Acid, OOB(2E)E ZSQ 450 1.30 1.14-1.49 0.03 1.25 0.00 1.35 0.01 1.31 0.0003 X1075 Phosphorodithioic Acid, OOZSa 161 1.42 1.17-1.71 0.23 1.22 0.00 1.59 0.02 1.44 0.0003 60713 POC - Gasoline (leaded)" 617 1.26 1.10-1.44 0.15 1.15 0.01 1.27 0.00 1.36 0.0002 X5263 POC - Jet Fuel & Gasoline, ULD" 557 1.28 1.12-1.46 0.03 1.24 0.02 1.24 0.00 1.37 0.0003 73075 SN, Tin - MF Unknown 420 1.30 1.13-1.48 0.06 1.22 0.01 1.30 0.00 1.38 0.0002 T1475 Solvent RD HVY PF DIST (Petroleum)" 535 1.24 1.08-1.41 0.09 1.18 0.38 1.09 0.00 1.45 0.0002 X2293 Sulfonic  Acids, Petroleum, CSO° 375 1.30 1.13-1.49 0.03 1.26 0.01 1.34 0.02 1.30 0.0006 X2308 Sulfonic  Acids, Petroleum, MS" 208 1.40 1.18-1.66 0.12 1.25 0.04 1.34 0.00 1.61 <.0001 " See appendix table B.l for  agent name abbreviations b p-value Figure 7.1: Histograms of  Positive Cumulative Exposures for  Top 30 Agents 36060 36955 60713 U) o c o £ S ra CL o "1 1 T" 0 10 30 "I 1 50 co Q. o o o co i r 0 5 "i—i—i—i r 15 25 I 35 CD Q. O o o •<t _ o - Cb_ 10 20 30 40 Cumulative Exposure Cumulative Exposure Cumulative Exposure 73075 83048 90320 U>  O —I I 8 -I ro Q. o -J i—i—i—i—r~ 0 5 10 20 -1 30 (0 Q_ o o o to 0.00 0.10 0.20 ra CL O o J o CO - 10 — r~ 20 Cumulative Exposure Cumulative Exposure Cumulative Exposure 90590 92500 M0984 ra 0. O -i o _ o ^ co —I—I—I—I 0 5 10 20 Cumulative Exposure ra Q. o o o <o r i—i—i—i i i i 0 2 4 6 8 12 Cumulative Exposure to CL o -l o _ o (O - "T" 10 ~I 15 Cumulative Exposure M1150 P0620 S2599 01 Q. O o o CD —i—r "I—I—I—I 0.0 1.0 2.0 3.0 Cumulative Exposure o o o co ra CL "i—r~ 15 n—i 25 0 5 Cumulative Exposure 01 CL o o o co T n — r 2 3 0 1 Cumulative Exposure T1475 T1909 X1075 uj o = § .CD S 16 CL —i—i—i—r~ 0 5 10 20 1 1 30 Cumulative Exposure m c .2 ra CL o o o co -I 1—I—I—I—I—I 0 1 2 3 4 5 6 Cumulative Exposure o o o co co CL T T 1 0.0 0.5 1.0 1.5 Cumulative Exposure X1401 X1894 X2293 ra Q. o o o (D —I— 0.10 T" ~I 0.20 0.00 Cumulative Exposure o o o to —r~ 10 15 Cumulative Exposure I I I I I I I I 0 5 15 25 35 Cumulative Exposure X2295 X2298 X2303 o o o <0 i—r 0 5 1—i—i—r 15 25 35 o o o tO -J CO CL l i i r 0 5 15 ~1 25 35 ra Q. i r 5 15 1 I 25 —I 35 Cumulative Exposure Cumulative Exposure Cumulative Exposure X2305 X2306 X2307 ra CL o _ o to - 10 ° 1 o _ .s (O -ra CL ~i—i—r 15 ~1 25 I 35 o _ o to _ i r 0 5 i—i—i—i—r 15 25 35 Cumulative Exposure Cumulative Exposure Cumulative Exposure X2308 X2689 X4267 W O —I ~l 1 1 6 8 10 Cumulative Exposure ° n o _ o to - r T T" n 0.0 0.2 0.4 0.6 Cumulative Exposure ra CL o _ o to - i—r o.oo -I—i—i—i—i—i 0.10 0.20 0.30 Cumulative Exposure X5263 X9078 Y1006 o o o (D -J " I 1 1 1 0 10 20 30 40 Cumulative Exposure o o (O J <0 CL i i i n 0 5 10 20 1 1 30 Cumulative Exposure ra CL ° -i o _ o to - —I—I—I—I—I 0.00 0.10 0.20 0.30 Cumulative Exposure Figure 7.2: Histograms of  Transformed  Posit ive Cumulative Exposures for  Top 30 Agents 36060 36955 60713 o -o -•o- - 1 L _ <2 o c o - .2 ® 15 Q_ I 1 1 1 1 1 1 -8 -4 0 2 4 . Transformed Cumulative Exposure 73075 i—i—r J=bi_ ~i—i—i—i 0 2 4 Transformed Cumulative Exposure 83048 o -o -oo -ra Q_ - d l £ L ^ i—i—i—i—i—i—i -8 -4 0 2 4 Transformed Cumulative Exposure 90320 o o CO J J i i r i i i 0 2 4 w c 15 CL Transformed Cumulative Exposure 3 — —i —i to c ® o n o -CD _ -r — Tn ra CL o - 1 r -5 -4 - 3 -2 Transformed Cumulative Exposure LL_ I—I—I—I—I—I—I -8 - 4 0 2 4 Transformed Cumulative Exposure 90590 92500 M0984 o -o -- i i r -8 -4 CO Q. I I I 0 2 4 Transformed Cumulative Exposure n r LL "i—r i -6 - 4 - 2 0 2 Transformed Cumulative Exposure o in CO CL QL I—I—I—I—I—I—I - 1 0 - 6 - 2 2 Transformed Cumulative Exposure M1150 P0620 S2599 o Tt o CM m=, c ffl 15 CL -6 - 4 - 2 0 2 Transformed Cumulative Exposure o o co i—i—r -8 -4 1—I 0 2 4 c o 15 o. Transformed Cumulative Exposure o o H CN r JZl 0 1 -8 -6 - 4 - 2 0 2 Transformed Cumulative Exposure T1475 T1909 X1075 o ° H i r -8 -4 On. c a> 15 CL I I I 0 2 4 Transformed Cumulative Exposure o •<t i r -4 1 1 0 2 c .9! 15 CL Transformed Cumulative Exposure o -m - n I I I I I - 1 0 - 6 - 2 0 Transformed Cumulative Exposure X1401 in CO in o -1 r d i i r T L i -7 -5 -3 Transformed Cumulative Exposure X1894 ro CL i I r -4 Transformed Cumulative Exposure X2293 « o n c a>  -co _ 15 0. o - i—r 10 n—r - 6 i . T—I—I—I -2 2 4 Transformed Cumulative Exposure X2295 X2298 X2303 o o CO i i r - 1 0 - 6 "L 0) 15 CL - 2 n l 2 4 Transformed Cumulative Exposure o o i r t h _ -4 "I—I—I 0 2 4 CO c o o o CO CO CL o Transformed Cumulative Exposure - r i - r f f "i—i—r - 1 0 - 6 Oh. 1—I 2 4 Transformed Cumulative Exposure X2305 X2306 X2307 -r-mT th^ as CL "1—i—I—I—I—I - 1 0 - 6 - 2 2 Transformed Cumulative Exposure o ° H co _ i i i r - 1 0 - 6 IK 1—i 2 4 w o c o 0) CO 15 CL o Transformed Cumulative Exposure -m-H" i i i i r - 1 0 - 6 - 2 O x \ I 2 4 Transformed Cumulative Exposure X2308 o -o - CM - - H T T T I—I—I—T i—i 2 .Si in I 15 Q. - 1 0 - 6 - 2 Transformed Cumulative Exposure X2689 i i r -7 -5 1—i— -3 o _ m CD c _ 2 o _ CO 15 0. o - Transformed Cumulative Exposure X4267 r T i , "1 -8 -6 -4 -2 Transformed Cumulative Exposure X5263 X9078 Y1006 o ° H CO 1—i—r -4 n 10 CL 0 2 4 Transformed Cumulative Exposure o o CO -J -rl~l I—I—r - 1 0 - 6 -2 2 4 Transformed Cumulative Exposure o _ o - T H ~ L T -5 -4 -3 -2 -1 Transformed Cumulative Exposure Table 7.6: Results for  Linear Fit of  Transformed"  Cumulative Exposure for  Top 30 Agents Transformed 0 Cumulative Exposure NIOSH Agent Name Cases p-value OR 95% CI 36060 HEPTANE 457 0.0001 1.003 1.001 - 1.004 36955 HEXANE 477 0.0001 1.003 1.001 - 1.004 60713 POC - GASOLINE (LEADED)6 617 0.0007 1.002 1.001 - 1.004 73075 SN, TIN - MF UNKNOWN 420 0.0002 1.003 1.001 - 1.004 83048 NONYLPHENOXYETHANOL 27 <.0001 1.010 1.005 - 1.014 90320 ASPHALT 499 0.0002 1.003 1.001 - 1.004 90590 CLAY, NEC 375 0.0003 1.003 1.001 - 1.004 92500 OIL, HYDRAULIC 51 0.0006 1.006 1.002 - 1.009 M0984 ETHANOL, 2-(2-(2-BE)E)-6 176 <.0001 1.004 1.002 - 1.006 M1150 CYCLOHEXYLAMINE, N - ETHYL - 28 0.0002 1.009 1.004 - 1.013 P0620 IMPACT NOISE 545 0.0001 1.003 1.001 - 1.004 S2599 OFW STEEL 221 0.0002 1.003 1.002 - 1.005 T1475 SOLVENT RD HVY PF DIST (PETROLEUM)6 535 0.0014 1.002 1.001 - 1.004 T1909 NONYLPHENOL ETHYLENE OA6 80 0.0002 1.005 1.002 - 1.008 X1075 PHOSPHORODITHIOIC ACID, OOZS6 161 0.0002 1.004 1.002 - 1.006 X1401 2-BUTENEDIOIC ACID (E)-, PW 1,3-B EB6' 35 <.0001 1.008 1.004 - 1.012 X1894 2-PROPENOIC ACID, 2M CEPWC26 48 0.0002 1.006 1.003 - 1.010 X2293 SULFONIC ACIDS, PETROLEUM, CSO6 375 0.0002 1.003 1.001 - 1.004 X2295 PHOSPHORODITHIOIC ACID, OOB(2E)E ZS6 450 0.0001 1.003 1.001 - 1.004 X2298 PHENOL, DODECYL-, SULFURIZED, CCSO6 390 <.0001 1.003 1.002 - 1.004 X2303 2,5-PYRROLIDINEDIONE, 12AE MPIB D6 361 <.0001 1.003 1.002 - 1.005 X2305 2,5-PYRROLIDINEDIONE, 12AE MPIB D RPh 206 0.0001 1.003 1.002 - 1.005 X2306 PHOSPHORODITHIOIC ACID, MOOB E ZS6 335 <.0001 1.003 1.002 - 1.005 X2307 ALKENES, C15-18 ALPHA-, RPW SDP CS Sb 301 <.0001 1.003 1.002 - 1.005 X2308 SULFONIC ACIDS, PETROLEUM, MS6 208 <.0001 1.004 1.002 - 1.005 X2689 1, 2-ETHANEDIAMINE, RP W C IB HP6 25 <.0001 1.011 1.006 - 1.016 X4267 ETHER, TERT - BUTYL METHYL 32 0.0010 1.007 1.003 - 1.011 X5263 POC - JET FUEL & GASOLINE, ULD6 557 0.0002 1.002 1.001 - 1.004 X9078 1 - PRQPENE, 2 - METHYL - , SULFURIZED 397 0.0006 1.002 1.001 - 1.004 Y1006 NATURAL GAS, LIQUIFIED 25 <.0001 1.012 1.007- 1.017 a Box-Cox transformation  with A = 0.01 6 See appendix table B.l for  agent name abbreviations Table 7.7: Component Scores" for  PCA of  the 30 Selected Agents Component Agent Name 1 2 3 4 5 6 7 8 9 10 ALKENES, C15-18 ALPHA-, RPW SDP CS S6 99 -2 -3 -5 -3 1 1 -1 -2 0 2,5-PYRROLIDINEDIONE, 12AE MPIB D6 . 98 0 -3 10 -2 1 0 0 -2 0 PHOSPHORODITHIOIC ACID, MOOB E ZS6 98 -1 -3 10 -3 1 0 -1 -2 0 PHENOL, DODECYL-, SULFURIZED, CCSO6 98 -1 -2 -5 1 2 1 8 -2 0 SULFONIC ACIDS, PETROLEUM, CSO6 97 -1 1 9 -4 1 -1 0 -4 1 PHOSPHORODITHIOIC ACID, OOB(2E)E ZS6 96 -2 3 10 -4 1 -1 0 -2 0 1 - PROPENE, 2 - METHYL - , SULFURIZED 76 0 56 -6 0 -2 3 -4 -2 9 POC - GASOLINE (LEADED)6 67 17 15 -3 20 -3 13 -4 30 -11 POC - JET FUEL & GASOLINE, ULD6 65 18 54 -6 11 -4 5 -5 15 -4 NATURAL GAS, LIQUIFIED 1 96 -1 2 4 -10 0 2 2 0 NONYLPHENOXYETHANOL 2 94 0 2 0 22 1 1 0 1 2-BUTENEDIOIC ACID (E)-, PW 1,3-B EB6 2 94 0 2 1 22 1 1 0 1 1, 2-ETHANEDIAMINE, RP W C IB HP6 1 92 -1 2 3 -12 0 1 3 0 ETHER, TERT - BUTYL METHYL 2 74 2 1 -3 55 2 0 -2 2 ASPHALT 12 1 90 2 -2 -5 4 -3 2 0 HEXANE -2 -1 82 2 1 2 6 -8 17 2 SN, TIN - MF UNKNOWN -2 -3 61 3 16 6 1 19 -10 -2 2,5-PYRROLIDINEDIONE, 12AE MPIB D RP 6 7 4 2 99 1 0 0 2 1 0 SULFONIC ACIDS, PETROLEUM, MS6 7 4 2 99 1 0 0 2 1 0 OFW STEEL 0 2 4 -1 75 4 -9 -4 -4 -5 HEPTANE 0 -4 8 0 70 13 11 0 5 12 PHOSPHORODITHIOIC ACID, OOZS6 4 43 9 8 52 -10 1 42 0 -4 CYCLOHEXYLAMINE, N - ETHYL - 1 22 5 1 -6 85 4 0 -4 3 ETHANOL, 2-(2-(2-BUTOXYETHOXY)ETHOXY)- 0 1 -2 0 15 46 -4 1 5 -4 CLAY, NEC -3 1 11 1 -2 -3 95 -1 1 -2 SOLVENT RD HVY PF DIST (PETROLEUM)6 65 0 -1 -3 12 3 65 9 -1 10 NONYLPHENOL ETHYLENE OA6 1 2 4 2 0 2 2 91 1 -4 IMPACT NOISE 1 0 15 2 11 0 7 -17 77 -20 OIL, HYDRAULIC 0 3 -5 -1 -13 5 -7 23 61 27 2-PROPENOIC ACID, 2M CEPWC26 0 2 1 1 7 -4 2 -5 0 92 Percentage of  Variance 26.30 15.40 8.49 6.75 4.87 4.29 4.09 3.69 3.42 3.35 Cumulative Percentage 26.30 41.70 50.19 56.94 61.82 66.11 70.20 73.89 77.31 80.66 a Component scores are multiplied by 100 and rounded to the nearest integer, component scores greater than 0.4 are highlighted 6 See appendix table B.l for  agent name abbreviations Table 7.8: Dose-Response Results for  Component Groups Dose-Response Low Medium High Ordinal Component Cases P-value OR P-value OR P-value OR P-value OR 1 666 0.65 1.05 <.01 1.32 <.01 1.39 <.0001 1.13 2 40 0.51 0.73 0.03 1.87 <.01 2.70 <.0001 1.36 3 637 <.01 1.37 <.01 1.32 <.01 1.40 <.0001 1.12 4 208 0.07 1.30 0.05 1.32 <.01 1.59 <.0001 1.17 5 501 0.03 1.24 <.01 1.38 <.01 1.31 0.0002 1.12 6 177 <.01 1.57 0.05 1.35 <.01 1.49 0.0002 1.17 7 585 <.01 1.34 0.03 1.23 <.01 1.37 0.0004 1.11 8 80 0.52 1.17 <.01 1.76 <.01 2.01 <.0001 1.28 9 546 0.19 1.14 <.01 1.40 <.01 1.35 <.0001 1.12 10 48 0.12 1.63 0.06 1.73 <.01 2.20 0.0002 1.32 Table 7.9: Multivariate Ordinal Results" for  Component Groups Ordinal Component Cases OR 95% CI 2 40 1.25 1.07- 1.47 4 208 1.11 1.02-1.20 9 546 1.10 1.04-1.17 a p-value < 5% Table 7.10: Results for  Ever Exposure to Any of  the Members of  each Component Group Any Exposure Component Cases p-value OR 95% CI 1 666 0.0014 1.25 1.09 - 1.44 2 40 0.0017 1.78 1.24 - 2.55 3 637 <.0001 1.36 1.19 - 1.56 4 208 <.0001 1.40 1.18 - 1.66 5 501 <.0001 1.31 1.15 - 1.50 6 177 <.0001 1.47 1.23 - 1.76 7 585 <.0001 1.31 1.15 - 1.50 8 80 0.0002 1.63 1.26 - 2.11 9 546 0.0001 1.30 1.14 - 1.48 10 48 0.0002 1.86 1.34 - 2.60 Table 7.11: Multivariate Any Results" for  Component Groups Any Component Cases OR 95% CI 3 637 1.26 1.09 - 1.45 6 177 1.25 1.03 - 1.52 10 48 1.54 1.10 - 2.17 " p-value < 5% Table 7.12: Results for  Ever Exposure to All of  the Members of  each Component Group All Exposed Component Cases p-value OR 95% CI 1 223 <.0001 1.42 1.21 - 1.67 2 25 <.0001 3.11 1.92 - 5.04 3 295 0.0113 1.21 1.04 - 1.40 4 206 0.0002 1.38 1.17 - 1.64 5 89 0.0020 1.47 1.15 - 1.87 6 27 <.0001 2.49 1.59 - 3.90 7 325 0.0052 1.23 1.06 - 1.41 8 80 0.0002 1.63 1.26 - 2.11 9 50 0.0008 1.74 1.26 - 2.41 10 48 0.0002 1.86 1.34 - 2.60 Table 7.13: Multivariate All Results" for  Component Groups All Exposed Component Cases OR 95% CI 1 223 1.32 1.12 - 1.57 2 25 2.47 1.49 - 4.08 " p-value < 5% Figure 7.3: Proportion of  Cumulative Exposure Due to Employment in Each US Job for  the Top 30 Agents • Gasoline service stations - related occupations Q Motor vehicles and equipment - Supervisors production IZ Gasoline service stations - Automobile mechanics H Hospitals - Power plant operators CD Construction - Insulation workers a Automotive repair shops - Automobile mechanics 13 Ship and boat building - Carpenters Q Pulp paper & paperboard mills - Misc. electronic equipment repairers • Automotive repair shops - Bus truck & stationary engine mechanics • Construction - Plumbers pipefitters and steamfitters • Other • Timber cutting and logging occupations • Ship & boat building - Lathe & turning machine set-up op. H Construction - Electrical power installers & repairers Ea Trucking service - Truck drivers heavy E3 Automotive repair shops - Automobile body repairers • Construction - Plumber pipefitter & steamfitter apprentices CI Apparel & accessories - Knitting looping taping & weaving mach. op. ffl Meat products - Bus truck & stationary engine mechanics • Water transportation - Marine engineers Q Sawmills planing mills and millwork - Stationary engineers Table 7.14: Properties of  the Selected 30 Agents Agent Name CAS IARC" Cases PC b Most Common US Job %c JEM%d ALKENES, C15-18 ALPHA-, RPW SDP CS Se 72275-86-6 301 1 Timber cutting & logging 85 52 2,5-PYRROLIDINEDIONE, 12AE MPIB D e 67762-72-5 361 1 Timber cutting & logging 79 52 PHOSPHORODITHIOIC ACID, MOOB E ZSe 68784-31-6 335 1 Timber cutting & logging 80 52 PHENOL, DODECYL-, SULFURIZED, CCSOe 68784-26-9 390 1 Timber cutting & logging 73 52 SULFONIC ACIDS, PETROLEUM, CSOe 68783-96-0 375 1 Timber cutting & logging 67 54 PHOSPHORODITHIOIC ACID, OOB(2E)E ZSe 4259-15-8 450 1 Timber cutting & logging 60 54 1 - PROPENE, 2 - METHYL - , SULFURIZED 68511-50-2 397 1 Timber cutting & logging 44 54 POC - GASOLINE (LEADED)6 2B 617 1 Timber cutting & logging 23 64 POC - JET FUEL & GASOLINE, ULDe 557 1 Timber cutting & logging 28 56 NATURAL GAS, LIQUIFIED 25 2 Gasoline service station related 100 1 NONYLPHENOXYETHANOL 27986-36-3 27 2 Gasoline service stations related 74 1 2-BUTENEDIOIC ACID (E)-, PW 1,3-B EBe 24938-12-3 35 2 Gasoline service station related 66 1 1, 2-ETHANEDIAMINE, RP W C IB HPe 68891-84-9 25 2 Gasoline service station related 94 3 ETHER, TERT - BUTYL METHYL 1634-04-4 3 32 2 Gasoline service station related 48 1 ASPHALT 8052-42-4 3 499 3 Con.-Plumber pipe & steam fitter  ap. 27 100 HEXANE 110-54-3 477 3 Con.-Plumber pipe & steam fitter  ap. 29 100 SN, TIN - MF UNKNOWN 7440-31-5 420 3 Con.-Plumbers pipe & steam fitter 16 40 2,5-PYRROLIDINEDIONE, 12AE MPIB D RP e 72269-41-1 206 4 Ship & boat building - Carpenter 26 18 SULFONIC ACIDS, PETROLEUM, MSe 61789-87-5 208 4 Ship & boat building - Carpenter 25 18 OFW STEEL 221 5 Automotive repair-body repairers 33 8 HEPTANE 142-82-5 457 5 Automotive repair-Bus truck & sta- tionary engine mechanic 20 95 PHOSPHORODITHIOIC ACID, OOZSe 26566-95-0 161 5 Trucking service - Truck drivers heavy 41 2 CYCLOHEXYLAMINE, N - ETHYL - 5459-93-8 3 28 6 Gasoline service stations-mechanic 59 23 ETHANOL, 2-(2-(2-BE)E)-e 143-22-6 176 6 Con.-Electrical power installer 53 100 CLAY, NEC 375 7 Construction-Insulation worker 39 32 SOLVENT RD HVY PF DIST (PETROLEUM)e 64741-88-4 1/3 535 7 Timber cutting & logging 37 52 NONYLPHENOL ETHYLENE OAe 26027-38-3 80 8 Ship & boat building - Lathe & turn- ing machine set-up op. 64 100 IMPACT NOISE 545 9 Construction - Carpenters 13 17 OIL, HYDRAULIC 1/3 51 9 Motor vehicles & equipment - Super- visor production 78 100 2-PROPENOIC ACID, 2M CEPWC26 66057-34-9 48 10 Hospitals - Power plant operator 51 100 ° IARC classification  where 1 is definitely  carcinogenic, 2B is possibly carcinogenic, and 3 is not classifiable. b Principal component number c Percentage of  total cumulative exposure of  9119 patients due to employment in the US Job in column 6. d Proportion of  NIOSH NOES study participants employed in the US job in column 6 exposed to the agent. e See appendix table B.l for  agent name abbreviations Dose-Response Chemical Name IARCQ CAS NIOSH Name Cases Ever Low Medium High Ordinal 4-Aminobiphenyl (xenylamine) 1 92-67-1 N,N-Bis(2-CE)-2-NL6 1 494-03-1 2-Naphthylamine 1 91-59-8 Naphthylamine, beta- + Benzidine 1 92-87-5 Benzidine + Cyclophosphamide 1 50-18-0 Oxazaphosphorine, 2-(bis6 + Magenta 1 632-99-5 CI Basic Violet 14, MHC6 12 0.69 0.54 0.49 1.07 0.43 Arsenic 1 7440-38-2 Arsenic 23 0.80 0.78 0.76 0.86 0.40 Coal tar pitches 1 65996-93-2 Pitch, Coal tar 89 1.04 1.10 1.03 1.01 0.84 Mineral oils6 1 8002-05-9 Petroleum 152 1.02 1.22 0.98 0.89 0.70 4,4'-Methylenebis(2-chloroaniline) 2A 101-14-4 Aniline, 4,4'-methylenebis6 + Adriamycin 2A 23214-92-8 Adriamycin + para-Chloro-ortho-Toluidine 2A 95-69-2 Toludine, 4-chloro-, ortho- + Phenacetin 2A 62-44-2 62-44-2 Acetophenetidide, para - Phenacetin, powder 4 + 0.78 Benzidine based dyes 2A 1937-37-7 2602-46-2 16071-86-6 C.I. Direct black 38, DS6 C.I. Direct blue 6, TS6 C.I. Direct brown 95, DS6 12 + + 1.36 1.55 1.14 1.46 0.39 Benz(a)anthracene 2A 56-55-3 Benz (a) anthracene 13 1.92* 2.65* 1.38 1.69 0.12 Trichloroethylene 2A 79-01-6 Ethylene, trichloro - 345 1.21** 1.10 1.33** 1.20 <.01 Tetrachloroethylene 2A 127-18-4 Tetrachloroethylene 470 1.25** 1.24* 1.32** 1.19 <.01 Diesel engine emissions 2A POC - Diesel fuels 605 1.18* 1.14 1.17 1.25* 0.01 Engine emissions 2A POC - Gasoline (leaded) 617 1.26** 1.15 1.27* 1.36** <.01 POC - Gasoline (lead CU)6 40 1.11 1.54 0.72 1.06 0.91 N-[4-(5-Nitro-2-Furyl)2TZ]A6 2B 531-82-8 N-Nitrosodi-n-butylamine 2B 924-16-3 Oil Orange SS 2B 2646-17-5 para-Cresidine 2B 120-71-8 2-(2-Formylhydrazino)-4(5N2F)T6 2B 3570-75-0 2-Nitroanisole 2B 91-23-6 3,3'-Dichlorobenzidine 2B 91-94-1 4-chloro-ortho-phenylenediamine 2B 95-83-0 Niridazole 2B 61-57-4 Citrus Red No. 2 2B 6358-53-8 Auramine 2B 492-80-8 Dose-Response Chemical Name IARC" CAS NIOSH Name Cases Ever Low Medium High Ordinal Chlordane 2B 12789-03-6 Chlordane + Disperse Blue 1 2B 2475-45-8 Anthraquinone, 1,4,5,8-ta-6 + Gasoline 2B 8006-61-9 Gasoline, natural + ortho-Aminoazotoluene 2B 97-56-3 C.I. Solvent yellow 3 + para-Dimethylaminobenzene 2B 60-11-7 C.I. Solvent yellow 2 + Ponceau 3R 2B 3564-09-8 Ponceau-3R + CI Basic Red 9 2B 569-61-9 C.I. Basic red 9, MHC6 4 2.44 3,3'-Dimethoxybenzidine 2B 119-90-4 Benzidene, 3, 3' dimethoxy - 8 1.69 1,3-Dichloropropene 2B 542-75-6 Propene, 1 ,3 - dichloro - 12 1.59 1.63 1.58 1.57 0.20 Chloroform  (in drinking water) 2B 67-66-3 Chloroform 27 0.99 0.69 1.26 1.06 0.81 Nitrilotriacetic Acid 2B 139-13-9 Nitrilotriacetic Acid 31 1.14 0.71 0.37 0.70 0.55 Sodium ortho-phenylphenate 2B 132-27-4 Biphenol, sodium salt, 2 - 114 1.02 0.96 1.18 0.92 0.88 Lead 2B 7439-92-1 7439-92-1 7439-92-1 7439-92-1 PB, lead - MF Unk PB, lead powder - MF Unk PB, lead - pure PB, lead fume  - MF Unk 396 14 + + 1.25** 1.18 1.30* 1.28* <.01 Carbon black 2B 1333-86-4 1333-86-4 Carbon black Carbon lampblack, powder 554 + 1.08 1.10 1.03 1.10 0.32 a IARC classification  where 1 is definitely  carcinogenic, 2A is probably carcinogenic, and 2B is possibly carcinogenic. 6 2-(2-Formylhydrazino)-4(5N2F)T = 2-(2-Formylhydrazino)-4-(5-Nitro-2-Furyl) Thiazole, N-[4-(5-Nitro-2-Furyl)2TZ]A = N-[4-(5-Nitro-2-Furyl)-2- Thiazolyl] Acetamide. Mineral oils = Mineral oils, untreated or mildly treated. Oxazaphosphorine, 2-(bis = Oxazaphosphorine, 2-(bis(2-chloroethyl) amino)tetrahydro-, 2-oxide, 2H-1, 3, 2-. MHC = monohydrochloride, DS = disodium salt, TS = tetrasodium salt, CU = content unknown, ta = tetraamino. N,N-bis(2-CE)-2-NL = N,N-bis(2-Chloroethyl)-2-naphthylamine (Chlornaphazine). + Less than 3 cases exposed * Significant  at a 5% alpha level ** Significant  at a 1% alpha level Siemiatycki Chemical Any Sub" CAS NIOSH Name Cases Ever Low Medium High Ordinal Natural gas comb, products 1.6* 3.8** Carbon tetrachloride 1.6 2.5** 56-23-5 Carbon tetrachloride 229 1.11 1.20 1.12 1.01 0.46 Diesel engine emissions 1.4 2.3** POC - Diesel fuels 605 1.18f 1.14 1.17 1.25f 0.01 Laboratory products 1.5 5.5* Cadmium compounds 1.6 4.9* Fabric dust 1 3.7* Fabric dust-synthetic + Photographic products 2.5 2.9* Chlorine 1 2.7* 7782-50-5 Chlorine 136 0.98 1.04 0.93 0.98 0.80 Polyester fibers 1.4 2.5* 80595-68-2 Polyester fibers  (MF Unk.) 12 0.98 0.90 1.00 1.05 0.98 Asphalt (bitumen) 0.9 2.2* 8052-42-4 Asphalt 499 1.29ft 1.40 ft 1.13 1.34ft <.01 Ammonia 1.2 2.1* 7664-41-7 Ammonia 336 1.06 1.15 1.09 0.94 0.90 Formaldehyde 1.2 1.7* 50-00-0 Formaldehyde 489 1.15| 1.18 1.07 1.18 0.08 Engine emissions 1.2* 1.3* POC - Gasoline (Leaded) 617 1.26ft 1.15 1.27f 1.36ft <.01 (leaded or unleaded) POC - Gasoline (Lead CU) 40 1.11 1.54 0.72 1.06 0.91 ionizing radiation 44** 0 Ionizing Radiation 18 0.57f 0.49 0.98 0.27f 0.02 Acrylic fibres 3.9** 3.3 Acrylic fibers  (MF Unk.) + Calcium carbonate 1.9** 1.6 471-34-1 Calcium carbonate, powder + 471-34-1 Carbonic acid, calcium salt 652 1.16f 1.19 1.11 1.18 0.09 471-34-1 Marble, dust 38 1.35 1.76f 1.34 0.97 0.34 1317-65-3 Limestone 470 1.22ft 1.29ft 1.12 1.24f 0.02 1317-65-3 Limestone, powder 290 1.21f 1.24 1.1 1.28f 0.02 Titanium dioxide ^ y** 4.5 13463-67-7 Titanium oxide (Tf02) 654 1.20ft 1.19 1.22f 1.20 0.02 Titanium compounds 7** 2.2 Hydrogen cyanide 3.4* 0 74-90-8 Hydrogen cyanide 13 1.23 1.14 1.21 1.34 0.48 Creosote 2.6* 2.6 Creosote 130 1.08 0.98 1.17 1.12 0.34 Polyethylene 2.5* 13 9002-88-4 Polyethylene wax 194 1.09 1.01 1.13 1.13 0.25 9002-88-4 Ethylene, polymers 470 1.11 1.11 1.06 1.14 0.15 9002-88-4 Polyethylene, fiber 11 1.67 3.08f 0.81 1.3 0.40 Carbon black 2.2* 1.8 1333-86-4 1333-86-4 Carbon black Carbon lampblack, powder 554 + 1.08 1.1 1.03 1.1 0.32 Clay dust 2.2* 1.8 Lead chromate 1.8* 2.2 7758-97-6 Lead chromate 98 1.12 0.99 1.15 1.22 0.23 7758-97-6 Chromic acid, lead(2+) salt 24 1.03 1.37 0.79 0.98 0.90 Aliphatic aldehydes 1.4* 1.6 Lead compounds 1.3* 1.1 a Odds-ratio for  substantial exposure. -I- Less than 3 cases * Significant  at p=0.10, one-sided, with at least 4 exposed exposed, f  Significant  at a 5% alpha level, ff  Significant  at a 1% alpha level ** Significant  at p=0.05, one-sided, with at least 5 exposed cases. Appendix A Table A.l: Odds-Ratios of  Ever Exposure and Dose-Response Results for  3,450 Agents" NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal 01038 107 0.88 0.83 0.86 0.93 0.35 11600 25 0.95 1.57 0.48 0.88 0.47 01568 407 1.05 1.13 1.12 0.91 0.92 11610 114 1.03 1.25 1.12 0.72 0.54 01600 205 1.08 1.18 1.16 0.89 0.86 11770 15 0.93 0.95 0.74 1.10 0.87 02740 280 1.24** 1.24 1.36** 1.13 0.02 - 12783 27 1.14 1.02 0.75 1.68 0.30 02820 559 1.25** 1.24* 1.33** 1.19 <.01 12845 214 1.16 1.10 1.13 1.24 0.06 02900 17 1.12 0.81 0.72 1.88 0.35 12960 324 1.03 1.16 1.13 0.81 0.50 03298 119 1.09 1.04 1.21 1.03 0.47 12963 165 0.97 0.83 1.05 1.03 0.92 03530 10 1.85 2.13 1.86 1.41 0.18 13100 61 1.22 1.27 0.90 1.49 0.14 03540 29 1.03 0.79 0.93 1.36 0.59 13410 61 1.12 1.25 0.94 1.17 0.53 03570 238 1.12 1.16 0.99 1.22 0.15 13480 459 1.15* 1.14 1.22* 1.09 0.11 03800 164 1.06 1.24 0.95 0.99 0.91 13850 435 1.18* 1.27* 1.13 1.15 0.08 04280 13 1.30 1.55 0.98 1.35 0.49 13980 514 1.21** 1.20 1.20 1.21* 0.01 04530 14 1.06 1.01 0.95 1.22 0.78 14380 441 1.13 1.12 1.06 1.21 0.06 04580 15 0.82 0.66 0.95 0.86 0.60 14400 210 1.12 0.94 1.25 1.18 0.09 04605 63 0.98 1.13 0.97 0.83 0.61 14410 181 1.01 0.91 1.13 0.97 0.88 04620 255 1.13 1.37** 0.95 1.09 0.42 14720 223 1.27** 1.10 1.29 1.44** <.01 04980 494 1.14 1.24* 1.05 1.13 0.20 14730 16 0.95 0.66 0.95 1.19 0.92 05250 336 1.06 1.15 1.09 0.94 0.90 15570 164 1.10 1.24 1.10 0.97 0.63 05270 308 1.05 1.05 1.10 1.01 0.62 15705 652 1.16* 1.19 1.11 1.18 0.09 06063 14 0.89 0.39 1.39 0.86 0.93 15720 317 1.14 1.25* 1.05 1.11 0.23 06145 481 1.11 1.13 1.06 1.14 0.17 15743 457 1.24** 1.27* 1.15 1.30** <.01 06163 105 1.04 1.03 1.07 1.04 0.72 15746 171 1.09 1.05 1.02 1.18 0.28 06175 238 1.13 1.24 1.13 1.03 0.34 15755 464 1.16* 1.10 1.21 1.18 0.03 06580 27 1.41 1.74 0.89 1.49 0.20 15765 378 1.17* 1.19 1.06 1.26* 0.03 07310 288 1.23** 1.13 1.28* 1.28* <.01 15800 103 1.04 1.09 1.04 0.98 0.90 07325 24 0.96 0.84 1.41 0.68 0.78 17366 476 1.21** 1.09 1.29** 1.25* <.01 07485 108 1.12 1.26 1.14 0.96 0.63 17367 502 1.21** 1.16 1.15 1.31** <.01 07545 23 0.80 0.78 0.76 0.86 0.40 17370 23 1.29 0.16 1.65 2.18* 0.03 08625 19 1.02 1.36 1.19 0.50 0.66 17385 29 0.95 0.60 1.12 1.11 0.88 08640 168 1.05 0.83 1.00 1.31* 0.17 17460 271 1.17* 1.20 1.01 1.29* 0.04 08650 71 1.07 1.03 1.09 1.09 0.59 17490 229 1.11 1.20 1.12 1.01 0.46 08655 234 1.10 0.99 1.22 1.08 0.23 17525 372 1.19* 1.18 1.28* 1.11 0.05 09070 285 1.16 1.39** 1.16 0.92 0.57 17683 437 1.15* 1.13 1.10 1.24* 0.03 10210 75 0.99 1.02 0.68 1.30 0.75 17695 248 1.17* 1.43** 0.96 1.12 0.26 11280 169 1.11 1.09 1.08 1.16 0.26 18040 136 0.98 1.04 0.93 0.98 0.80 11360 165 1.09 1.17 0.94 1.16 0.41 18045 29 1.07 0.85 1.40 0.97 0.70 11590 45 1.34 1.45 1.36 1.23 0.15 18190 55 1.42* 1.10 1.80* 1.50 0.01 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal 18260 119 1.28* 1.33 1.23 1.28 0.04 24615 418 1.15* 1.22* 1.22* 1.01 0.26 18500 27 0.99 0.69 1.26 1.06 0.81 24680 14 1.13 0.98 1.42 0.96 0.71 19130 15 0.84 0.40 0.84 1.20 0.90 24930 73 1.05 1.17 1.04 0.94 0.97 19360 88 0.99 1.01 0.83 1.13 0.95 25145 390 1.08 1.06 1.09 1.11 0.23 19380 32 0.93 0.96 1.28 0.53 047 25210 259 1.24** 1.33* 1.02 1.35* 0.01 19395 363 1.16* 1.27* 1.24* 0.97 0.30 25544 545 1.23** 1.15 1.17 1.37** <.01 19425 10 1.18 2.42* 0.00 1.22 0.94 25820 168 1.00 0.93 1.05 1.01 0.90 19430 71 1.05 1.22 0.98 0.93 0.97 26075 207 1.11 0.95 1.18 1.21 0.10 19540 58 0.95 1.18 0.90 0.78 0.43 26095 23 1.09 0.99 0.90 1.39 0.54 19680 387 1.12 1.13 1.22* 1.00 0.32 26130 101 1.01 1.08 0.77 1.17 0.85 19710 81 1.03 1.04 1.14 0.93 0.93 26335 67 0.99 1.07 0.83 1.03 0.88 19767 94 0.92 1.18 0.83 0.78 0.22 26560 237 1.17 1.06 1.20 1.26 0.03 19770 142 1.05 1.00 1.13 1.02 0.63 26615 590 1.19* 1.24* 1.13 1.20 0.04 19935 22 1.02 0.68 1.08 1.32 0.61 26880 10 1.39 1.23 1.86 0.97 0.43 19985 187 1.18 1.31* 1.07 1.17 0.14 26940 56 1.30 1.63* 1.08 1.19 0.23 20115 495 1.21** 1.25* 1.15 1.24* 0.01 27590 179 1.16 1.27 1.03 1.18 0.19 20155 16 1.10 1.73 1.02 0.60 0.78 27615 288 1.16* 1.09 1.18 1.22 0.03 20245 13 1.19 0.98 1.31 1.33 0.49 27760 22 1.26 2.09* 1.28 0.49 0.99 20265 560 1.23** 1.29** 1.14 1.25* 0.01 27780 130 1.07 1.08 1.17 0.98 0.64 20340 117 1.07 1.04 1.10 1.06 0.57 28510 36 1.12 1.32 1.21 0.83 0.87 20380 193 1.26** 1.29 1.13 1.36* 0.01 29010 206 1.18 1.13 1.18 1.23 0.05 20810 237 1.16 1.13 1.20 1.16 0.08 29325 29 0.93 1.24 1.03 0.49 0.34 20850 16 1.16 1.18 1.58 0.73 0.81 29930 565 1.16* 1.20 1.27** 1.03 0.21 20900 117 0.97 1.01 0.97 0.94 0.72 31350 248 1.01 1.06 0.98 0.97 0.87 21190 87 1.11 0.97 1.37 0.99 0.41 31470 418 1.25** 1.17 1.32** 1.26* <.01 21560 66 1.03 0.98 1.05 1.05 0.79 31490 186 1.15 1.26 1.12 1.08 0.27 21660 364 1.16* 1.26* 0.93 1.29* 0.05 31500 497 1.11 1.19 1.17 0.98 0.54 22734 38 1.11 1.32 1.02 0.98 0.80 31830 300 1.11 1.22 1.06 1.06 0.38 23180 10 0.67 0.30 1.94 0.38 0.35 31900 18 1.25 1.41 1.02 1.33 0.46 23275 34 1.56* 1.08 1.67 1.93* 0.01 31970 136 1.02 1.23 0.98 0.87 0.69 23360 9 1.57 1.77 2.32 0.52 0.46 32220 153 1.05 0.97 1.16 1.03 0.55 24003 207 1.17 0.97 1.39* 1.14 0.05 32385 619 1.23** 1.33** 1.17 1.19 0.05 24006 12 0.73 1.33 0.38 0.51 0.16 32500 50 0.97 1.53 0.76 0.66 0.29 24095 349 1.11 1.18 1.14 1.02 0.39 32550 115 0.87 0.81 0.81 0.97 0.35 24130 47 0.92 0.99 0.95 0.81 0.51 32590 342 1.23** 1.17 1.29* 1.23 <.01 24235 26 1.27 1.31 1.53 0.93 0.44 32925 52 0.95 0.99 1.17 0.77 0.58 24425 11 1.18 1.04 0.69 1.78 0.44 32940 61 0.96 0.89 1.26 0.82 0.72 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal 33115 325 1.14 1.35** 1.10 0.97 0.56 38530 13 1.23 1.14 1.21 1.34 0.48 33160 62 1.04 0.78 1.07 1.29 0.40 38550 126 1.24* 1.24 1.26 1.22 0.07 33165 25 0.91 0.92 1.06 0.76 0.62 38575 23 1.02 0.99 1.35 0.61 0.89 33230 39 0.82 0.60 0.63 1.26 0.69 38580 446 1.12 1.20 1.12 1.06 0.30 33235 243 1.09 1.23 1.13 0.91 0.83 38585 161 1.20 1.29 1.11 1.19 0.11 33307 273 1.18* 1.24 1.20 1.09 0.12 38605 327 1.13 1.22 1.04 1.13 0.20 33350 146 1.08 0.98 1.12 1.13 0.33 38620 243 1.29** 1.30* 1.25 1.32* <.01 33370 20 0.86 0.56 0.28 1.62 0.85 38670 37 1.03 0.80 1.72* 0.52 0.94 33415 18 1.03 0.88 1.19 0.99 0.86 38950 14 1.35 1.06 2.13 0.89 0.38 33565 223 1.02 0.87 1.09 1.09 0.48 40030 54 0.93 1.14 0.87 0.74 0.35 33595 203 1.27** 1.39* 1.26 1.14 0.04 40297 337 1.21* 1.19 1.21 1.21 0.02 33635 58 0.83 0.79 0.84 0.85 0.26 40370 123 1.27* 1.32 1.44* 1.05 0.08 33640 489 1.15* 1.18 1.07 1.18 0.08 40380 52 1.27 1.01 1.56 1.23 0.11 33675 277 1.16 1.16 1.14 1.18 0.08 40410 193 1.24* 1.21 1.35* 1.17 0.03 33720 201 1.00 0.93 0.91 1.16 0.59 40430 274 1.05 1.01 1.15 1.00 0.56 33850 10 1.15 0.72 1.41 1.28 0.56 40910 28 0.97 1.11 0.57 1.27 0.98 33940 107 0.89 1.01 0.86 0.81 0.20 40984 132 1.31** 1.65** 1.22 1.05 0.13 34120 218 1.16 1.09 1.13 1.26 0.05 40987 664 1.24** 1.27** 1.40** 1.07 0.08 34370 201 1.10 1.13 1.09 1.09 0.34 41775 516 1.26** 1.31** 1.19 1.28** <.01 34715 142 0.97 1.17 0.95 0.79 0.31 42355 44 1.45* 1.82* 1.72 0.90 0.20 35085 511 1.15* 1.09 1.26* 1.11 0.06 42490 396 1.25** 1.18 1.30* 1.28* <.01 35120 10 1.31 0.45 1.30 2.12 0.21 42685 10 0.91 0.81 0.78 1.17 0.93 35260 51 0.81 0.85 0.79 0.79 0.18 43040 30 0.87 1.18 0.65 0.82 0.35 35455 31 1.18 1.14 1.65 0.78 0.62 43320 274 1.19* 1.03 1.42** 1.12 0.02 35505 213 1.15 1.24 1.12 1.09 0.23 43360 179 1.20* 1.19 1.24 1.17 0.07 35755 13 1.07 1.07 0.90 1.22 0.77 43410 189 1.13 1.33* 1.09 0.96 0.59 35925 49 0.96 0.82 1.03 1.01 0.94 43660 100 1.12 1.00 1.28 1.05 0.34 35927 128 1.13 1.30 1.03 1.04 0.54 44000 453 1.18* 1.23* 1.13 1.17 0.05 36060 457 1.30** 1.35** 1.22 1.33** <.01 44030 93 1.03 1.02 0.87 1.19 0.63 36330 29 1.29 1.57 1.19 1.13 0.40 44440 22 0.88 0.76 1.26 0.66 0.56 36340 82 1.16 1.32 0.85 1.32 0.28 44870 38 0.98 1.19 0.97 0.80 0.66 36710 19 0.76 1.17 0.56 0.51 0.12 45655 18 0.77 0.64 0.82 0.85 0.42 36955 477 1.30** 1.44** 1.25* 1.21 <.01 45850 92 1.15 1.08 1.21 1.16 0.24 37330 12 1.25 1.80 0.82 0.92 0.85 45930 584 1.16* 1.19 1.17 1.12 0.12 37510 450 1.17* 1.23* 1.18 1.12 0.08 46240 130 1.04 1.12 0.96 1.04 0.85 37630 394 1.08 1.07 1.01 1.15 0.24 46470 11 1.65 0.52 2.42 1.65 0.09 38110 62 1.04 1.52* 1.00 0.62 0.44 46935 13 1.30 0.29 2.18 1.51 0.19 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal 46970 595 1.20** 1.23* 1.26* 1.10 0.08 54185 380 1.14 1.03 1.37** 1.02 0.11 47030 128 1.10 1.47* 0.88 0.97 0.93 54243 126 1.18 1.10 1.13 1.31 0.08 47270 537 1.18* 1.20 1.22* 1.14 0.05 54480 110 1.14 0.96 1.20 1.24 0.15 47700 103 1.02 1.01 0.92 1.15 0.70 54790 470 1.25** 1.24* 1.32** 1.19 <.01 48320 309 1.08 1.19 1.13 0.93 0.81 55460 463 1.16* 1.16 1.14 1.18 0.05 48535 395 1.26** 1.44** 0.95 1.39** <.01 56240 76 1.19 1.01 1.68** 0.91 0.26 48625 282 1.14 1.15 1.16 1.10 0.15 57210 66 1.18 1.38 0.88 1.31 0.30 48910 307 1.14 1.18 1.22 1.02 0.25 57280 22 1.37 1.57 1.69 0.89 0.39 49600 241 1.18* 1.27 1.21 1.06 0.15 57340 194 1.15 1.07 1.27 1.10 0.14 50195 170 1.11 1.49** 1.01 0.81 0.81 57740 10 2.08 2.54 2.15 1.62 0.11 50420 397 1.20** 1.24* 1.20 1.16 0.04 58520 513 1.13 1.11 1.17 1.10 0.13 50440 13 0.76 1.37 0.33 0.58 0.18 59115 129 0.99 1.07 0.98 0.93 0.76 50470 12 1.34 2.67* 0.64 0.71 0.98 59185 58 1.29 1.23 1.38 1.27 0.11 50480 34 0.98 0.71 1.11 1.12 0.80 59210 109 1.28* 1.43* 1.29 1.12 0.11 50510 53 1.18 1.15 1.12 1.27 0.27 59230 113 1.20 1.37 1.06 1.16 0.22 50742 176 1.04 0.90 1.14 1.09 0.42 59450 61 1.22 1.42 0.82 1.38 0.24 50795 9 1.14 0.39 1.36 1.68 0.42 59465 161 1.14 1.29 1.01 1.11 0.35 50865 258 1.00 0.95 1.00 1.04 0.83 60122 145 1.18 1.54** 0.82 1.16 0.38 50870 136 1.27* 1.49* 0.98 1.34 0.05 60125 29 1.56* 2.11* 1.61 0.96 0.19 50888 135 1.31** 1.49* 1.06 1.40* 0.02 60297 33 1.04 1.15 0.89 1.08 0.91 50890 57 0.91 1.00 1.03 0.70 0.31 60315 26 0.90 0.75 0.78 1.14 0.87 50910 308 1.12 1.12 1.13 1.11 0.17 60350 374 1.12 1.14 1.21 1.01 0.29 51090 17 1.01 1.14 1.10 0.77 0.87 60360 208 1.12 1.00 1.23 1.13 0.14 51100 104 1.18 1.30 0.98 1.26 0.21 60370 29 2.01** 2.55** 2.62** 0.92 0.02 51118 171 1.14 1.44** 0.95 1.03 0.57 60400 21 0.85 1.38 0.35 0.86 0.30 51705 102 1.24 1.28 1.18 1.25 0.09 60410 35 1.42 2.03* 1.41 0.90 0.34 51910 24 0.99 1.34 0.96 0.71 0.64 60420 119 1.28* 1.10 1.44* 1.31 0.02 52132 335 1.15 1.12 1.19 1.13 0.08 60440 519 1.17* 1.26* 1.17 1.08 0.18 52136 17 1.15 1.28 1.23 0.95 0.76 60490 52 1.03 1.41 0.75 0.95 0.76 52138 30 1.96** 2.29** 1.54 1.86 <.01 60540 204 1.18 1.24 1.01 1.29 0.06 52141 605 1.27** 1.29** 1.24* 1.27* <.01 60570 18 1.22 1.77 1.17 0.65 0.93 52142 38 1.08 0.99 1.04 1.18 0.59 60711 87 1.01 1.25 0.73 1.03 0.82 52145 117 1.18 1.05 1.21 1.30 0.08 60712 180 1.38** 1.30 1.49** 1.34 <.01 52190 477 1.15* 1.22* 1.20 1.03 0.25 60713 617 1.26** 1.15 1.27* 1.36** <.01 52480 137 0.95 1.00 0.90 0.95 0.55 60714 40 1.11 1.54 0.72 1.06 0.91 53615 21 1.19 1.28 1.18 1.10 0.58 60717 68 0.99 1.07 1.00 0.88 0.72 54160 141 1.28* 1.59** 1.17 1.06 0.15 60721 184 1.21* 1.25 1.15 1.21 0.07 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal 62000 45 0.90 1.25 1.00 0.49* 0.17 69090 203 0.97 0.92 1.20 0.79 0.49 62460 34 1.18 1.43 0.90 1.25 0.49 69120 11 0.49* 0.56 0.26 0.67 0.05 63265 71 1.11 1.22 1.25 0.88 0.75 69220 434 1.17* 0.97 1.39** 1.16 <.01 63525 603 1.11 1.15 1.13 1.06 0.33 69230 481 1.21** 1.24* 1.35** 1.06 0.06 63550 189 1.09 1.10 1.18 0.99 0.48 69270 110 1.20 1.34 1.18 1.09 0.25 65080 20 1.11 1.11 0.65 1.56 0.52 69330 11 0.65 0.58 1.13 0.31 0.15 66495 518 1.22** 1.32** 1.05 1.31** <.01 69375 29 0.91 0.62 1.18 0.95 0.87 66950 61 1.46* 1.60* 1.06 1.69* 0.02 69445 118 1.02 0.83 1.13 1.10 0.53 67220 37 0.75 0.67 0.85 0.74 0.16 69460 49 1.14 1.22 1.27 0.96 0.62 67405 23 0.93 0.99 1.51 0.34 0.41 69470 168 0.98 1.00 1.01 0.94 0.77 67410 30 0.94 0.45 0.70 1.62 0.55 69715 253 1.15 1.14 1.23 1.09 0.13 67537 300 1.26** 1.35** 1.15 1.27* 0.01 69730 20 0.83 0.93 1.18 0.36 0.26 67680 46 0.93 1.21 0.88 0.70 0.34 69740 490 1.21** 1.26* 1.16 1.22* 0.02 67915 314 1.21** 1.35** 1.12 1.17 0.06 69855 620 1.21** 1.22* 1.26* 1.14 0.05 67918 10 1.90 2.36 2.64 0.61 0.27 70130 310 1.04 1.04 0.94 1.14 0.46 68208 12 1.16 1.67 0.55 1.29 0.82 70131 242 1.03 1.11 0.85 1.12 0.73 68295 16 0.85 1.03 1.16 0.41 0.32 70845 431 1.18* 1.12 -1.24* 1.18 0.02 68508 32 1.36 1.69 1.10 1.24 0.27 70860 200 1.23* 1.38* 1.19 1.12 0.09 68509 65 0.99 0.97 0.96 1.04 1.00 70865 10 1.96 1.97 1.93 1.97 0.08 68512 110 0.93 0.93 0.95 0.92 0.54 70870 379 1.06 1.03 1.08 1.08 0.35 68657 584 1.16* 1.25* 1.06 1.17 0.12 70995 117 1.13 0.97 1.27 1.16 0.18 68695 469 1.23** . 1.21 1.22* 1.27* <.01 71025 11 1.14 0.95 1.33 1.15 0.66 68730 204 1.12 1.20 0.95 1.21 0.22 71055 536 1.13 1.18 1.09 1.12 0.17 68765 298 1.12 1.00 1.20 1.17 0.07 71058 20 0.69 1.02 0.70 0.40 0.06 68766 188 1.20* 1.40* 1.05 1.14 0.18 71095 25 1.00 1.23 0.72 1.07 0.92 68768 12 1.70 0.00 1.30 4.83** <.01 71640 20 1.30 1.68 0.96 1.22 0.48 68770 185 1.00 1.23 0.86 0.90 0.47 71695 292 1.17* 1.11 1.04 1.37** 0.01 68820 13 0.80 0.81 0.53 1.10 0.61 71900 30 0.90 1.02 0.68 0.99 0.61 68850 601 1.18* 1.29** 1.18 1.06 0.28 72200 13 0.87 0.24 1.14 1.12 0.99 68870 28 0.90 0.91 0.52 1.30 0.86 73075 420 1.30** 1.22 1.30* 1.38** <.01 68880 515 1.16* 1.24* 1.14 1.11 0.14 73255 271 1.16 1.15 1.27* 1.07 0.12 68900 100 1.22 1.37 1.60** 0.68 - 0.50 73300 573 1.21** 1.28** 1.15 1.20 0.04 68905 204 1.15 1.24 1.07 1.12 0.24 73390 45 1.60** 2.00* 1.27 1.56 0.02 68950 39 1.09 1.27 1.34 0.66 0.92 73470 39 1.01 1.49 0.46 1.10 0.79 69000 434 1.14 1.20 1.16 1.05 0.23 73525 168 1.16 1.28 1.23 1.00 0.32 69055 292 1.03 0.99 1.10 1.00 0.74 73730 135 1.20 1.28 1.08 1.23 0.12 69070 599 1.12 1.11 1.16 1.08 0.21 73790 345 1.21** 1.10 1.33** 1.20 <.01 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal 74010 204 1.27** 1.54** 1.31 0.99 0.12 80071 115 1.01 1.24 0.81 0.99 0.74 74175 78 1.18 1.13 1.29 1.10 0.26 80073 90 1.00 0.72 1.09 1.16 0.57 74430 14 1.42 1.38 1.01 1.96 0.19 80076 394 1.15 1.18 1.26* 1.01 0.22 74635 118 1.15 0.96 1.19 1.29 0.10 80079 417 1.10 1.09 1.14 1.06 0.27 74655 40 1.32 0.80 1.61 1.54 0.05 80083 296 1.06 1.05 1.10 1.02 0.58 74795 342 1.11 1.24* 1.16 0.94 0.64 80090 50 1.17 1.72* 0.77 1.02 0.78 74980 103 1.03 0.99 0.92 1.18 0.58 80092 220 1.18 1.40** 1.12 1.03 0.30 74990 295 1.28** 1.23 1.23 1.39** <.01 80094 74 1.02 1.07 0.75 1.14 0.83 75158 94 1.03 1.01 0.98 1.10 0.72 80105 78 1.40* 1.21 1.98** 1.09 0.03 76165 128 1.08 1.14 1.02 1.10 0.52 80109 550 1.18* 1.15 1.15 1.24* 0.01 76210 10 0.85 1.24 0.52 0.76 0.48 80123 88 1.26 1.08 1.41 1.31 0.05 76355 209 1.13 1.24 0.92 1.23 0.21 80133 15 0.65 0.77 0.71 0.49 0.10 76445 90 1.01 0.90 1.22 0.92 0.92 80140 197 1.15 1.22 1.12 1.12 0.20 76510 180 1.11 1.29 0.95 1.09 0.51 80142 64 1.01 0.88 1.68* 0.52* 0.71 76720 624 1.21** 1.23* 1.21* 1.18 0.04 80143 17 2.16** 2.65* 2.04 1.70 0.03 77115 413 1.21** 1.22 1.29* 1.14 0.03 80144 365 1.15 1.37** 1.05 1.02 0.45 77150 252 1.06 0.95 1.12 1.10 0.32 80145 11 0.63 0.70 0.39 0.76 0.19 77190 402 1.16* 1.22 1.23* 1.04 0.16 80148 9 0.67 0.93 0.61 0.49 0.19 77215 353 1.16* 1.20 1.16 1.10 0.12 80153 89 1.11 1.29 0.77 1.27 0.45 77220 45 1.12 1.29 1.24 0.84 0.84 80157 34 1.10 0.85 1.17 1.28 0.42 77265 106 1.12 1.27 0.80 1.31 0.32 80158 187 1.04 1.09 0.92 1.13 0.61 80004 13 1.08 1.57 0.52 1.06 0.94 80164 366 1.15 1.17 1.25* 1.03 0.18 80017 605 1.18* 1.14 1.17 1.25* 0.01 80165 292 1.16* 1.31* 1.03 1.14 0.17 80032 215 1.07 1.30* 0.82 1.08 0.82 80169 34 0.85 0.52 0.69 1.38 0.96 80037 328 1.13 1.26* 1.00 1.13 0.24 80175 100 1.10 0.90 1.04 1.36 0.16 80041 237 1.19* 1.23 1.20 1.15 0.07 80177 12 0.90 0.74 0.24 1.61 0.88 80047 164 1.11 1.16 1.04 1.15 0.30 80181 15 1.54 0.81 2.31 1.62 0.08 80048 147 1.06 1.14 1.10 0.93 0.89 80182 10 1.20 0.93 2.21 0.69 0.69 80049 62 0.91 0.85 0.99 0.89 0.55 80194 15 0.93 1.17 0.50 1.16 0.81 80051 438 1.19* 1.20 1.32** 1.04 0.08 80197 37 0.76 0.64 0.77 0.87 0.25 80053 382 1.17* 1.20 1.16 1.14 0.07 80199 18 1.00 0.73 1.51 0.69 0.98 80056 318 1.12 1.00 1.22 1.13 0.09 80200 45 0.95 0.88 1.31 0.70 0.62 80058 31 1.14 1.37 0.79 1.32 0.55 80201 38 0.78 1.14 0.31* 0.91 0.13 80059 150 1.23* 1.16 1.41* 1.13 0.06 80202 88 0.92 0.92 0.85 0.98 0.59 80061 342 1.16* 1.17 1.17 1.14 0.08 80203 23 1.00 0.74 0.87 1.37 0.68 80064 68 1.22 1.17 1.09 1.44 0.11 80206 23 1.13 1.32 1.30 0.76 0.89 80069 386 1.05 1.04 1.08 1.04 0.51 80214 296 1.07 1.17 1.09 0.96 0.76 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal 80215 42 1.70** 0.84 2.75** 1.43 <.01 80310 11 0.71 0.81 0.90 0.49 0.24 80216 442 1.14 1.23* 1.08 1.12 0.17 80314 19 1.01 0.51 1.03 1.43 0.56 80218 45 0.95 0.79 1.09 0.94 0.87 80323 16 0.70 0.79 0.96 0.36 0.12 80219 69 1.14 1.25 1.05 1.11 0.49 80331 9 0.97 0.25 2.84* 0.33 0.96 80220 64 1.24 1.07 1.50 1.17 0.13 80332 140 1.02 1.23 0.97 0.88 0.70 80223 434 1.10 1.25* . 1.09 0.98 0.67 80341 143 1.14 1.18 1.09 1.14 0.27 80224 305 1.08 1.01 1.41** 0.83 0.62 80343 436 1.08 1.00 1.13 1.11 0.17 80231 361 1.08 1.24* 1.00 1.00 0.80 80346 50 1.30 1.48 1.28 1.13 0.24 80235 188 1.08 0.85 1.11 1.25 0.12 80347 75 1.35* 1.15 1.74** 1.14 0.04 80237 43 1.23 1.36 1.09 1.21 0.34 80349 245 1.05 1.11 1.07 0.97 0.85 80243 455 1.21** 1.34** 1.13 1.16 0.05 80350 60 1.33 1.18 1.53 1.28 0.06 80244 277 1.09 1.17 1.00 1.12 0.35 80354 236 1.12 1.13 1.16 1.08 0.24 80248 357 1.12 1.30* 1.06 1.00 0.56 80358 17 0.79 0.32 1.06 0.94 0.65 80249 468 1.17* 1.08 1.16 1.25* 0.01 80365 118 1.08 0.98 1.22 1.06 0.41 80251 125 1.09 1.21 1.12 0.93 0.78 80368 228 1.08 1.02 1.29* 0.94 0.47 80257 50 0.82 1.14 0.50* 0.81 0.11 80369 45 0.82 0.55 1.00 0.92 0.48 80258 195 1.06 1.07 1.14 0.97 0.67 80371 14 0.85 1.56 0.35 0.68 0.32 80260 12 1.42 1.08 1.49 1.67 0.22 80372 14 0.99 1.00 0.70 1.26 0.91 80261 51 1.13 1.57 1.03 0.84 0.97 80381 15 0.74 0.18 1.07 0.88 0.56 80265 64 1.15 0.95 1.47 1.04 0.33 80389 28 0.83 0.92 0.24* 1.37 0.63 80268 124 1.05 1.07 1.12 0.97 0.77 80390 321 1.11 1.00 1.14 1.18 0.08 80270 22 1.11 1.09 1.16 1.10 0.67 80393 45 1.18 1.15 1.34 1.06 0.39 80273 17 1.21 1.21 1.57 0.94 0.62 80417 24 1.28 0.92 1.48 1.44 0.20 80276 11 1.24 1.35 1.71 0.64 0.75 80419 16 0.64 1.06 0.47 0.38 0.04 80282 26 1.42 1.65 1.51 0.99 0.28 80421 11 1.10 0.93 2.38 0.46 1.00 80283 232 1.03 0.98 1.06 1.04 0.67 80439 134 1.13 1.27 0.93 1.19 0.33 80285 46 1.47* 2.05** 0.86 1.54 0.08 80441 526 1.17* 1.29** 1.06 1.15 0.13 80286 133 1.12 1.20 1.11 1.03 0.47 80447 125 1.03 1.05 0.93 1.11 0.71 80287 97 1.07 1.23 0.94 1.04 0.77 80452 20 1.02 0.86 1.05 1.14 0.82 80288 274 1.13 1.29* 1.01 1.10 0.33 80461 141 1.05 1.20 1.01 0.95 0.99 80293 165 1.09 1.25 1.00 1.03 0.63 80487 18 0.94 0.75 0.97 1.11 1.00 80295 13 1.59 2.55* 0.95 1.34 0.35 80488 32 0.98 0.90 0.80 1.24 0.86 80298 466 1.20** 1.24* 1.15 1.21 0.02 80496 167 1.03 1.11 1.02 0.97 0.94 80299 332 1.17* 1.12 1.15 1.26* 0.02 80507 157 1.05 0.93 1.06 1.16 0.37 80300 20 0.98 0.48 1.42 0.96 0.82 80517 66 0.98 1.24 0.98 0.72 0-45 80301 13 0.89 1.17 0.87 0.62 0.51 80527 17 0.62 0.94 0.22* 0.72 0.06 80305 57 1.34 1.69* 1.22 1.04 0.26 80530 313 1.16* 1.22 1.26* 1.02 0.19 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal 80531 38 1.12 1.05 1.30 0.98 0.61 81085 11 0.76 0.43 0.64 1.21 0.73 80538 9 1.37 0.48 2.60* 0.60 0.38 81115 17 1.06 1.82 0.60 0.83 0.73 80542 333 1.20* 1.28* 1.14 1.17 0.05 81120 46 1.08 1.16 1.61 0.56 0.85 80545 19 0.93 0.77 0.29 1.79 0.70 81125 114 1.02 0.96 1.18 0.92 0.88 80549 17 1.21 0.67 1.71 1.15 0.37 81135 9 1.09 2.82* 0.38 0.27 0.40 80563 9 1.45 1.94 1.49 0.99 0.51 81350 252 1.15 1.13 1.09 1.25 0.06 80564 26 0.81 0.66 0.69 1.09 0.56 81355 11 0.85 0.47 1.15 0.91 0.80 80570 202 1.15 1.17 1.06 1.22 0.12 81390 242 0.97 1.11 1.03 0.77 0.20 80573 16 1.17 0.79 1.79 1.04 0.51 81440 13 0.57 0.53 0.46 0.69 0.10 80574 14 0.78 0.70 0.78 0.87 0.50 81455 18 0.62 0.84 0.41 0.60 0.05 80579 27 0.99 0.95 0.60 1.45 0.77 81460 279 1.19* 1.35* 1.10 1.12 0.13 80585 9 0.65 0.42 0.43 1.10 0.44 81510 140 1.10 1.06 1.10 1.16 0.28 80587 80 1.07 1.02 1.30 0.88 0.77 81515 225 1.06 1.14 0.86 1.18 0.43 80588 101 1.08 1.20 1.17 0.88 0.85 81560 10 1.03 1.03 1.28 0.80 0.97 80589 216 1.08 1.19 0.91 1.15 0.43 81650 302 1.14 1.25 1.15 1.02 0.33 80595 39 0.83 0.86 0.79 0.83 0.30 81651 74 0.99 0.81 1.13 1.03 0.84 80596 17 0.84 1.46 0.79 0.28 0.14 81663 497 1.18* 1.18 1.21* 1.13 0.05 80602 284 0.99 1.14 1.08 0.77* 0.24 81664 66 1.08 1.28 1.02 0.93 0.91 80611 291 1.14 1.11 1.19 1.12 0.11 81667 22 0.88 0.80 0.62 1.21 0.82 80612 134 1.10 1.17 1.13 1.01 0.54 81668 387 1.18* 1.26* 1.18 1.09 0.11 80625 378 1.18* 1.23 1.12 1.19 0.05 81671 128 1.30* 1.43* 1.48* 1.00 0.09 80675 290 1.21* 1.31* 1.22 1.09 0.08 81675 33 1.06 1.02 0.74 1.44 0.57 80680 71 1.10 0.99 1.16 1.16 0.39 81676 162 1.13 1.02 1.07 1.31 0.08 80685 94 1.12 1.51* 0.75 1.14 0.67 81679 103 1.09 0.94 1.04 1.29 0.24 80705 413 1.14 1.09 1.05 1.27* 0.02 81680 31 0.93 1.01 0.74 1.04 0.76 80720 12 1.36 2.26 0.69 1.05 0.72 81683 83 0.95 0.93 1.08 0.83 0.58 80725 94 1.15 1.04 1.26 1.14 0.24 81684 9 0.91 0.91 0.92 0.90 0.80 80780 23 1.27 1.74 0.91 1.21 0.51 81692 10 1.24 1.71 1.87 0.00 0.93 80785 99 1.23 1.59** 0.98 1.08 0.35 81695 12 0.84 0.23 1.27 0.93 0.87 80790 21 1.27 2.01 1.12 0.75 0.81 81696 17 0.73 0.99 0.69 0.53 0.15 80828 47 1.00 0.96 0.85 1.20 0.81 81698 24 1.16 1.65 1.23 0.59 0.95 80836 137 1.09 1.48** 1.11 0.69 0.59 81700 11 0.76 0.69 0.92 0.70 0.44 80900 23 1.26 1.01 1.28 1.47 0.24 81702 15 1.07 1.12 1.45 0.64 0.98 80945 21 0.88 0.80 1.02 0.83 0.64 81710 11 0.99 0.29 1.46 1.18 0.70 81005 20 1.07 1.37 0.96 0.84 0.93 81711 21 1.29 1.86 1.22 0.90 0.66 81040 248 1.08 1.17 1.07 1.00 0.63 81713 80 1.21 1.43 1.30 0.88 0.47 81080 42 0.95 0.61 1.06 1.24 0.72 81715 66 1.08 1.22 0.93 1.07 0.75 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal 81720 32 1.14 1.21 1.43 0.77 0.78 81887 205 1.03 1.02 1.11 0.96 0.88 81721 105 1.39** 1.50* 1.19 1.49* <.01 81891 13 0.95 0.27 1.27 1.20 0.78 81724 16 0.79 0.97 0.78 0.61 0.29 81894 36 1.09 1.37 1.02 0.92 0.93 81731 24 1.32 1.33 2.01 0.80 0.43 81905 13 0.97 1.24 0.91 0.72 0.70 81736 22 1.12 1.90* 0.86 0.63 0.74 81908 21 0.92 0.94 0.80 1.03 0.77 81741 101 1.04 0.99 1.16 0.98 0.75 81914 14 0.67 0.77 0.56 0.70 0.19 81751 169 1.23* 1.32 1.32 1.05 0.11 81915 252 1.08 1.07 1.02 1.16 0.25 81753 13 1.21 1.15 1.29 1.20 0.56 81921 144 1.03 1.14 0.90 1.05 0.91 81754 172 1.19 1.27 1.05 1.26 0.08 81922 30 1.16 0.82 1.52 1.13 0.37 81755 105 1.19 1.38 1.10 1.07 0.34 81931 17 1.91* 1.10 2.09 2.48* <.01 81763 9 1.16 0.90 1.93 0.84 0.70 81935 15 1.04 0.74 0.97 1.42 0.61 81767 25 0.85 0.79 0.71 1.05 0.62 81945 20 0.97 0.98 0.61 1.28 0.94 81770 10 1.36 0.84 2.43 0.50 0.47 81949 14 0.92 0.75 0.79 1.40 0.97 81777 13 0.98 1.28 0.51 1.08 0.85 81953 37 0.94 0.92 0.99 0.90 0.72 81779 422 1.13 1.08 1.21 1.10 0.11 81957 12 0.89 0.81 1.07 0.83 0.74 81787 27 1.46 1.13 1.48 1.78 0.05 81963 10 0.88 0.74 0.53 1.40 0.96 81800 28 1.36 1.48 1.62 0.95 0.31 81964 103 1.09 1.06 1.15 1.07 0.48 81806 108 0.97 0.81 0.99 1.10 0.84 81971 147 1.22* 1.41* 1.15 1.11 0.16 81811 54 0.97 0.98 1.17 0.83 0.70 81974 19 0.82 0.73 0.44 1.24 0.71 81815 357 1.09 0.97 1.19 1.09 0.17 81975 29 1.32 1.49 1.32 1.17 0.29 81821 54 0.98 0.98 1.17 0.87 0.80 81986 13 0.93 1.26 0.61 0.93 0.67 81826 10 0.83 1.40 0.55 0.49 0.32 81987 233 0.99 0.96 0.93 1.07 0.90 81830 16 0.96 1.77 0.66 0.40 0.34 81990 396 1.14 1.19 1.25* 0.99 0.28 81836 259 1.10 1.25 0.91 1.13 0.45 81991 338 1.20* 1.23 1.25* 1.12 0.05 81843 24 0.93 1.24 0.59 0.93 0.58 81992 275 1.24** 1.24 1.00 1.47** <.01 81851 557 1.20** 1.26* 1.13 1.20* 0.04 81993 233 1.17 1.28 1.08 1.14 0.16 81853 19 1.66 1.92 1.65 1.41 0.11 81999 103 0.98 1.00 1.01 0.93 0.78 81855 30 0.76 0.83 0.98 0.47 0.11 82001 12 1.23 0.95 1.94 0.66 0.63 81857 24 1.03 0.75 0.95 1.39 0.57 82002 41 1.12 1.13 1.32 0.88 0.68 81873 24 0.87 1.04 0.81 0.78 0.44 82006 12 1.70 0.40 2.69* 2.20 0.04 81876 24 1.03 1.37 0.79 0.98 0.90 82009 204 1.11 1.20 1.13 1.02 0.43 81877 33 1.34 1.53 1.26 1.19 0.26 82013 12 0.79 0.63 0.86 0.87 0.57 81879 295 1.13 1.06 1.20 1.14 0.09 82030 23 1.12 0.56 1.60 1.22 0.38 81882 119 1.18 1.38 1.10 1.07 0.31 82035 12 1.32 1.25 1.87 0.92 0.51 81884 47 0.80 0.82 0.87 0.73 0.16 82037 282 1.19* 1.11 1.15 1.31* 0.01 81885 44 1.07 1.21 0.96 1.05 0.81 82056 34 1.03 0.92 0.88 1.31 0.66 81886 13 0.97 0.68 1.05 1.17 0.88 82057 51 1.26 1.16 1.23 1.40 0.12 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal 82065 28 1.40 0.42 1.47 2.44** <.01 82806 49 0.84 0.98 0.93 0.63 0.15 82078 10 0.88 1.96 0.00 0.73 0.35 82807 91 1.08 1.05 0.84 1.36 0.31 82082 10 0.72 0.42 0.67 1.10 0.60 82815 13 1.08 1.61 0.83 0.84 0.90 82097 28 0.80 1.01 1.10 0.28* 0.10 82819 69 0.94 0.98 1.11 0.72 0.45 82100 186 1.12 1.11 1.22 1.03 0.31 82834 164 1.17 1.26 1.39* 0.91 0.34 82101 9 1.12 2.12 0.81 0.38 0.66 82840 74 1.03 1.01 0.95 1.13 0.74 82113 218 1.17 1.18 1.40** 0.95 0.22 82841 178 1.14 1.18 1.13 1.11 0.23 82118 15 1.38 1.40 1.20 1.59 0.27 82849 226 1.21* 1.37* 1.28 0.98 0.20 82120 18 0.64 0.67 0.54 0.69 0.10 82859 9 1.53 2.81* 0.00 1.61 0.54 82127 380 1.07 1.18 0.98 1.04 0.70 82861 187 1.24* 1.28 1.32* 1.12 0.05 82134 110 1.05 0.84 1.06 1.25 0.31 82869 10 1.07 0.58 1.27 1.44 0.57 82135 23 0.83 0.99 0.64 0.85 0.37 82871 36 0.93 0.95 1.03 0.81 0.64 82136 11 1.15 1.06 1.76 0.67 0.82 82880 276 1.13 1.05 1.13 1.20 0.08 82156 29 0.87 1.10 0.79 0.71 0.33 82886 33 1.16 1.89* 0.64 0.92 0.98 82164 11 0.66 0.37 0.91 0.72 0.34 82889 19 1.04 0.82 1.17 1.12 0.76 82177 25 0.74 1.09 0.50 0.67 0.10 82897 38 0.83 0.58 1.31 0.64 0.38 82181 140 1.15 1.48** 1.08 0.89 0.76 82905 397 1.19* 1.26* 1.24* 1.07 0.10 82184 130 1.09 0.94 1.11 1.22 0.22 82907 40 1.19 0.88 1.45 1.27 0.22 82187 18 0.64 0.50 0.87 0.54 0.10 82917 19 1.07 1.55 0.86 0.80 0.82 82206 20 1.24 0.36 1.69 1.75 0.13 82920 163 1.07 1.19 1.01 1.02 0.71 82207 12 1.12 0.53 1.91 0.73 0.64 82924 100 1.03 1.28 1.13 0.67 0.53 82208 46 1.28 1.09 1.77* 0.95 0.21 82927 74 1.01 0.98 1.20 0.89 0.99 82210 42 1.02 0.78 1.44 0.82 0.92 82934 81 0.86 0.75 1.11 0.72 0.24 82214 48 1.22 1.01 1.09 1.57 0.11 82942 115 0.96 0.68 1.02 1.17 0.66 82224 16 1.20 0.88 1.70 0.95 0.55 82946 130 1.14 1.37* 1.14 0.88 0.67 82233 12 1.20 0.64 1.60 1.30 0.43 82948 49 0.95 1.07 1.21 0.58 0.42 82253 176 1.05 0.95 1.12 1.08 0.46 82949 18 1.42 2.08 0.99 1.17 0.43 82254 194 1.24* 1.05 1.42** 1.24 0.01 82951 27 0.97 1.38 0.52 1.07 0.74 82256 21 0.97 1.56 0.78 0.58 0.44 82953 20 1.03 1.05 0.46 1.58 0.66 82272 34 1.13 1.32 1.80* 0.28* 0.81 82955 275 1.16 1.13 1.06 1.29* 0.03 82274 14 1.45 1.85 0.66 2.42 0.21 82960 25 1.73* 2.93** 1.05 1.42 0.11 82276 36 1.07 1.17 0.90 1.12 0.78 82963 169 1.29** 1.26 1.48** 1.13 0.02 82786 153 1.13 1.13 1.31 0.97 0.37 82967 9 1.26 1.93 1.14 0.81 0.84 82789 422 1.10 1.12 1.20 0.98 0.46 82978 112 1.16 1.24 1.23 1.00 0.37 82792 36 1.30 1.39 1.42 1.08 0.29 82994 9 0.91 0.89 0.57 1.33 0.95 82795 58 0.99 1.00 0.89 1.04 0.96 82995 305 1.14 1.21 1.11 1.09 0.20 82798 26 1.02 1.20 1.01 0.85 0.85 82998 143 1.08 1.31 1.08 0.84 0.93 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal 83002 104 1.11 1.18 1.07 1.09 0.46 83184 121 1.11 1.10 1.25 0.98 0.49 83007 99 1.13 1.08 1.05 1.26 0.23 83185 50 1.13 0.97 1.61* 0.81 0.58 83017 128 1.10 0.99 1.20 1.10 0.30 83186 12 0.86 0.26 1.23 1.00 0.93 83019 140 1.14 1.10 1.32 0.99 0.33 83189 11 1.90 2.08 1.57 2.06 0.09 83024 205 1.18* 1.33* 1.15 1.08 0.19 83190 123 1.34** 1.40 1.62** 0.99 0.05 83030 63 1.38* 1.32 1.67* 1.15 0.06 83193 15 2.04* 2.81* 1.97 1.47 0.07 83032 243 1.16 1.03 1.18 1.25 0.04 83194 10 1.61 1.52 3.03* 0.43 0.38 83033 60 1.19 0.93 1.52 1.13 0.19 83196 13 0.98 1.56 0.90 0.46 0.52 83038 117 1.22 1.37 1.04 1.25 0.12 83197 85 1.03 1.00 1.01 1.07 0.77 83046 280 1.05 1.08 1.04 1.04 0.62 83198 19 1.00 0.80 1.23 0.95 0.94 83048 27 2.49** 2.08 2.61* 2.81** <.01 83199 54 1.03 1.67* 0.62 0.88 0.55 83062 14 1.06 0.48 0.97 1.68 0.43 83200 37 1.06 0.91 1.18 1.09 0.66 83065 10 1.12 0.65 1.65 1.11 0.61 83201 39 1.21 1.18 1.31 1.14 0.35 83066 90 1.00 1.18 1.01 0.80 0.57 83204 101 1.29* 1.56* 1.13 1.20 0.10 83079 23 1.34 1.02 1.09 1.92 0.11 83205 15 1.20 1.22 1.76 0.53 0.77 83085 128 1.09 1.03 1.16 1.06 0.44 83207 10 1.38 1.54 1.11 1.52 0.40 83102 409 1.15* 1.16 1.18 1.10 0.11 83208 409 1.23** 1.24* 1.19 1.25* <.01 83104 26 1.17 0.80 1.57 1.13 0.38 83209 87 0.89 1.00 0.76 0.90 0.30 83105 52 1.34 1.32 1.18 1.51 0.06 83213 9 1.61 0.52 2.46 1.83 0.12 83110 261 1.25** 1.29* 1.26 1.21 0.01 83217 39 0.93 0.95 0.92 0.92 0.68 83111 156 1.11 0.89 1.24 1.19 0.13 83218 23 1.45 2.02 1.67 0.72 0.42 83115 16 1.41 0.80 0.55 2.81** 0.05 83224 150 1.06 1.24 1.16 0.79 0.80 83124 25 1.27 1.11 1.22 1.46 0.24 83233 10 1.73 1.97 2.14 1.06 0.24 83128 10 1.69 1.04 2.44 1.53 0.14 83248 9 1.16 0.66 2.56 0.39 0.77 83138 19 0.84 1.18 0.58 0.69 0.30 83252 30 0.79 0.86 0.43* 1.12 0.38 83140 19 1.09 1.20 0.95 1.13 0.79 83258 45 0.85 1.02 0.67 0.86 0.27 83142 274 1.20* 1.19 1.29* 1.11 0.06 83262 11 1.16 0.96 0.67 1.85 0.43 83150 9 1.84 1.48 1.12 2.89 0.08 83265 ' 108 1.16 1.05 1.10 1.31 0.12 83151 24 1.31 1.56 1.14 1.29 0.34 83271 218 1.20* 1.22 1.22 1.16 0.06 83152 14 0.96 1.34 0.21 1.33 0.91 83275 118 1.11 1.04 1.29 1.02 0.37 83162 14 1.79 1.54 2.35 1.37 0.09 83276 129 1.13 1.14 1.16 1.08 0.34 83166 133 1.10 1.28 1.13 0.87 0.88 83277 169 1.09 1.29 0.87 1.10 0.63 83170 43 1.65** 1.74 1.48 1.71 <.01 83278 122 1.04 1.16 0.94 1.02 0.93 83177 98 1.03 1.24 1.08 0.76 0.68 83279 22 1.11 0.84 1.36 1.15 0.54 83180 11 0.77 0.27 0.81 1.11 0.74 83280 269 1.08 1.19 1.13 0.91 0.87 83181 33 1.07 0.83 1.34 1.03 0.64 83290 152 1.11 1.21 1.13 0.99 0.55 83182 156 1.06 1.23 1.05 0.90 1.00 83293 304 1.20* 1.34** 1.03 1.25 0.04 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal 83299 99 1.01 0.99 1.02 1.03 0.87 83496 16 1.01 0.74 1.25 1.18 0.74 83302 10 2.22* 0.66 4.47** 0.86 0.03 83497 40 1.33 1.42 1.24 1.33 0.16 83307 199 1.14 0.99 1.25 1.18 0.08 83506 10 1.36 0.92 1.17 1.95 0.25 83308 14 1.07 0.63 1.54 1.05 0.65 83508 308 1.23** 1.23 1.26* 1.22 0.01 83323 24 1.23 1.22 0.83 1.73 0.26 83509 388 1.18* 1.14 1.27* 1.13 0.04 83331 152 1.11 1.18 1.03 1.12 0.38 83512 279 1.10 1.02 1.19 1.10 0.17 83332 146 1.36** 1.48* 1.42* 1.17 0.02 83513 35 0.92 1.08 0.95 0.73 0.47 83335 12 1.10 0.53 1.30 1.56 0.45 83514 177 1.06 1.00 1.34* 0.84 0.81 83351 136 1.10 0.93 .1.22 1.15 0.22 83515 23 1.19 0.66 1.74 1.27 0.27 83353 19 1.26 0.66 1.39 1.67 0.19 83517 15 0.89 1.16 0.84 0.70 0.52 83354 16 0.59* 0.83 0.30* 0.65 0.05 83551 9 0.95 1.08 1.16 0.61 0.74 83355 81 1.26 1.25 1.31 1.20 0.11 83553 163 1.04 0.84 1.03 1.24 0.24 83364 22 0.84 0.77 1.10 0.62 0.41 83554 304 1.19* 1.17 1.38** 1.03 0.08 83365 35 1.05 1.02 1.15 0.97 0.85 83562 136 1.33** 1.53** 1.25 1.20 0.03 83369 28 0.80 1.01 1.10 0.28* 0.10 83571 47 1.23 1.50 1.15 1.03 0.46 83376 15 1.00 0.72 1.24 1.01 0.90 83572 30 1.19 1.07 1.00 1.52 0.28 83379 18 1.13 1.23 1.04 1.11 0.72 83573 20 1.16 0.74 1.96 0.95 0.46 83383 463 1.12 1.08 1.24* 1.03 0.22 83574 27 1.46 1.50 1.49 1.41 0.12 83404 12 1.10 0.70 1.46 1.20 0.61 83581 119 1.12 1.29 1.07 0.99 0.61 83413 74 1.30 1.02 1.71** 1.15 0.05 83587 30 1.09 1.39 1.05 0.84 0.98 83433 79 1.14 1.29 0.97 1.18 0.41 83589 40 0.93 0.62 1.53 0.67 0.73 83434 272 1.17* 1.14 1.11 1.25 0.04 83596 23 1.13 1.47 0.59 1.31 0.74 83435 17 1.08 1.60 0.80 0.88 0.90 83598 9 0.84 0.63 0.57 1.31 0.90 83436 34 1.25 1.52 0.91 1.27 0.38 83600 9 0.58 1.09 0.00 0.56 0.08 83440 119 1.31* 1.20 1.42* 1.32 0.01 83609 68 1.26 1.67* 1.01 1.09 0.35 83441 87 1.27 1.54* 1.18 1.06 0.23 83626 243 1.10 1.13 1.11 1.06 0.35 83444 142 1.28* 1.50** 1.09 1.24 0.06 83628 34 1.07 0.70 1.05 1.43 0.38 83446 16 0.85 0.76 0.96 0.84 0.62 83629 52 1.29 1.45 1.27 1.13 0.23 83447 103 1.32* 1.28 1.30 1.37 0.02 83639 106 1.24 1.21 1.29 1.22 0.08 83449 175 1.12 1.20 1.03 1.15 0.27 83641 13 1.28 1.61 1.32 0.87 0.70 83451 97 0.98 0.96 0.92 1.05 0.97 83643 68 0.97 1.01 1.04 0.88 0.73 83453 235 1.09 1.02 1.19 1.07 0.27 83646 80 1.20 1.21 0.82 1.60* 0.08 83461 42 1.11 1.19 0.83 1.29 0.53 83649 19 1.01 0.51 1.03 1.43 0.56 83475 208 1.11 1.15 1.10 1.08 0.34 83660 62 1.05 1.00 1.28 0.88 0.85 83477 53 1.22 1.09 1.38 1.20 0.20 83664 17 0.93 1.41 0.78 0.64 0-48 83480 38 1.48* 1.91* 1.16 1.41 0.10 83665 188 0.96 1.13 0.94 0.81 0.25 83495 155 1.28* 1.20 1.22 1.42* <.01 83669 154 1.12 1.11 1.28 0.99 0.38 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal 83676 9 1.02 0.61 1.56 1.00 0.81 83937 68 1.18 1.31 1.34 0.90 0:48 83678 155 1.06 1.00 1.11 1.08 0.47 83946 258 1.17* 1.05 1.15 1.32* 0.01 83681 104 1.27* 1.50* 1.29 1.01 0.20 83951 14 1.41 2.71* 1.34 0.31 0.93 83685 189 1.21* 1.11 1.28 1.25 0.03 83952 29 0.85 1.10 0.72 0.73 0.28 83705 18 1.02 1.62 0.66 0.83 0.71 83967 12 0.78 0.61 0.94 0.80 0.53 83718 21 1.08 0.70 1.02 1.48 0.46 83987 49 1.30 1.75* 0.80 1.20 0.33 83726 12 0.42** 0.52 0.35* 0.38 <.01 84001 139 1.31** 1.29 1.24 1.39* <.01 83731 13 1.08 1.57 0.91 0.78 0.89 84030 141 1.19 1.05 1.33 1.18 0.08 83732 14 0.69 0.80 0.30 0.97 0.29 84031 9 1.26 0.40 1.94 1.55 0.32 83734 51 1.23 1.74* 1.15 0.80 0.75 84035 12 1.07 1.78 0.80 0.73 0.78 83736 356 1.22** 1.15 1.35** 1.16 0.01 84037 11 0.48* 0.24* 0.74 0.52 0.06 83739 19 1.10 0.54 0.86 1.85 0.29 84048 17 1.11 1.29 1.61 0.41 0.90 83741 70 1.26 0.96 1.78** 1.05 0.09 84063 18 0.92 1.36 0.59 0.82 0.51 83748 159 1.17 1.11 1.16 1.23 0.09 84077 69 1.27 1.25 1.07 1.49 0.07 83758 138 1.07 1.00 1.01 1.19 0.35 84081 9 1.95 1.76 2.04 2.11 0.09 83760 28 0.92 0.87 0.80 1.08 0.83 84086 14 0.85 0.68 1.02 0.81 0.64 83765 14 1.53 0.36 3.21** 1.08 0.13 84090 25 1.28 1.81 1.61 0.45 0.84 83770 78 1.03 0.92 0.99 1.18 0.57 84093 98 1.25 1.50* 1.02 1.21 0.18 83786 90 1.19 1.17 1.22 1.18 0.19 84097 15 0.98 0.83 1.02 1.07 0.96 83788 108 1.28* 1.16 1.32 1.37 0.02 84100 56 0.95 0.96 1.09 0.86 0.67 83800 18 0.76 0.65 0.65 1.07 0.48 84105 14 1.30 1.38 0.00 2.46* 0.22 83818 142 1.06 1.23 1.06 0.91 0.98 84116 206 1.13 1.27 1.07 1.06 0.35 83820 172 1.22* 1.29 1.32 1.05 0.11 84118 12 0.69 0.54 0.49 1.07 0.43 83823 25 1.12 1.24 1.16 0.97 0.76 84133 9 0.96 0.95 0.72 1.19 0.99 83830 16 1.11 1.30 0.87 1.14 0.79 84153 28 1.26 1.18 1.86 0.82 0.44 83831 127 1.10 1.19 1.09 1.03 0.54 84154 351 1.16* 0.98 1.30* 1.21 0.01 83835 22 0.96 0.80 1.49 0.55 0.74 84160 23 1.04 0.77 1.52 0.85 0.85 83844 50 1.13 1.53 1.26 0.67 0.90 84180 9 0.59 0.64 0.33 0.87 0.21 83849 15 0.85 0.41 0.98 1.11 0.90 84183 69 1.28 1.19 1.39 1.25 0.09 83866 10 0.90 0.80 1.36 0.54 0.69 84192 10 0.83 0.91 0.66 0.93 0.62 83871 168 1.22* 1.56** 1.03 1.07 0.25 84195 28 0.90 0.54 0.74 1.45 0.78 83872 79 1.04 0.94 0.81 1.33 0.44 84203 46 1.15 1.01 1.17 1.27 0.31 83889 36 1.09 1.27 0.97 1.03 0.83 84204 352 1.12 1.15 1.07 1.15 0.15 83904 175 1.12 1.39* 1.01 0.97 0.71 84233 125 1.21 1.26 1.10 1.25 0.10 83906 15 0.93 1.01 0.84 0.94 0.79 84235 41 1.15 1.41 1.01 1.00 0.71 83911 19 0.95 1.25 0.78 0.79 0.61 84238 45 1.14 1.76* 0.95 0.69 0.81 83919 13 0.96 0.39 1.43 1.15 0.77 84240 21 1.30 0.77 1.59 1.55 0.16 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal 84269 30 1.43 1.33 1.47 1.50 0.09 84475 14 1.52 1.36 1.97 1.30 0.21 84274 22 1.27 1.39 0.98 1.40 0.37 84477 133 1.21 1.44* 1.24 0.96 0.33 84287 28 1.91** 2.64** 1.75 1.38 0.04 84478 11 1.37 1.60 1.09 1.38 0.46 84295 196 1.16 1.30 1.01 1.17 0.18 84479 14 0.68 0.69 0.46 0.88 0.26 84296 9 0.91 2.32 0.00 0.49 0.29 84480 14 0.84 1.94 0.69 0.15 0.12 84313 11 0.84 0.66 0.87 1.01 0.76 84494 141 1.08 1.19 1.09 0.94 0.79 84314 9 1.25 1.95 1.33 0.47 0.94 84495 73 1.31* 1.34 1.20 1.38 0.06 84318 122 1.18 1.37 1.13 1.04 0.35 84499 59 0.98 0.99 0.98 0.97 0.87 84326 10 1.15 1.26 0.69 1.52 0.64 84505 35 1.01 0.96 0.86 1.21 0.81 84330 24 1.40 1.90 0.98 1.33 0.30 84508 37 1.13 1.11 1.44 0.84 0.71 84335 65 1.32 1.21 1.56 1.21 0.07 84513 85 1.03 0.78 1.11 1.19 0.47 84341 83 1.29* 1.56* 1.20 1.10 0.18 84515 9 1.10 1.48 0.97 0.99 0.93 84349 33 0.77 1.18 0.44* 0.71 0.08 84526 194 1.08 1.14 1.02 1.08 0.51 84352 21 1.22 1.78 0.94 0.97 0.75 84535 12 0.91 0.89 0.81 1.06 0.85 84364 32 1.27 0.83 2.06* 0.96 0.22 84537 256 1.19* 1.07 1.22 1.29* 0.01 84370 9 1.51 0.48 2.37 1.68 0.16 84544 39 1.05 1.01 0.79 1.40 0.55 84376 264 1.13 1.20 1.19 1.01 0.31 84549 70 1.34* 1.36 1.42 1.23 0.07 84381 24 1.04 1.11 1.22 0.82 0.96 84563 318 1.15 1.14 1.18 1.11 0.11 84383 139 1.27* 1.24 1.26 1.30 0.03 84566 10 1.60 2.20 1.09 1.74 0.27 84386 135 1.04 1.29 1.06 0.80 0.68 84569 16 0.70 0.68 0.83 0.58 0.20 84407 13 0.93 1.38 0.65 0.71 0.54 84613 16 1.30 1.55 1.15 1.23 0.47 84414 10 0.86 0.74 0.73 1.14 0.82 84620 130 1.13 1.08 1.25 1.05 0.32 84425 182 1.22* 1.04 1.25 1.37* <.01 84628 18 0.78 0.89 0.28 1.15 0.48 84426 37 1.02 0.74 0.96 1.36 0.52 84646 20 1.11 1.29 0.40 1.53 0.59 84427 94 1.15 0.89 1.50* 1.08 0.19 .84662 150 1.26* 1.32 1.08 1.38* 0.02 84428 34 1.11 1.17 1.29 0.86 0.79 84674 136 1.10 1.15 1.10 1.05 0.47 84443 74 1.04 1.20 1.08 0.81 0.78 84696 217 1.16 1.27 1.17 1.06 0.22 84445 • 11 1.23 0.90 1.34 1.51 0.40 84705 10 0.91 0.30 1.51 0.84 0.99 84446 9 2.20* 1.74 2.96* 1.61 0.06 84716 210 1.13 1.17 1.03 1.21 0.15 84447 12 0.96 1.33 0.29 1.02 0.73 84718 13 1.32 1.57 1.40 0.85 0.56 84458 17 1.06 1.14 0.42 1.54 0.65 84736 25 1.32 0.67 1.65 1.59 0.10 84462 15 2.01* 2.73* 1.38 2.09 0.05 84743 138 1.15 1.03 1.20 1.23 0.11 84463 9 1.10 0.66 1.46 1.25 0.62 84745 26 0.82 0.77 0.58 1.08 0.55 84468 73 0.83 0.79 0.90 0.81 0.21 84754 10 1.28 1.29 1.70 0.85 0.62 84470 13 1.54 1.20 3.16** 0.27 0.38 84755 97 1.08 0.95 1.07 1.22 0.34 84472 45 0.94 0.98 1.08 0.77 0.57 84758 38 1.12 0.92 1.25 1.18 0.44 84473 9 1.92 3.72** 1.35 0.61 0.48 84765 12 0.93 0.87 1.07 0.85 0.83 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal 84772 124 1.08 1.08 1.09 1.08 0.50 92740 117 0.92 1.05 0.83 0.88 0.31 84789 46 1.34 1.16 2.15* 1.12 0.11 92780 248 1.19* 1.22 1.15 1.22 0.04 84805 15 0.93 0.68 1.02 1.12 0.98 92850 82 1.41** 1.54* 1.65* 1.04 0.05 84809 45 1.23 1.59 0.73 1.38 0.34 92910 15 1.39 1.59 1.35 1.22 0.37 84830 72 1.27 1.51 1.00 1.30 0.17 92930 12 1.56 1.26 2.02 1.41 0.19 84832 23 1.02 0.93 1.40 0.70 0.95 92960 10 1.32 0.91 0.78 2.14 0.25 90310 337 1.20* 1.34** 1.07 1.18 0.08 92980 9 0.87 0.81 0.30 1.51 0.96 90320 499 1.29** 1.40** 1.13 1.34** <.01 94140 313 1.03 1.03 0.97 1.08 0.64 90340 471 1.26** 1.33** 1.21 1.24* <.01 94220 481 1.21** 1.18 1.15 1.30** <.01 90410 10 1.35 0.00 1.78 2.66 0.10 A1021 97 1.12 0.83 1.62** 0.92 0.30 90590 375 1.29** 1.33** 1.27* 1.28* <.01 A1049 22 0.65 0.72 1.15 0.16* 0.02 90620 39 1.42 1.50 1.27 1.49 0.08 A1053 10 1.29 0.88 1.48 1.46 0.38 90800 38 1.35 1.33 1.46 1.28 0.14 A1065 184 1.08 1.00 1.09 1.15 0.29 90820 215 1.13 1.14 1.06 1.20 0.14 A1070 10 0.93 1.24 0.00 1.74 0.97 90870 291 1.24** 1.13 1.46** 1.13 <.01 A1073 169 1.21* 1.35* 1.16 1.15 0.12 90880 552 1.23** 1.15 1.23* 1.31** <.01 A1075 33 1.28 1.31 1.23 1.28 0.26 90883 148 1.20 1.14 1.10 1.37* 0.04 A1082 19 0.89 0.39 1.66 0.76 0.86 90885 507 1.27** 1.26* 1.28* 1.27* <.01 A1091 17 1.39 0.64 1.15 2.75* 0.04 90900 156 1.08 1.08 1.09 1.06 0.49 A1112 149 1.10 1.35* 0.98 0.97 0.82 90980 228 1.19* 1.19 1.19 1.20 0.05 A1165 24 1.26 1.46 0.62 1.70 0.28 91095 559 1.25** 1.40** 1.17 1.19 0.04 A1167 18 1.20 0.96 1.62 1.11 0.45 91110 83 1.31* 1.08 1.27 1.60* 0.01 A1179 131 1.13 1.34 1.17 0.89 0.68 91115 15 1.20 0.97 1.09 1.59 0.38 A1200 79 1.39* 1.65* 0.96 1.54* 0.03 91120 45 0.83 0.59 0.98 0.92 0.47 A1204 223 1.10 1.08 1.01 1.21 0.18 91150 13 0.76 0.54 1.57 0.31 0.31 A1214 67 1.12 1.08 1.09 1.19 0.39 91190 15 0.97 1.32 0.68 0.98 0.78 A1216 17 0.87 0.94 0.77 0.90 0.60 92150 54 0.93 0.99 1.08 0.75 0.45 A1220 87 1.14 1.33 1.15 0.92 0.66 92255 56 0.98 0.97 1.19 0.76 0.71 A1221 17 1.57 2.12 1.23 1.39 0.22 92290 35 1.23 1.18 1.41 1.11 0.33 A1242 58 1.28 1.21 1.48 1.15 0.14 92310 12 1.06 1.23 0.94 1.01 0.94 A1259 20 1.05 1.99* 0.49 0.72 0.55 92320 15 0.82 0.73 1.02 0.72 0.50 A1262 10 0.60 1.25 0.00 0.51 0.06 92355 133 1.14 1.41* 0.99 1.03 0.55 A1278 9 1.22 0.77 2.57 0.46 0.68 92470 44 1.08 1.04 0.97 1.21 0.57 A1279 93 1.10 1.05 1.43* 0.83 0.66 92500 51 1.74** 0.80 2.30** 2.15** <.01 A1324 19 1.29 1.17 1.50 1.20 0.36 92630 54 0.97 0.99 1.16 0.84 0.73 A1328 15 1.42 1.97 1.30 0.88 0.56 92650 24 1.30 1.99* 0.86 0.92 0.73 A1329 23 1.23 1.05 0.63 2.07* 0.17 92685 42 1.05 1.47 1.30 0.47 0.50 A1337 12 1.44 2.09 2.07 0.49 0.72 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal A1339 93 1.21 1.21 1.50* 0.93 0.26 M0238 15 0.64 0.77 0.93 0.24 0.06 A1346 207 1.10 1.11 1.17 1.01 0.44 M0239 98 1.29* 1.32 1.27 1.28 0.05 A1353 66 1.23 1.26 1.14 1.29 0.18 M0244 25 1.04 1.46 0.44 1.29 0.96 A1355 44 0.86 0.91 1.02 0.66 0.27 M0256 147 0.97 0.90 0.90 1.12 0.88 A1356 22 1.06 1.33 1.56 0.39 0.69 M0259 176 1.14 1.31 1.02 1.10 0.33 A1357 15 0.80 0.49 1.63 0.42 0.42 M0260 513 1.16* 1.15 1.27* 1.08 0.09 A1358 81 1.17 1.21 1.20 1.09 0.34 M0262 209 1.05 0.90 1.14 1.10 0.34 A1359 17 1.29 1.43 1.17 1.25 0.44 M0264 23 1.04 0.95 0.53 1.68 0.52 A1360 13 1.03 1.59 0.47 1.02 0.82 M0321 16 . 0.84 0.38 1.19 0.86 0.73 A1458 15 1.10 0.65 0.91 1.79 0.39 M0327 29 1.27 1.12 1.83 0.97 0.35 A1463 65 0.98 1.04 0.85 1.02 0.85 M0347 321 1.13 1.11 1.27* 1.02 0.20 A1466 31 1.27 0.93 1.23 1.72 0.11 M0377 59 1.20 1.33 1.38 0.85 0.50 A1481 11 1.19 1.19 0.36 1.94 0.42 M0386 71 0.98 0.89 1.12 0.92 0.92 A1491 12 1.69 2.82* 0.80 1.55 0.28 M0421 149 1.14 1.27 0.79 1.37* 0.16 A1515 41 1.28 0.92 1.78* 1.12 0.15 M0430 115 1.17 1.10 1.14 1.26 0.13 A1604 79 1.27 1.20 1.83** 0.85 0.19 M0451 20 1.22 0.54 1.61 1.56 0.20 A1642 32 0.98 1.01 0.81 1.12 0.98 M0461 21 0.98 1.31 0.78 0.86 0.70 A1667 11 1.29 0.71 1.09 2.05 0.23 M0462 55 0.93 0.98 0.79 1.02 0.68 A1728 29 1.24 1.03 1.63 1.05 0.33 M0478 16 0.95 1.23 0.72 0.89 0.71 A1771 27 1.54* 1.94* 1.59 1.00 0.18 M0527 209 1.27** 1.46** 1.27 1.08 0.06 A1827 25 1.25 0.96 1.74 1.04 0.33 M0529 109 1.20 1.46* 1.08 1.07 0.32 A1874 18 0.90 0.72 0.50 1.55 0.92 M0538 74 1.31* 1.65* 1.12 1.09 0.23 B0043 9 0.64 0.44 0.65 0.82 0.34 M0539 85 1.23 1.32 0.75 1.56* 0.08 B0044 175 1.14 1.20 1.01 1.21 0.18 M0577 176 1.12 1.26 0.88 1.19 0.33 B0045 13 1.92* 2.65* 1.38 1.69 0.12 M0578 94 1.12 1.18 0.79 1.38 0.26 B0105 317 1.10 1.24* 1.15 0.91 0.81 M0579 306 1.15 1.16 1.24 1.06 0.13 L0035 20 0.95 0.54 0.98 1.37 0.73 M0599 106 1.03 1.32 0.93 0.83 0.61 L0112 20 0.98 0.97 0.68 1.36 0.88 M0600 423 1.11 1.08 1.12 1.12 0.15 M0006 95 0.88 0.78 0.91 0.96 0.48 M0602 334 1.13 1.27* 1.05 1.07 0.35 M0073 153 1.11 1.40* 1.00 0.95 0.80 M0603 467 1.16* 1.19 1.25* 1.03 0.18 M0125 85 1.05 1.06 0.90 1.20 0.59 M0609 138 1.06 1.06 1.01 1.12 0.52 M0126 114 1.12 0.97 1.21 1.18 0.21 M0626 18 1.43 1.69 0.84 1.83 0.20 M0130 145 1.10 1.01 1.26 1.02 0.37 M0627 11 1.10 1.01 1.15 1.13 0.75 M0132 59 0.91 1.10 0.68 0.95 0.43 M0628 454 1.25** 1.27* 1.20 1.28* <.01 M0155 111 0.92 1.06 0.79 0.91 0.36 M0644 71 1.04 1.29 0.89 0.90 0.80 M0156 97 0.94 0.90 0.88 1.03 0.75 M0645 90 1.33* 1.63** 1.28 1.04 0.16 M0218 265 1.14 1.20 1.09 1.13 0.19 M0646 21 1.13 1.02 1.02 1.37 0.51 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal M0647 554 1.08 1.10 1.03 1.10 0.32 M0789 15 1.20 1.61 0.00 2.10 0.43 M0648 15 1.17 0.87 0.99 1.68 0.36 M0794 166 1.38** 1.18 1.51** 1.44* <.01 M0650 41 1.03 1.23 0.75 1.08 0.99 M0799 245 1.10 0.98 1.25 1.07 0.20 M0651 368 1.20* 1.15 1.29* 1.15 0.02 M0812 71 1.10 1.12 0.72 1.48 0.31 M0652 19 0.86 0.96 0.66 0.97 0.57 M0826 32 1.32 1.07 1.74 1.16 0.17 M0653 363 1.13 1.18 1.20 1.01 0.30 M0833 131 1.10 1.05 1.16 1.11 0.32 M0661 512 1.22** 1.16 1.20 1.30** <.01 M0850 190 1.04 1.27 0.95 0.92 0.79 M0662 149 1.08 1.03 1.08 1.12 0.38 M0863 10 0.89 1.86 0.24 0.76 0.42 M0674 100 1.30* 1.52* 1.18 1.19 0.10 M0867 16 0.92 1.00 1.15 0.64 0.60 M0675 9 0.86 0.59 1.71 0.28 0.59 M0870 94 1.21 1.00 1.48* 1.16 0.10 M0679 56 1.19 1.27 0.91 1.34 0.27 M0873 92 1.08 0.96 1.09 1.18 0.38 M0680 80 1.04 0.95 0.93 1.22 0.54 M0877 9 1.29 0.44 1.47 2.08 0.24 M0682 415 1.20** 1.25* 1.20 1.16 0.03 M0879 10 1.43 0.89 1.36 2.17 0.19 M0683 59 1.11 1.19 1.29 0.83 0.80 M0881 386 1.14 1.14 1.17 1.12 0.11 M0689 35 1.25 1.16 1.35 1.22 0.26 M0888 78 1.26 1.13 1.40 1.29 0.07 M0692 107 1.27* 1.51* 1.54* 0.79 0.29 M0892 17 1.18 0.98 0.46 2.03 0.27 M0698 39 1.11 1.66* 0.95 0.73 0.77 M0894 31 1.00 0.47 1.19 1.37 0.47 M0699 18 1.50 2.21 1.84 0.48 0.51 M0899 37 1.17 1.08 0.96 1.46 0.28 M0700 464 1.17* 1.34** 1.10 1.07 0.26 M0900 161 1.09 1.06 0.94 1.28 0.21 M0701 398 1.25** 1.26* 1.44** 1.07 0.02 M0905 14 0.74 0.68 0.63 0.90 0.41 M0708 58 1.04 0.85 1.35 0.93 0.73 M0909 26 1.03 1.07 1.26 0.76 0.93 M0716 20 1.12 0.80 1.34 1.24 0.49 M0912 328 1.08 1.10 1.03 1.12 0.30 M0717 16 1.54 2.45* 1.00 1.27 0.35 M0913 654 1.20** 1.19 1.22* 1.20 0.02 M0720 12 0.81 1.10 0.20 1.15 0.56 M0916 225 1.10 1.07 1.14 1.10 0.26 M0725 176 1.13 0.96 1.35* 1.06 0.18 M0918 16 1.20 1.61 1.16 0.79 0.83 M0745 235 1.24** 1.50** 0.98 1.25 0.06 M0920 31 1.30 1.74 1.01 1.10 0.45 M0747 366 1.12 1.15 1.11 1.11 0.17 M0926 323 1.11 1.16 1.15 1.03 0.34 M0749 24 0.73 0.87 0.61 0.72 0.15 M0927 19 0.91 0.48 0.90 1.30 0.89 M0752 29 1.33 0.90 1.17 2.05* 0.05 M0928 15 0.95 1.05 0.62 • 1.26 0.95 M0756 71 1.01 0.99 0.82 1.21 0.73 M0930 624 1.23** 1.33** 1.15 1.23* 0.03 M0760 288 1.18* 1.34* 1.04 1.17 0.11 M0937 9 0.91 1.19 0.60 0.95 0.72 M0773 331 1.23** 1.31* 1.08 1.29* 0.01 M0939 88 1.05 1.46* 0.78 0.91 0.71 M0774 39 1.13 1.20 0.74 1.49 0.39 M0947 39 1.34 1.02 1.81* 1.19 0.10 M0779 27 1.03 1.02 1.15 0.92 0.96 M0950 47 1.20 0.98 1.32 1.32 0.19 M0783 22 0.85 1.02 0.72 0.82 0.42 M0951 132 1.12 0.92 1.53** 0.91 0.33 M0785 21 0.99 0.57 1.43 1.14 0.69 M0952 10 0.64 0.56 0.58 0.80 0.28 M0787 40 1.07 0.70 1.24 1.25 0.41 M0959 16 1.15 1.03 0.82 1.57 . 0.45 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal M0960 83 1.08 1.07 1.08 1.10 0.52 M1190 66 1.26 1.11 1.49 1.21 0.10 M0961 17 1.12 1.17 1.60 0.61 0.93 M1199 26 1.22 0.99 1.10 1.59 0.23 M0969 235 1.11 1.14 1.17 1.01 0.41 M1202 118 1.11 1.38 1.01 0.97 0.76 M0983 11 1.15 2.04 0.91 0.60 0.83 M1203 74 1.01 1.01 1.10 0.93 0.99 M0984 176 1.48** 1.57** 1.34 1.52** <.01 M1205 14 1.50 0.31 1.93 2.36 0.04 M0985 101 1.03 0.94 1.10 1.06 0.67 M1207 89 1.12 1.19 0.99 1.16 0.44 M1000 744 1.20* 1.22* 1.15 1.21* 0.06 M1211 235 1.22* 1.43** 1.15 1.09 0.13 M1002 77 1.41** 1.95** 1.33 0.97 0.15 M1217 82 1.03 1.51* 0.99 0.63 0.35 M1010 12 0.92 1.17 1.22 0.42 0.52 M1218 125 1.10 1.24 1.14 0.94 0.69 M1023 46 0.89 0.90 1.12 0.66 0.34 M1226 245 1.18* 1.08 1.14 1.30* 0.02 M1026 29 0.95 0.58 1.11 1.16 0.83 M1232 19 0.97 0.76 0.16 1.97* 0.52 M1027 90 1.23 1.21 1.41 1.06 0.17 M1289 80 1.16 1.32 1.04 1.10 0.43 M1028 21 1.15 0.68 1.63 1.13 0.43 M1300 96 1.11 1.64** 0.82 0.90 0.89 M1030 261 1.06 1.08 1.09 1.02 0.56 M1309 174 1.08 1.22 0.95 1.07 0.62 M1042 88 1.04 0.86 1.22 1.02 0.62 M1312 225 1.11 1.21 0.98 1.12 0.36 M1047 207 1.15 1.13 0.99 1.34* 0.05 M1320 11 0.72 1.81 0.19 0.19 0.06 M1051 16 0.77 0.17 1.02 1.00 0.67 M1327 100 0.99 0.91 0.90 1.16 0.74 M1055 187 1.00 1.21 0.94 0.85 0.45 M1341 67 1.06 1.38 0.62 1.22 0.83 M1102 24 1.21 0.83 0.99 1.85 0.16 M1342 212 1.05 1.09 1.00 1.05 0.69 M1105 104 1.24 1.33 1.32 1.06 0.17 M1348 79 0.94 1.03 1.04 0.77 0.44 M1112 160 1.14 1.05 1.28 1.09 0.19 M1351 12 0.96 0.82 0.71 1.33 0.90 M1114 48 0.99 0.91 1.07 0.98 1.00 M1381 64 1.10 0.67 1.22 1.37 0.18 M1128 59 0.88 0.98 1.04 0.65 0.20 M1392 20 1.13 0.90 1.21 1.26 0.51 M1130 21 1.07 0.89 0.79 1.54 0.50 M1407 44 1.25 1.08 1.14 1.57 0.10 M1137 393 1.26** 1.30* 1.21 1.26* <.01 M1410 12 1.36 1.55 1.14 1.46 0.39 M1142 24 1.51 1.34 1.93 1.37 0.10 M1419 19 1.28 1.40 0.84 1.58 0.33 M1145 93 1.14 1.05 1.24 1.12 0.28 M1422 20 1.03 0.67 0.44 1.93* 0.39 M1150 28 2.29** 2.31* 3.59** 1.03 <.01 M1423 93 0.98 1.08 0.72 1.14 0.97 M1152 53 0.90 0.95 1.05 0.75 0.39 M1428 9 1.61 2.82 1.39 0.61 0.59 M1155 23 1.12 1.55 0.68 1.14 0.88 M1429 49 1.23 1.36 1.07 1.26 0.28 M1164 43 0.92 1.05 0.79 0.92 0.56 M1431 54 1.36* 1.48 1.42 1.19 0.11 M1173 35 0.93 1.01 1.19 0.57 0.45 M1432 15 1.07 0.46 1.38 1.34 0.51 M1174 229 1.10 1.13 1.09 1.08 0.36 M1433 22 1.22 1.04 1.60 1.01 0.45 M1176 46 1.19 1.17 1.42 0.97 0.43 M1436 173 1.20* 1.35* 1.16 1.09 0.17 M1183 206 1.03 1.15 1.12 0.82 0.65 M1438 79 1.29 1.25 1.77** 0.89 0.16 M1184 214 1.18 1.15 1.35* 1.04 0.13 M1439 28 1.28 1.62 1.49 0.77 0.60 M1187 184 1.15 1.00 1.42** 1.05 0.12 M1441 9 1.08 1.52 0.69 1.08 0.97 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal Ml 444 17 0.84 1.62 0.33 0.55 0.19 M1650 60 1.37* 1.50 1.13 1.50 0.05 M1448 13 1.02 1.89 1.24 0.18 0.40 M1659 10 1.13 1.07 0.96 1.37 0.66 M1450 31 0.98 0.99 1.29 0.63 0.71 M1662 29 1.02 0.97 0.99 1.09 0.85 M1456 20 1.31 2.23* 0.83 1.00 0.70 M1673 10 1.22 1.36 1.44 0.59 0.78 M1462 11 0.83 0.35 1.18 1.18 0.98 M1677 18 0.80 0.98 0.96 0.48 0.23 M1463 560 1.20** 1.28** 1.12 1.20 0.04 M1687 58 1.24 1.59 1.25 0.92 0.47 M1467 19 1.13 0.76 1.14 1.44 0.41 M1702 52 1.14 1.30 1.14 0.99 0.62 M1469 16 0.63 0.49 0.84 0.54 0.12 M1710 27 1.20 0.83 1.35 1.40 0.25 M1475 79 1.03 0.82 1.23 1.04 0.61 M1711 207 1.04 1.13 0.92 1.09 0.69 M1485 9 1.09 1.14 0.62 1.65 0.68 M1717 34 1.01 0.89 1.37 0.78 0.95 M1492 36 0.84 0.89 0.53 1.11 0.50 Ml 720 83 1.04 1.01 0.93 1.16 0.64 M1495 26 0.93 1.13 0.69 0.92 0.62 M1721 10 1.51 1.55 1.51 1.47 0.30 M1515 119 1.06 0.85 1.13 1.17 0.34 M1726 61 1.30 1.86** 0.98 1.00 0.42 M1517 25 0.86 0.90 0.91 0.75 0-44 M1737 124 1.26* 1.17 0.96 1.64** <.01 M1519 72 1.31* 1.49 1.25 1.18 0.13 Ml 743 18 1.03 1.14 0.48 1.48 0.75 M1525 470 1.22** 1.29** 1.12 1.24* 0.02 Ml 765 40 1.14 0.78 1.07 1.56 0.19 M1527 11 1.77 0.89 3.04* 1.51 0.08 Ml 766 83 1.16 1.16 1.13 1.19 0.26 M1528 133 0.99 0.89 1.08 0.97 0.99 M1772 127 1.15 1.10 1.37 0.99 0.28 M1529 182 1.19 1.38* 1.07 1.11 0.21 M1800 34 0.96 1.16 0.65 1.08 0.79 M1531 13 2.25* 2.78* 2.75 1.09 0.07 M1806 167 1.05 1.20 1.03 0.93 1.00 M1532 186 1.07 1.01 1.12 1.09 0.39 M1812 24 1.12 0.82 0.98 1.54 0.38 M1540 16 0.98 0.63 1.06 1.21 0.81 M1813 218 1.21* 1.28 1.02 1.35* 0.03 M1541 23 0.78 0.76 0.70 0.89 0.38 M1818 18 1.68 1.15 1.85 2.00 0.03 M1542 436 1.24** 1.35** 1.13 1.24* 0.01 M1821 63 1.02 0.78 1.23 1.01 0.72 M1545 23 1.20 1.31 0.98 1.30 0.48 M1832 11 0.70 0.79 0.56 0.75 0.29 M1548 9 1.01 0.66 0.93 1.51 0.70 M1833 30 1.24 1.27 1.21 1.24 0.35 M1566 22 1.04 0.79 1.25 1.14 0.70 M1839 43 0.96 0.99 0.64 1.26 0.95 M1569 249 1.17* 1.19 1.00 1.33* 0.04 M1842 18 0.80 0.98 0.96 0.48 0.23 M1577 270 1.13 1.16 1.02 1.20 0.13 M1844 42 1.40 2.03** 0.85 1.38 0.20 M1596 61 0.99 0.92 1.29 0.79 0.84 M1850 49 0.99 0.89 1.13 0.97 0.95 M1598 13 0.96 0.84 0.47 1.56 0.80 M1851 18 1.49 1.71 0.82 1.84 0.15 M1604 17 1.03 0.87 0.82 1.40 0.69 M1859 23 1.00 0.95 1.00 1.03 0.97 M1608 54 1.07 0.93 1.09 1.19 0.52 M1866 27 1.73* 1.26 1.26 2.68** <.01 M1609 114 1.05 1.16 1.01 1.00 0.85 M1872 29 1.13 0.83 1.41 1.15 0.43 M1633 17 0.66 0.61 0.48 0.92 0.23 M1874 9 0.83 0.38 2.74 .0.52 0.77 M1634 60 1.08 0.85 1.34 1.05 0.48 Ml 884 54 1.20 1.35 1.34 0.94 0.48 Ml 643 9 3.20** 1.08 3.48 5.20** <.01 M1910 14 0.74 1.01 0.38 0.96 0.34 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal M1913 62 0.88 0.96 0.93 0.75 0.28 M2130 19 1.16 0.54 1.67 1.32 0.32 M1915 123 1.03 1.24 0.98 0.88 0.76 M2131 18 1.06 1.25 1.56 0.37 0.75 M1920 387 1.22** 1.18 1.33** 1.16 0.01 M2135 143 1.01 1.38* 0.84 0.84 0.40 M1922 84 1.27 1.54* 1.17 1.07 0.23 M2138 11 1.13 2.09 0.62 0.39 0.63 M1931 47 1.17 1.65* 0.69 1.10 0.75 M2140 12 1.35 0.68 2.02 1.20 0.29 M1936 83 1.26 1.27 1.61* 0.90 0.21 M2141 19 1.34 1.60 0.84 1.58 0.30 M1937 111 1.37** 1.41 1.08 1.65** <.01 M2142 37 1.03 1.12 1.15 0.82 0.89 M1941 14 0.87 1.03 0.00 1.34 0.81 M2148 171 1.05 1.22 1.18 0.74 0.62 M1951 278 1.14 1.16 1.18 1.07 0.20 M2152 84 1.15 1.21 1.10 1.16 0.33 M1956 38 1.05 1.02 1.08 1.05 0.79 M2153 78 0.97 0.99 0.82 1.06 0.87 M1957 15 1.14 1.83 0.26 1.17 0.98 M2154 85 1.25 1.46 1.24 1.03 0.27 M1959 24 1.38 1.20 1.98 1.12 0.21 M2161 12 0.87 0.23 0.91 1.40 0.85 M1962 15 1.55 1.76 2.01 0.89 0.29 M2162 16 2.27** 4.18** 2.02 0.47 0.15 M1966 23 1.28 1.39 1.32 1.16 0.40 M2163 14 0.96 0.23 0.84 1.68 0.54 M1992 80 0.98 1.04 0.76 1.12 0.98 M2167 9 0.72 0.89 0.41 0.92 0.38 M1993 90 1.12 0.94 1.18 1.24 0.21 M2171 20 1.30 1.54 0.73 1.68 0.29 M2025 9 2.15* 2.60 3.50* 0.00 0.21 M2175 12 1.13 0.32 0.81 2.16 0.27 M2038 16 1.41 2.44* 0.60 1.02 0.68 M2181 78 1.11 0.99 1.05 1.29 0.26 M2040 12 1.13 0.51 0.71 2.16 0.26 M2187 19 0.86 0.50 0.69 1.43 0.93 M2073 13 0.75 0.72 0.56 0.95 0.46 M2193 9 0.85 0.55 0.76 1.32 0.95 M2074 48 0.96 1.22 0.75 0.92 0.58 M2194 128 1.26* 1.12 1.20 1.46* 0.01 M2075 11 0.99 0.54 0.57 1.81 0.58 M2196 33 0.74 0.84 0.96 0.44* 0.06 M2078 23 1.11 1.32 1.04 0.99 0.84 M2198 37 1.05 1.13 0.73 1.30 0.71 M2089 17 1.24 0.99 0.40 2.61** 0.14 M2201 28 1.12 0.62 0.81 1.93* 0.17 M2090 14 0.63 0.85 0.42 0.62 0.10 M2203 19 1.07 0.68 0.86 1.67 0.41 M2098 86 1.04 0.91 1.00 1.21 0.49 M2206 10 1.17 0.84 1.09 1.75 0.44 M2100 31 1.07 1.42 0.53 1.22 0.88 M2208 45 1.11 0.98 1.18 1.19 0.44 M2101 181 1.01 0.93 1.21 0.90 1.00 M2209 40 1.25 1.52 1.06 1.17 0.38 M2105 21 0.97 1.69 0.80 0.51 0.37 M2210 17 1.06 2.10* 0.39 0.69 0.53 M2109 104 0.99 0.95 1.04 0.97 0.93 M2211 14 0.92 1.58 0.42 0.61 0.38 M2111 22 1.39 1.67 1.21 1.34 .0.27 M2212 11 1.08 0.74 0.94 1.54 0.59 M2113 39 1.26 1.16 1.40 1.22 0.21 M2216 „12 1.13 0.51 0.71 2.16 0.26 M2125 14 1.32 1.34 1.15 1.46 0.37 M2218 15 0.90 1.06 1.88 0.00 0.29 M2126 20 1.20 1.29 0.73 1.56 0.42 M2221 55 1.06 1.17 1.03 0.98 0.88 M2127 15 1.30 1.76 1.06 0.94 0.68 M2223 114 1.06 1.01 0.92 1.23 0.40 M2128 30 1.00 1.20 0.61 1.16 0.96 M2225 13 0.88 0.68 1.02 0.92 0.79 M2129 71 1.13 1.27 1.25 0.86 0.71 M2226 15 1.05 0.79 1.57 0.84 0.85 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal M2227 14 1.13 0.76 1.66 0.94 0.65 M2553 100 1.11 1.38 1.19, 0.77 0.98 M2233 25 1.07 1.23 0.78 1.20 0.79 M2566 112 1.09 1.00 1.07 1.19 0.32 M2235 130 1.01 1.19 0.92 0.91 0.69 M2571 123 1.20 1.53** 0.96 1.12 • 0.30 M2238 50 1.20 1.85** 0.93 0.86 0.86 M2576 13 1.24 1.67 1.03 1.10 0.67 M2241 12 0.97 0.29 0.75 1.72 0.56 M2589 17 1.23 0.57 1.24 2.09 0.14 M2244 11 1.10 1.20 0.98 1.10 0.83 M2606 12 0.58 0.56 0.77 0.42 0.08 M2248 13 0.85 0.67 0.98 0.89 0.70 M2609 134 1.32** 1.18 1.57** 1.17 0.01 M2255 152 1.14 1.39* 1.16 0.89 0.68 M2629 14 1.37 0.92 2.35* 0.88 0.35 M2259 29 1.33 1.43 0.80 1.80 0.14 M2632 10 0.78 0.85 0.71 0.78 0.48 M2263 145 1.10 0.95 1.18 1.16 0.23 M2634 10 0.92 0.70 1.67 0.55 0.77 M2266 18 1.04 1.40 0.96 0.70 0.76 M2637 265 1.09 1.30* 0.91 1.05 0.72 M2268 25 1.18 1.68 0.77 1.16 0.70 M2648 52 0.96 0.99 1.17 0.78 0.61 M2271 11 1.35 0.38 1.45 2.16 0.14 M2651 22 0.98 0.54 1.19 1.20 0.72 M2274 79 1.04 1.04 1.13 0.95 0.88 M2653 11 1.00 1.16 0.50 1.39 0.90 M2278 135 1.20 1.22 1.22 1.16 0.12 M2673 9 0.78 0.72 0.26 1.39 0.76 M2280 405 1.16* 1.15 1.26* 1.08 0.09 M2677 19 0.86 1.06 0.94 0.56 0.37 M2284 13 1.52 2.65* 0.34 1.65 0.38 M2690 24 1.06 1.07 1.53 0.62 0.94 M2286 58 1.16 1.13 1.24 1.11 0.37 M2698 11 1.55 1.93 1.30 1.37 0.33 M2287 43 1.17 1.40 0.84 1.26 0.48 M2703 18 0.70 0.72 0.59 0.78 0.21 M2301 90 1.26 1.00 1.35 1.43 0.03 M2704 15 1.39 0.93 1.87 1.85 0.13 M2307 12 1.05 1.59 0.28 1.16 0.90 M2709 135 1.17 1.00 1.35 1.14 0.11 M2309 154 1.15 0.96 1.24 1.24 0.07 M2710 176 1.22* 1.39* 1.07 1.21 0.09 M2310 16 1.04 1.16 0.39 1.58 0.73 M2716 116 1.02 1.06 0.81 1.16 0.76 M2325 79 0.98 0.95 0.91 1.06 0.99 M2725 36 1.15 0.85 1.26 1.33 0.29 M2326 11 1.31 0.35 2.27 1.21 0.28 M2761 22 1.01 1.22 0.72 1.08 0.95 M2327 113 1.04 1.15 0.90 1.06 0.87 M2766 40 1.37 1.54 1.49 1.06 0.20 M2347 12 0.98 1.23 1.22 0.50 0.67 M2776 31 0.93 1.56 0.67 0.56 0.22 M2379 46 1.25 1.00 1.46 1.34 0.13 M2779 60 1.15 0.89 1.34 1.23 0.23 M2386 29 1.16 0.64 1.33 1.50 0.22 M2848 14 0.87 0.62 0.69 1.33 0.97 M2395 27 1.00 0.89 1.38 0.82 0.95 M2878 74 1.03 0.92 0.83 1.33 0.49 M2398 13 1.01 1.13 0.86 1.04 0.98 M2880 251 1.09 1.04 1.11 1.12 0.25 M2441 10 0.70 0.67 1.00 0.42 0.26 M2891 20 1.35 2.15* 0.66 1.09 0.63 M2452 25 2.21** 2.37** 2.41 1.79 <.01 M2894 152 1.19 1.19 1.18 1.19 0.10 M2498 14 1.16 1.29 0.27 1.93 0.46 M2900 87 1.31* 1.61* 1.14 1.16 0.14 M2509 115 1.08 1.06 1.05 1.13 0.44 M2954 40 1.20 0.88 1.69* 0.99 0.29 M2524 11 1.22 1.43 0.99 1.22 0.65 M2955 92 1.09 1.14 1.12 1.00 0.65 M2544 15 0.96 0.57 1.43 1.15 0.79 M2958 27 1.14 1.58 0.49 1.39 0.65 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal M2972 143 0.94 1.01 0.86 0.96 0.52 M3288 25 1.28 1.42 0.57 1.94* 0.18 M2975 14 1.05 1.51 1.14 0.47 0.70 M3289 39 1.36 1.03 1.80* 1.28 0.09 M2989 80 1.04 0.98 1.18 0.98 0.78 M3295 56 1.26 1.46 1.55 0.79 0.44 M2994 69 0.98 1.16 0.80 0.92 0.63 M3297 35 1.17 1.00 1.14 1.37 0.29 M2997 29 0.88 0.81 1.14 0.67 0.47 M3299 18 1.03 1.13 1.41 0.60 0.80 M3004 14 1.12 0.46 1.13 1.88 0.30 M3304 194 1.19 1.48** 1.09 0.99 0.38 M3019 12 0.97- 0.65 1.00 1.33 0.79 M3305 25 1.16 0.70 1.49 1.28 0.33 M3032 22 1.25 1.65 1.31 0.65 0.75 M3308 16 0.89 0.83 0.35 1.45 0.98 M3033 77 1.09 0.84 1.28 1.16 0.32 M3309 22 1.02 0.70 1.76 0.58 1.00 M3050 12 1.16 1.12 1.29 1.09 0.68 M3374 27 0.99 1.10 1.01 0.85 0.81 M3057 73 1.19 1.70** 0.91 1.00 0.64 M3389 19 1.05 0.89 0.85 1.46 0.62 M3058 49 1.08 0.96 1.23 1.05 0.60 M3390 86 1.27 1.26 1.25 1.30 0.07 M3091 44 1.22 1.42 0.93 1.37 0.29 . M3397 52 0.96 0.99 1.17 0.78 0.61 M3095 76 1.10 0.96 1.34 1.00 0.48 M3405 20 1.19 0.51 1.66 1.46 0.24 M3098 14 0.85 1.51 0.72 0.35 0.23 M3417 78 0.87 1.12 0.99 0.51* 0.06 M3107 54 0.89 1.00 0.93 0.73 0.29 M3821 250 1.19* 1.23 1.15 1.20 0.05 M3114 10 1.18 1.46 1.51 0.42 0.99 M3823 64 1.02 0.99 1.03 1.05 0.84 M3116 11 1.51 0.86 1.42 2.31 0.11 M3829 48 1.21 1.49 1.18 0.93 0.59 M3118 219 1.10 0.90 1.37* 1.05 0.17 M3832 82 1.33* 1.13 1.34 1.51* 0.01 M3125 31 1.03 0.96 1.62 0.46 0.84 M3835 68 1.19 1.56* 0.93 1.00 0.61 M3135 9 1.06 0.34 1.59 1.25 0.60 M3841 118 1.08 1.35 0.75 1.14 0.72 M3162 12 1.13 0.51 0.71 2.16 0.26 M3846 144 1.07 1.47** 1.03 0.74 0.52 M3187 140 1.12 0.95 1.19 1.22 0.13 M3847 18 0.72 1.45 0.47 0.32 0.04 M3190 64 1.30 1.52 1.39 0.97 0.26 M3848 12 1.17 0.65 2.70* 0.36 0.64 M3191 507 1.25** 1.22* 1.28* 1.25* <.01 M3849 27 0.99 1.14 0.68 1.12 0.94 M3192 139 1.15 1.38* 1.03 1.03 0.51 M3850 26 1.06 1.20 0.94 1.06 0.87 M3195 12 1.05 1.59 0.28 1.16 0.90 M3855 60 1.08 1.09 0.95 1.20 0.55 M3207 10 0.53 0.71 0.42 0.49 0.06 M3856 57 1.22 1.53 1.01 1.11 0.42 M3210 17 1.13 1.71 1.12 0.59 0.87 M3860 34 1.29 0.69 1.34 1.82* 0.05 M3213 21 0.90 0.95 0.72 1.04 0.72 M3862 16 1.41 1.41 0.86 1.72 0.21 M3220 158 1.14 1.11 0.96 1.34* 0.10 M3864 46 1.04 1.09 0.93 1.09 0.82 M3225 46 1.03 0.61 1.11 1.42 0.33 M3866 31 1.04 1.07 0.76 1.29 0.73 M3232 12 0.88 0.76 0.43 1.44 0.98 M3873 189 1.16 1.13 1.04 1.29 0.07 M3249 58 1.20 1.64* 0.83 1.17 0.49 M3881 210 1.18* 1.24 1.05 1.27 0.06 M3252 63 0.91 0.99 0.96 0.80 0.40 M3886 102 1.16 0.96 1.59** 0.95 0.24 M3258 17 0.86 0.80 1.09 0.69 0.53 M3891 15 1.30 2.62* 1.07 0.24 0.75 M3275 19 1.09 0.82 1.42 1.18 0.57 M3895 164 1.23* 1.29 1.16 1.23 0.06 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal M3898 217 1.16 1.05 1.18 1.24 0.05 M4242 152 1.02 1.22 0.98 0.89 0.70 M3899 184 1.25* 1.45** 1.19 1.13 0.08 M4250 95 1.17 1.55* 0.92 1.03 0.59 M3901 34 0.99 0.97 1.67 0.46 0.60 M4258 9 1.27 2.48 1.15 0.00 0.76 M3914 9 0.84 0.97 1.19 0.46 0.47 M4279 31 1.33 1.72 1.06 1.16 0.37 M3939 18 1.05 2.01 0.81 0.52 0.48 M4292 130 1.17 0.99 1.33 1.17 0.10 M3981 333 1.11 1.18 1.12 1.03 0.40 M4308 67 0.99 1.16 0.79 0.93 0.66 M3983 9 1.31 2.62 1.16 0.38 0.94 M4318 342 1.19* 1.13 1.25* 1.20 0.02 M3984 17 1.08 1.27 0.84 1.14 0.87 M4320 16 1.14 1.46 0.84 1.11 0.81 M3985 17 1.10 1.71 0.65 1.04 0.98 M4331 266 1.22* 1.26 1.23 1.18 0.03 M3986 17 1.06 1.32 0.38 1.50 0.78 M4338 17 1.08 0.75 1.22 1.29 0.58 M4011 33 0.75 0.92 0.70 0.63 0.09 M4353 42 1.12 0.70 1.20 1.47 0.21 M4016 379 1.19* 1.25* 1.22 1.12 0.06 M4358 16 1.30 0.95 0.86 2.22 0.16 M4022 90 1.06 0.97 1.01 1.20 0.45 M4370 14 0.87 1.03 0.00 1.34 0.81 M4037 15 1.06 1.45 0.77 0.99 0.96 M4376 254 1.09 1.01 0.97 1.27* 0.12 M4039 105 1.23 1.16 1.41 1.13 0.10 M4385 20 1.00 1.28 0.86 0.89 0.81 M4051 9 1.50 2.50 1.45 0.52 0.72 M4392 95 0.96 1.13 0.89 0.86 0.46 M4052 11 2.72** 2.57 3.49* 2.09 0.02 M4401 68 0.89 0.98 1.07 0.64 0.21 M4056 66 1.13 0.97 1.15 1.28 0.25 M4404 186 1.05 1.11 0.96 1.08 0.66 M4058 146 1.22* 1.21 1.11 1.32 0.05 M4408 12 1.13 0.51 0.71 2.16 0.26 M4063 200 1.06 1.14 1.10 0.95 0.82 M4410 11 0.80 0.85 1.17 0.42 0.38 M4077 9 0.83 0.59 1.45 0.50 0.58 M4438 29 1.55* 2.79** 0.98 1.09 0.30 M4088 16 1.03 0.52 1.38 1.26 0.59 M4444 9 1.62 2.58 1.54 0.62 0.56 M4090 22 1.18 0.94 0.95 1.60 0.31 M4448 58 1.14 1.48 0.86 1.10 0.66 M4097 34 1.11 0.87 1.19 1.28 0.40 M4487 14 1.32 1.57 1.38 0.95 0.56 M4102 12 0.88 0.74 1.09 0.81 0.75 M4491 12 0.95 0.81 1.00 1.03 0.96 M4115 15 1.08 0.90 0.82 1.61 0.55 M4513 11 1.14 0.95 1.33 1.15 0.66 M4117 30 1.04 0.65 1.01 1.46 0.43 M4514 19 1.26 0.39 1.53 1.96 0.11 M4132 27 1.11 1.08 1.68 0.53 0.93 M4517 14 1.00 1.27 0.72 1.04 0.90 M4134 10 0.73 0.70 0.22 1.24 0.58 M4554 23 1.03 0.82 1.53 0.75 0.97 M4188 83 1.12 0.97 1.10 1.30 0.22 M4558 16 0.99 1.23 0.76 0.96 0.83 M4210 9 0.52 0.54 0.92 0.15 0.05 M4575 16 1.10 2.01 0.83 0.55 0.64 M4214 57 1.06 1.35 0.85 0.95 0.92 M4598 22 1.73* 3.19** 0.49 1.62 0.15 M4215 11 1.17 0.90 1.02 1.63 0.45 M4618 9 0.80 0.35 2.74 0.52 0.72 M4219 17 0.98 0.73 1.06 1.22 0.79 M4624 46 1.06 0.79 1.49 0.95 0.63 M4220 14 0.96 0.65 0.93 1.32 0.82 M4636 11 0.83 0.90 0.69 0.90 0.60 M4232 10 0.84 0.83 0.41 1.46 0.85 M4649 42 1.45* 1.58 1.77* 1.03 0.13 M4235 20 1.24 0.94 1.80 0.94 0.42 M4663 9 0.72 0.31 2.61 0.47 0.55 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal M4673 18 1.14 0.57 1.51 1.36 0.37 P2001 641 1.24** 1.21* 1.24* 1.27* <.01 M4692 29 1.28 0.78 0.84 2.17** 0.05 P2002 700 1.19* 1.14 1.20 1.23* 0.02 M4718 22 1.18 1.25 0.45 1.91 0.32 P2003 727 1.20* 1.13 1.10 1.37** <.01 M4719 173 1.30** 1.00 1.40* 1.48** <.01 P2004 693 1.17* 1.15 1.13 1.22* 0.04 M4743 38 1.03 1.23 1.15 0.75 0.79 P2005 735 1.19* 1.15 1.19 1.24* 0.02 M4755 53 1.27 1.50 1.25 1.07 0.29 P2006 643 1.15* 1.18 1.21* 1.07 0.22 M4756 12 1.58 2.26 1.86 0.72 0.45 P2007 332 1.04 1.02 1.03 1.07 0.52 M4839 106 1.05 0.96 1.10 1.07 0.60 P2008 420 1.07 1.22* 0.94 1.04 0.81 M4884 40 1.19 1.05 1.81* 0.84 0.49 P2009 223 1.03 0.95 1.10 1.03 0.64 M4891 11 1.47 1.51 1.16 1.77 0.26 P2011 100 0.96 1.34 0.77 0.81 0.27 M4896 11 0.52* 1.04 0.30 0.25 0.01 P2013 56 0.94 1.01 1.07 0.78 0.52 M4897 222 1.06 1.11 0.96 1.12 0.49 S0001 150 1.17 1.41* 0.86 1.23 0.23 M4905 39 1.10 0.71 1.28 1.29 0.34 S0002 35 1.28 1.21 1.71 0.94 0.30 M4946 11 0.98 1.05 1.35 0.54 0.77 S0005 102 1.11 1.00 0.98 1.34 0.19 M4982 177 1.06 1.22 0.97 0.97 0.95 S0006 9 0.82 1.21 0.26 1.05 0.57 M4987 24 1.23 0.80 1.29 1.62 0.19 S0008 86 1.20 1.09 1.38 1.14 0.15 M4999 52 0.96 0.99 1.17 0.78 0.61 S0009 40 1.41 1.58 1.06 1.58 0.08 M5090 64 0.95 1.00 0.96 0.90 0.66 S0012 46 1.28 1.63 1.29 0.93 0.45 M5222 282 1.16 1.20 1.11 1.16 0.11 S0019 136 1.19 1.09 1.20 1.27 0.06 P0120 14 0.81 0.32 0.89 1.25 0.92 S0020 44 1.28 1.13 1.57 1.14 0.17 P0310 148 1.10 1.12 1.19 0.99 0.49 S0022 68 1.02 0.94 0.92 1.19 0.68 P0410 102 1.11 0.91 1.27 1.15 0.25 S0024 66 1.33* 1.70* 1.38 0.95 0.26 P0412 18 0.57* 0.49 0.98 0.27* 0.02 S0026 37 1.21 1.46 1.10 1.05 0.53 P0418 36 1.17 0.69 1.29 1.55 0.16 S0028 81 1.08 0.95 1.06 1.25 0.34 P0420 25 0.83 1.04 0.76 0.69 0.27 S0030 71 1.35* 1.70* 1.31 1.05 0.17 P0430 141 1.04 1.04 1.01 1.07 0.68 S0037 66 1.01 0.97 0.98 1.07 0.85 P0431 79 1.01 1.04 0.93 1.04 0.98 S0042 13 1.14 1.69 1.36 0.48 0.86 P0432 16 0.75 0.12* 0.74 1.54 1.00 S0045 13 1.14 1.69 1.36 0.48 0.86 P0450 56 0.83 0.68 1.13 0.65 0.23 S0049 22 1.05 1.39 0.43 1.28 0.96 P0610 705 1.18* 1.19 1.07 1.28** 0.02 S0050 239 1.27** 1.21 1.42** 1.19 <.01 P0620 545 1.30** 1.18 1.36** 1.34** <.01 S0051 136 1.11 1.19 0.99 1.13 0.40 P0640 81 1.11 1.54* 0.59 1.21 0.76 S0056 14 1.14 1.30 0.79 1.28 0.68 P0651 630 1.15* 1.12 1.13 1.21* 0.03 S0057 31 1.27 1.66 1.06 1.04 0.52 P0652 601 1.18* 1.16 1.14 1.24* 0.02 S0058 12 0.98 1.26 1.03 0.68 0.72 P0710 493 1.12 1.14 1.01 1.19 0.11 S1030 16 1.33 0.95 1.33 1.81 0.19 P0720 163 1.05 0.86 1.20 1.07 0.40 S1031 10 1.38 0.90 2.01 1.53 0.26 P2000 718 1.19* 1.24* 1.18 1.15 0.11 S2002 111 1.18 1.22 1.19 1.11 0.23 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal S2003 229 1.19* 1.00 1.42** 1.14 0.03 S2124 29 1.17 1.81 1.11 0.67 0.93 S2004 153 1.13 1.02 0.98 1.39* 0.07 S2126 15 1.29 0.73 1.38 1.85 0.20 S2009 22 0.89 1.04 0.56 1.02 0.64 S2129 9 0.86 1.17 0.57 0.84 0.58 S2010 97 0.99 0.92 1.03 1.02 0.94 S2133 32 1.31 1.38 1.11 1.48 0.19 S2031 25 1.39 0.83 1.81 1.69 0.07 S2134 146 1.07 0.94 1.27 0.98 0.52 S2035 17 1.22 1.07 1.13 1.53 0.37 S2139 92 1.21 1.43 1.18 1.02 0.35 S2042 42 1.28 1.94** 0.94 0.88 0.68 S2140 19 1.18 1.20 0.68 1.58 0.43 S2044 90 1.22 1.17 1.33 1.16 0.13 S2156 39 1.12 0.85 1.32 1.28 0.33 S2045 107 1.12 1.09 0.90 1.34 0.21 S2167 28 1.23 1.73 0.65 1.32 0.54 S2046 9 1.81 2.90* 0.00 • 1.86 0.30 S2184 58 1.18 1.37 1.05 1.11 0.44 S2049 18 1.29 1.38 1.54 0.92 0.50 S2186 44 1.21 1.17 1.40 1.04 0.36 S2059 9 1.13 0.85 2.01 0.79 0.76 S2199 58 0.98 0.93 1.12 0.90 0.87 S2063 31 1.33 1.72 1.06 1.16 0.37 S2205 58 1.30 1.66* 1.06 1.18 0.24 S2065 60 1.11 0.81 1.37 1.15 0.32 S2206 88 1.21 1.48* 1.03 1.08 0.36 S2068 270 1.20* 1.10 1.24 1.27* 0.01 S2207 67 1.22 1.29 1.39 0.96 0.33 S2069 78 1.21 1.28 1.19 1.16 0.22 S2209 214 1.20* 1.18 1.26 1.17 0.05 S2075 9 1.26 2.07 0.35 1.68 0.62 S2210 81 1.32* 1.46 1.46 1.03 0.12 S2077 204 1.13 1.15 1.16 1.09 0.24 S2213 12 1.34 1.34 0.85 2.20 0.28 S2080 19 1.21 1.07 1.00 1.58 0.34 S2215 35 1.26 1.64 0.77 1.34 0.38 S2084 58 1.04 1.53 0.61 0.99 0.75 S2226 12 1.34 1.34 0.85 2.20 0.28 S2085 21 0.89 1.06 0.76 0.86 0.56 S2227 12 1.19 1.26 0.72 1.99 0.47 S2088 18 1.12 1.74 1.10 0.62 0.80 S2228 12 1.34 1.34 0.85 2.20 0.28 S2090 24 1.00 0.78 1.13 1.12 0.81 S2257 119 1.04 1.14 0.91 1.06 0.85 S2091 14 0.85 0.92 0.95 0.70 0.52 S2258 29 1.45 1.50 1.37 1.51 0.10 S2092 218 1.11 1.11 1.10 1.11 0.27 S2259 86 1.00 0.67 1.25 1.07 0.58 S2093 17 1.05 1.03 0.91 1.19 0.81 S2260 135 1.11 0.94 1.17 1.20 0.18 S2094 58 1.24 1.30 1.33 1.06 0.28 S2261 35 1.10 1.33 1.04 0.91 0.92 S2095 426 1.21** 1.26* 1.14 1.24* 0.01 S2262 12 1.34 1.34 0.85 2.20 0.28 S2099 78 1.25 1.15 1.15 1.46 0.06 S2263 83 0.94 0.88 0.98 0.97 0.75 S2100 14 0.70 1.28 0.14 0.77 0.14 S2265 32 0.99 0.82 1.05 1.10 0.86 S2101 269 1.16 1.14 1.14 1.20 0.06 S2269 16 1.34 1.01 1.23 2.01 0.16 S2105 29 1.26 1.73 0.89 1.10 0.58 S2271 15 1.22 0.99 1.38 1.40 0.39 S2106 32 1.24 1.70 0.90 1.08 0.60 S2312 64 1.03 1.29 1.00 0.79 0.69 S2113 64 0.93 0.96 1.01 0.81 0.49 S2315 17 1.07 1.09 1.43 0.65 0.99 S2114 56 0.98 0.97 1.24 0.82 0.74 S2316 25 1.07 1.01 1.29 0.85 0.86 S2120 93 1.18 1.53* 0.82 1.20 0.40 S2317 14 1.38 2.09 0.64 1.70 0.44 S2123 55 1.23 1.29 1.15 1.24 0.23 S2318 17 1.12 1.99 0.45 0.89 0.83 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal S2323 15 1.10 1.21 0.64 1.31 0.75 S2475 45 1.29 1.67* 1.24 0.95 0.44 S2325 22 1.25 0.67 1.57 1.49 0.18 S2479 12 1.05 2.01 0.53 0.71 0.65 S2326 106 1.14 1.03 1.09 1.31 0.14 S2482 68 0.99 0.95 0.75 1.25 0.77 S2393 169 1.13 1.09 1.16 1.12 0.22 S2483 37 1.03 1.31 0.68 1.08 0.95 S2394 54 1.33 1.46 1.25 1.26 0.13 S2484 58 1.39* 1.23 1.55 1.41 0.03 S2395 209 1.20* 1.03 1.33* 1.24 0.02 S2487 52 1.28 1.69* 1.22 0.90 0.47 S2396 131 1.09 0.98 0.97 1.32 0.20 S2489 234 1.12 1.12 1.19 1.06 0.25 S2397 50 1.25 1.51 1.07 1.16 0.34 S2490 45 1.43* 1.54 1.74* 1.00 0.12 S2399 31 1.18 1.60 0.87 1.04 0.75 S2494 9 1.70 1.28 0.48 3.61* 0.06 S2400 12 0.90 0.91 1.08 0.70 0.68 S2499 9 1.70 1.28 0.48 3.61* 0.06 S2401 21 0.79 1.03 0.63 0.72 0.25 S2501 29 1.19 1.73 0.83 1.00 0.79 S2404 163 1.10 1.15 1.11 1.04 0.46 S2503 17 1.12 1.15 0.69 1.76 0.52 S2405 31 1.33 1.72 1.06 1.16 0.37 S2507 114 1.01 1.02 0.99 1.01 0.97 S2410 12 0.69 0.75 0.94 0.44 0.19 S2511 56 0.98 1.29 0.95 0.72 0.46 S2414 86 1.08 0.94 0.95 1.35 0.27 S2515 14 0.86 0.48 0.76 1.46 0.92 S2421 57 1.15 1.40 1.14 0.87 0.76 S2517 9 0.84 0.56 1.01 0.97 0.81 S2425 12 0.89 1.04 0.58 1.10 0.73 S2524 63 1.23 1.43 1.05 1.20 . 0.28 S2427 35 1.00 1.49 0.98 0.57 0.44 S2531 38 1.15 1.58 1.00 0.83 0.92 S2430 24 1.16 1.30 1.41 0.74 0.79 S2532 17 1.32 1.61 1.27 1.12 0.47 S2431 322 1.15 1.19 1.01 1.25* 0.06 S2540 115 1.02 0.94 1.13 0.99 0.80 S2434 9 1.70 1.28 0.48 3.61* 0.06 S2541 12 1.28 1.63 0.64 2.20 0.41 S2436 33 1.27 1.71 0.98 1.08 0.51 S2544 66 1.01 0.97 0.98 1.07 0.85 S2437 45 1.43* 1.54 1.74* 1.00 0.12 S2545 109 1.15 1.19 0.96 1.30 0.19 S2438 75 1.06 1.16 1.02 0.99 0.85 S2547 19 0.91 0.63 0.84 " 1.23 0.97 S2439 113 0.98 0.92 1.10 0.91 0.82 S2552 31 1.33 1.72 1.06 1.16 0.37 S2442 16 1.50 0.62 1.36 2.38* 0.04 S2555 42 1.00 0.63 1.04 1.31 0.53 S2443 9 1.70 1.28 0.48 3.61* 0.06 S2558 224 1.11 0.97 1.32* 1.06 0.18 S2447 14 1.06 1.10 0.65 1.44 0.74 S2561 33 1.27 1.71 0.98 1.08 0.51 S2453 112 1.27* 1.47* 0.94 1.40 0.05 S2569 31 1.32 1.70 . 1.10 1.10 0.40 S2459 69 1.15 1.22 1.20 1.00 0.49 S2581 120 1.28* 1.24 1.36 1.24 0.04 S2460 95 0.96 0.93 0.97 0.97 0.78 S2582 18 0.91 0.65 0.92 1.11 0.94 S2465 59 1.16 1.33 1.38 0.72 0.76 S2583 116 1.28* 1.31 1.52* 1.00 0.10 S2467 71 1.16 1.52* 1.12. 0.84 0.81 S2584 62 1.59** 1.75* 1.52 1.50 <.01 S2471 339 1.19* 1.22 1.18 1.17 0.04 S2586 44 1.32 1.07 1.60 1.40 0.08 S2472 14 1.80 1.54 1.26 2.51* 0.04 S2591 12 1.34 1.34 0.85 2.20 0.28 S2473 33 1.27 1.71 0.98 1.08 0.51 S2594 249 1.17* 1.09 1.31* 1.12 0.06 S2474 45 1.43* 1.54 1.74* 1.00 0.12 S2596 12 1.34 1.34 0.85 2.20 0.28 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal S2597 99 1.07 1.13 1.00 1.06 0.69 T0481 42 1.39 1.07 1.43 1.69 0.03 S2598 278 1.19* 1.11 1.27* 1.18 0.03 T0482 10 0.86 0.83 0.57 1.14 0.82 S2599 221 1.37** 1.45** 1.37* 1.28 <.01 T0508 92 1.37** 1.70** 1.21 1.16 0.08 S2601 13 1.29 1.26 0.77 2.14 0.29 T0532 12 1.50 0.85 1.37 2.22 0.11 S2603 41 1.02 0.85 1.38 0.83 0.95 T0535 89 1.04 1.10 1.03 1.01 0.84 S2604 13 0.75 0.61 0.66 0.96 0.50 T0550 104 1.17 1.59** 0.88 1.03 0.63 S2627 257 1.14 1.20 1.14 1.09 0.20 T0624 147 1.22* 1.53** 0.97 1.17 0.19 S2630 56 0.98 0.99 1.17 0.86 0.79 T0641 16 0.81 0.82 0.93 0.65 0.42 T0016 31 1.08 0.90 1.42 0.95 0.70 T0670 30 0.80 0.98 0.76 0.66 0.18 T0021 50 1.14 1.05 1.51 0.88 0.55 T0763 18 1.95* 1.42 2.58* 1.90 0.02 T0027 56 1.22 1.16 1.24 1.26 0.19 TO 79 5 23 1.36 2.53** 0.81 0.75 0.84 TOO 38 70 1.33* 1.31 1.36 1.32 0.06 T0798 10 0.91 0.62 2.19 0.51 0.79 T0052 11 1.87 0.48 3.50** 1.67 0.04 T0819 19 1.20 1.34 1.35 0.93 0.64 T0055 9 1.45 1.67 1.22 1.49 0.39 T0837 11 0.81 1.03 0.54 0.79 O.46 T0059 32 1.20 1.77* 0.82 1.04 0.74 T0861 18 1.49 1.89 1.78 0.98 0.35 T0062 46 1.31 1.39 L20 1.35 0.15 T0890 9 1.03 1.50 0.00 1.48 0.99 T0084 91 1.30* 1.17 1.50* 1.24 0.04 T0892 276 1.01 0.90 1.15 0.99 0.71 T0118 142 1.09 0.97 1.11 1.20 0.23 T0902 12 0.56 0.14 1.18 0.40 0.12 T0166 23 1.14 0.69 1.20 1.47 0.33 T0962 225 1.16 1.30* 1.09 1.11 0.19 T0176 299 1.12 1.18 1.15 1.03 0.34 T0981 69 1.27 1.47 0.98 1.31 0.16 TO 180 590 1.25** 1.32** 1.25* 1.17 0.02 T0995 9 0.94 0.25 2.13 0.88 0.85 TO 183 16 1.23 1.39 1.40 0.91 0.66 T1017 16 1.03 1.47 0.70 0.94 0.86 T0189 17 1.09 1.36 0.91 1.01 0.91 T1055 14 1.07 0.54 1.34 1.24 0.60 T0202 9 0.94 0.25 2.13 0.88 0.85 T1063 130 1.17 0.89 1.51** 1.11 0.09 T0204 61 1.23 1.23 1.43 1.05 0.26 T1074 12 1.05 1.26 0.25 1.68 0.75 T0245 34 1.27 1.62 1.04 1.11 0.45 T1076 130 1.14 1.08 1.29 1.04 0.27 T0262 148 1.30** 1.29 1.29 1.31 0.02 T1120 12 0.80 1.86 0.34 0.41 0.16 T0263 18 0.94 0.66 1.38 0.69 0.84 T1124 27 1.05 1.11 0.69 1.32 0.72 T0265 17 1.24 1.41 1.09 1.16 0.56 T1150 25 1.12 0.69 1.15 1.52 0.32 T0269 240 1.18* 1.24 1.05 1.25 0.06 T1153 269 1.20* 1.15 1.08 1.35* 0.01 T0345 125 1.09 0.90 1.40* 1.00 0.35 T1155 191 1.13 1.32* 1.12 0.96 0.55 T0362 45 0.96 0.53 1.34 0.98 0.84 T1185 364 1.28** 1.29* 1.25* 1.30* <.01 T0375 10 1.06 0.93 1.12 1.13 0.82 T1186 209 1.14 1.29* 1.32* 0.82 0.64 T0379 34 1.25 0.81 1.57 1.33 0.16 T1187 641 1.22** 1.26* 1.13 1.27** 0.01 T0420 47 1.29 1.13 1.39 1.37 0.10 T1188 569 1.23** 1.16 1.21 1.32** <.01 T0430 21 0.95 0.43 1.39 1.00 0.86 T1192 19 1.07 1.06 0.68 1.63 0.60 T0453 13 1.02 1.04 1.51 0.50 0.83 T1194 190 1.13 1.09 1.11 1.17 0.17 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal T1214 43 1.24 1.26 1.29 1.18 0.27 T1624 66 1.04 1.38 0.86 0.86 0.72 T1269 503 1.18* 1.38** 1.00 1.16 0.16 T1628 290 1.12 1.18 1.04 1.13 0.23 T1270 482 1.22** 1.27* 1.16 1.22* 0.02 T1649 10 1.02 0.00 1.61 1.56 0.46 T1271 411 1.13 1.21 1.12 1.06 0.29 T1650 257 1.09 1.19 1.16 0.93 0.68 T1272 430 1.16* 1.15 1.14 1.19 0.04 T1651 247 1.10 1.03 1.15 1.12 0.20 T1274 12 1.07 0.83 1.36 1.03 0.77 T1652 30 1.41 0.66 1.99* 1.67 0.03 T1293 10 0.73 0.42 1.27 0.48 0.41 T1676 97 1.06 1.06 1.10 1.01 0.71 T1307 29 0.88 0.80 1.07 0.76 0.55 T1706 14 1.22 1.33 1.16 1.18 0.58 T1341 125 1.10 1.37 1.03 0.90 0.93 T1720 44 1.09 1.00 1.19 1.09 0.58 T1364 9 1.40 0.82 2.98* 0.51 0.43 T1722 34 0.86 1.21 0.44 0.98 0.36 T1366 27 1.11 1.57 1.14 0.64 0.86 T1734 37 1.11 0.98 1.07 1.26 0.46 T1378 81 1.14 1.25 1.18 1.00 0.50 T1764 408 1.19* 1.23* 1.19 1.15 0.04 T1379 10 0.90 1.42 0.00 1.21 0.72 T1768 552 1.17* 1.21* 1.13 1.17 0.06 T1460 141 1.11 1.12 0.92 1.28 0.23 T1792 15 1.22 0.75 1.68 1.28 0.37 T1473 9 0.73 0.71 0.84 0.65 0.39 T1799 194 1.09 1.01 1.13 1.13 0.25 T1474 165 0.97 0.90 1.17 0.85 0.68 T1816 11 1.62 1.91 0.46 2.50 0.15 T1475 535 1.24** 1.18 1.09 1.45** <.01 T1833 216 1.13 1.22 1.10 1.06 0.32 T1486 42 1.25 1.18 1.28 1.29 0.21 T1854 151 1.18 1.42* 1.25 0.85 0.55 T1488 21 1.21 1.03 1.66 0.92 0.49 T1857 183 1.09 0.99 0.95 1.34* 0.13 T1492 65 1.32* 1.37 1.49 1.12 0.11 T1867 102 0.94 1.03 1.03 0.77 0.35 T1493 65 1.30 1.27 1.53 1.11 0.12 T1870 227 1.14 1.05 1.30* 1.07 0.14 T1500 73 1.11 0.90 1.52* 0.92 0.46 T1872 104 1.07 1.01 1.13 1.09 0.49 T1505 9 1.27 0.83 0.87 2.09 0.30 T1873 14 1.08 1.61 1.01 0.64 0.78 T1516 11 2.38* 2.42 3.25* 1.43 0.04 T1876 134 1.01 0.97 0.87 1.21 0.59 T1523 69 0.89 1.07 0.85 0.74 0.20 T1880 230 1.13 1.12 1.06 1.21 0.11 T1525 36 1.03 0.81 1.34 0.94 0.79 T1887 295 1.06 0.97 1.14 1.07 0.32 T1531 73 1.32* 1.32 1.45 1.18 0.08 T1890 21 0.86 1.47 0.39 0.86 0.32 T1542 495 1.13 1.16 1.10 1.12 0.15 T1891 21 0.84 1.22 0.45 0.87 0.35 T1554 170 0.99 0.92 1.32* 0.75 0.68 T1892 10 0.72 0.69 1.07 0.41 0.29 T1557 423 1.10 1.12 1.13 1.04 0.37 T1909 80 1.63** 1.17 1.76** 2.01** <.01 T1558 171 1.21* 1.46** 1.17 1.01 0.26 T1912 229 1.22* 1.24 1.23 1.18 0.04 T1575 253 1.18* 1.21 1.08 1.24 0.05 T1941 23 0.98 1.31 0.67 1.02 0.79 T1577 47 1.10 1.03 1.28 0.97 0.63 T1947 13 1.18 0.61 0.87 1.92 0.29 T1583 73 1.35* 1.47 1.06 1.50 0.04 T1949 17 0.76 0.87 0.15 1.18 0.47 T1585 101 1.01 1.02 1.02 0.99 0.99 T1956 304 1.00 1.05 1.01 0.94 0.73 T1587 30 1.10 1.19 0.72 1.40 0.56 T1966 16 0.92 0.77 1.19 0.78 0.76 T1595 23 0.89 0.71 0.68 1.29 0.96 T1998 16 1.30 0.53 1.14 2.26* 0.10 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal T2003 16 1.03 2.49** 0.19 0.52 0.32 X1080 177 1.19 0.97 1.51** 1.08 0.06 T2051 65 1.06 0.96 0.79 1.40 0.39 X1100 12 1.40 0.35 2.02 1.83 0.13 T2060 18 1.04 0.57 1.08 1.41 0.54 X1102 12 1.44 0.35 1.30 2.77* 0.06 T2072 18 1.03 0.99 0.84 1.23 0.81 X1103 18 1.19 0.69 1.30 1.70 0.24 T2076 109 1.10 0.97 0.99 1.34 0.19 X1105 56 1.32 1.13 1.52 1.28 0.08 T2080 50 1.20 1.03 1.32 1.31 0.19 X1106 20 0.99 0.65 1.62 0.74 0.98 T2086 25 1.02 0.75 1.46 0.84 0.89 X1107 14 0.82 0.00 0.98 1.40 0.88 T2096 .13 0.87 0.77 1.24 0.60 0.59 X1108 266 1.19* 1.20 1.24 1.14 0.06 T2097 19 1.01 0.34 1.34 1.29 0.56 X1109 20 1.26 2.09* 1.29 0.34 0.85 T2099 9 0.93 1.26 0.00 1.40 0.86 X1112 67 0.91 0.97 0.97 0.81 0.41 X0001 61 0.98 0.94 0.99 1.02 0.99 X1114 9 0.86 0.47 1.55 0.66 0.81 X0029 55 1.05 1.20 0.99 0.98 0.93 X1118 183 1.10 1.19 0.89 1.22 0.28 X0063 57 1.15 1.71* 0.90 0.82 0.96 X1120 9 0.82 0.89 1.09 0.50 0.49 X0074 192 1.04 0.87 1.49** 0.77 0.86 X1121 12 1.13 0.51 0.71 2.16 0.26 X0089 46 1.04 1.03 1.18 0.90 0.94 X1133 43 1.25 0.90 1.82* 1.27 0.11 X0093 78 1.14 1.25 1.19 0.98 0.53 X1134 28 1.25 0.79 1.70 1.27 0.19 X0105 9 1.66 1.13 2.42 1.51 0.18 X1135 14 0.73 0.75 0.17 1.24 0.49 X0108 22 0.99 0.81 1.43 0.88 0.97 X1139 10 0.67 0.59 0.38 1.07 0.41 X0145 64 0.92 0.97 0.96 0.84 0.47 XI140 47 1.08 1.22 0.86 1.15 0.71 X0150 11 0.89 0.51 1.35 0.75 0.82 X1141 25 1.22 0.83 1.08 1.72 0.17 X0158 91 1.25 1.34 1.00 1.40 0.08 X1146 131 0.99 1.01 0.83 1.13 0.91 X0167 14 0.76 0.91 1.03 0.41 0.21 X1156 33 1.22 1.96* 0.71 1.04 0.76 X0182 45 1.09 0.93 1.17 1.18 0.49 X1162 15 1.27 1.64 0.25 1.94 0.39 XI009 53 1.38* 1.25 1.34 1.54 0.03 X1165 149 1.14 1.29 0.92 1.20 0.30 X1017 23 1.12 1.35 1.15 0.90 0.86 XI166 18 1.49 1.89 1.78 0.98 0.35 X1021 103 1.07 0.87 1.15 1.15 0.36 X1167 45 1.27 1.41 1.09 1.30 0.24 X1028 31 1.32 1.71 1.05 1.14 0.39 X1170 45 1.07 0.69 1.28 1.26 0.37 X1031 427 1.15* 1.08 1.20 1.17 0.04 X1172 28 1.22 0.75 1.38 1.58 0.16 X1036 117 1.10 1.03 0.92 1.33 0.22 X1173 111 1.08 1.00 1.07 1.16 0.38 X1038 440 1.25** 1.17 1.29* 1.27* <.01 X1174 18 1.25 0.99 1.27 1.52 0.28 X1041 16 1.78* 1.75 1.39 2.17 0.04 X1175 39 1.29 1.50 1.48 0.89 0.42 XI049 17 1.03 1.59 1.37 0.28 0.53 X1184 51 1.18 1.16 1.42 1.01 0.42 X1056 28 1.05 0.82 0.94 1.39 0.52 X1185 156 1.15 1.10 1.15 1.21 0.12 X1067 16 1.11 1.58 0.40 1.36 0.83 X1187 85 0.96 1.08 0.78 0.96 0.60 X1068 257 1.25** 1.34* 1.16 1.26 0.02 X1197 63 0.90 0.98 0.87 0.84 0.37 X1075 161 1.42** 1.22 1.59** 1.44* <.01 XI209 11 0.67 0.88 0.86 0.20 0.12 X1079 321 1.20* 1.10 1.21 1.29* <.01 X1217 42 1.23 1.13 1.78* 0.80 0.40 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal X1219 250 1.19* 1.24 1.08 1.25 0.05 X1355 9 1.06 1.64 0.36 1.16 0.94 X1222 50 0.96 0.97 0.96 0.93 0.76 X1357 15 0.85 0.66 0.67 1.27 0.87 X1224 26 1.23 0.71 1.43 1.55 0.17 X1367 199 1.09 1.18 1.02 1.06 0.51 X1226 11 1.71 2.41 2.57 0.00 0.49 X1376 27 1.25 0.73 1.57 1.43 0.17 X1228 121 1.07 1.05 0.95 1.21 0.39 X1379 157 1.25* 1.36* 1.08 1.32 0.04 X1230 10 1.05 0.67 1.82 1.01 0.73 X1380 92 1.01 0.91 1.19 0.94 0.93 X1231 352 1.22** 1.09 1.36** 1.21 <.01 XI394 127 1.36** 1.38 1.60** 1.11 0.02 X1239 15 1.34 0.77 1.48 1.85 0.16 X1395 221 1.23* 0.94 1.51** 1.23 <.01 XI240 31 0.73 0.86 0.63 0.68 0.09 X1396 177 1.35** 1.50** 1.49** 1.07 0.02 , X1255 35 0.94 0.85 0.72 1.23 0.96 X1401 35 2.18** 1.62 2.68** 2.25* <.01 X1256 12 1.00 0.53 0.97 1.45 0.67 X1402 67 . 1.00 1.19 0.81 0.93 0.70 X1257 15 1.18 1.48 . 0.62 1.55 0.57 X1411 56 1.19 1.16 1.04 1.36 0.22 X1258 10 0.73 0.26 0.54 1.39 0.76 XI424 9 2.83* 1.10 4.48* 2.70 <.01 X1266 9 1.32 0.43 1.31 2.65 0.18 X1425 218 1.21* 1.26 1.19 1.18 0.06 X1267 19 1.13 0.58 1.26 1.50 0.35 X1442 127 1.09 1.13 1.11 1.04 0.52 X1268 61 1.21 0.89 1.56* 1.20 0.12 X1447 151 1.01 0.98 1.20 0.84 0.84 X1280 37 1.08 1.21 0.78 1.23 0.67 XI448 14 1.82 2.00 1.46 2.04 0.07 X1281 18 1.07 0.61 1.16 1.56 0.40 X1449 102 1.03 0.87 1.04 1.16 0.52 X1302 14 1.60 0.66 1.42 2.77* 0.03 X1450 15 1.33 2.45* 1.40 0.27 0.98 X1303 52 0.95 1.00 1.17 0.76 0.56 X1454 24 1.38 1.81 1.09 1.24 0.33 XI304 12 1.07 1.37 0.00 1.96 0.70 X1456 170 1.10 1.25 1.04 1.03 0.56 XI306 23 0.94 1.08 1.03 0.69 0.61 X1457 94 1.29* 1.41 1.26 1.22 0.08 XI308 14 1.08 1.16 1.75 0.48 0.89 X1458 . 41 1.12 0.87 1.00 1.55 0.25 X1312 285 1.18* 1.10 1.30* 1.13 0.04 X1459 17 1.04 1.13 0.41 1.52 0.73 X1314 15 1.26 1.38 1.42 0.99 0.60 X1460 87 1.26 1.50* 1.07 1.19 0.18 X1322 176 1.13 1.00 1.24 1.14 0.15 X1463 50 1.19 1.17 1.26 1.16 0.32 X1329 61 1.24 1.40 1.29 1.01 0.34 X1464 18 1.14 0.43 2.04* 0.88 0.49 X1333 24 1.10 0.45 1.40 1.38 0.35 X1468 25 1.51 1.36 1.51 1.68 0.07 X1334 15 1.38 1.50 0.88 1.75 0.26 X1471 39 1.13 0.70 0.96 1.76* 0.13 X1335 104 1.12 1.09 1.27 1.00 0.44 X1475 10 1.12 1.68 0.58 1.12 0.96 X1336 14 1.13 0.55 1.53 1.24 0.48 X1484 17 1.37 2.48* 0.60 1.14 0.66 X1340 13 1.15 0,54 1.57 1.35 0.43 X1485 90 1.14 1.22 1.12 1.07 0.43 XI342 10 0.91 0.62 2.19 0.51 0.79 X1486 101 1.17 1.04 1.13 1.32 0.11 XI343 15 0.95 0.36 1.03 1.47 0.64 X1488 15 1.01 0.45 1.59 0.86 0.81 X1350 19 0.94 1.57 0.45 0.84 0.50 X1490 10 1.38 2.89* 0.80 0.41 0.92 X1351 12 1.10 0.51 0.69 2.04 0.31 X1496 83 1.39* 1.61* 1.43 1.13 0.06 X1354 18 1.03 0.50 1.67 0.84 0.74 X1497 34 1.20. 0.70 1.27 1.63 0.13 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal XI500 105 1.31* 1.33 1.23 1.38 0.03 X1653 28 1.27 1.36 1.22 1.24 0.34 X1503 108 1.26* 1.49* 1.13 1.15 0.15 X1654 166 1.16 1.25 1.24 1.00 0.32 X1505 115 1.03 0.98 1.18 0.93 0.86 X1655 26 1.20 0.90 1.28 1.39 0.28 X1506 86 1.06 0.95 0.89 1.31 0.38 X1656 170 1.34** 1.52** 1.48** 1.04 0.03 X1507 10 0.95 1.28 0.34 1.11 0.81 X1657 343 1.12 1.19 0.99 1.20 0.14 X1508 297 1.24** 1.05 1.26 1.41** <.01 X1658 34 1.22 0.86 1.58 1.27 0.21 X1509 370 1.25** 1.19 1.26* 1.31* <.01 X1667 9 0.99 0.88 0.00 2.10 0.63 X1511 34 0.97 1.23 1.15 0.53 0.47 X1673 12 1.13 0.51 0.71 2.16 0.26 X1512 287 1.13 1.28* 1.16 0.96 0.51 X1674 18 1.26 1.51 0.94 1.28 0.50 X1513 117 1.03 1.31 0.78 1.02 0.88 X1680 14 1.65 0.88 1.67 2.78* 0.03 X1516 41 1.15 0.67 1.79* 0.89 0.34 X1688 13 1.86* 4.50** 0.38 1.23 0.40 X1550 101 1.26* 1.53* 0.98 1.29 0.12 X1696 13 0.97 0.46 1.23 1.20 0.79 X1551 111 1.15 0.99 1.23 1.23 0.14 X1698 131 1.11 1.24 1.12 0.99 0.57 X1569 59 1.21 0.97 1.39 1.28 0.14 X1699 12 1.05 1.03 0.77 1.40 0.76 X1573 242 1.16 1.24 1.01 1.23 0.10 XI700 32 1.15 1.03 1.52 0.89 0.57 X1576 186 1.15 0.90 1.19 1.36* 0.02 X1712 15 1.19 0.70 1.60 1.30 0.39 X1579 249 1.10 1.12 1.10 1.08 0.33 X1718 87 1.29* 1.29 1.18 1.39 0.05 X1580 151 1.12 0.93 1.18 1.24 0.12 X1752 128 1.26* 1.23 1.20 1.37 0.02 X1585 83 1.29* 1.56* 1.20 1.10 0.18 X1783 13 0.58 0.43 0.62 0.67 0.12 XI586 106 1.30* 1.55* 1.28 1.09 0.12 X1791 9 1.17 0.90 1.63 1.04 0.62 X1588 19 1.20 0.44 1.98 1.12 0.33 XI794 15 0.65 0.65 0.25 1.10 0.28 X1590 24 1.34 1.10 1.37 1.54 0.16 X1808 176 1.18 1.43** 0.99 1.13 0.25 X1594 15 0.78 0.45 0.89 1.03 0.67 X1814 10 0.76 1.24 0.50 0.56 0.27 X1595 14 1.02 0.43 0.88 1.88 0.43 X1827 63 0.93 0.91 1.14 0.81 0.54 X1597 9 2.08 2.80 0.57 3.32 0.08 X1829 14 1.06 1.52 0.98 0.67 0.78 XI598 89 1.00 1.01 0.75 1.23 0.74 X1830 13 1.08 1.57 0.91 0.78 0.89 X1613 19 0.95 0.51 0.95 1.34 0.78 X1833 13 0.66 1.21 0.41 0.44 0.07 X1636 11 0.83 0.76 1.04 0.69 0.58 X1835 24 1.05 1.34 1.39 0.47 0.70 X1638 23 1.27 0.70 1.03 2.02* 0.10 X1836 116 1.23 1.33 1.03 1.33 0.08 X1639 326 1.10 1.12 1.16 1.03 0.36 X1837 203 1.18 1.28 0.98 1.27 0.08 X1641 17 1.23 0.56 1.46 1.88 0.17 X1841 129 1.16' 0.96 1.44* 1.07 0.15 X1642 26 1.08 1.12 0.93 1.22 0.69 X1842 46 1.30 1.67* 1.09 1.12 0.32 XI643 12 1.29 0.32 1.25 2.34 0.13 X1850 45 1.29 1.67* 1.24 0.95 0.44 XI646 74 1.12 0.98 1.46 0.91 0.51 X1867 21 1.34 0.73 1.66 1.71 0.11 X1650 23 1.35 0.68 1.49 1.90 0.07 X1868 10 1.02 2.53* 0.62 0.24 0.35 X1651 118 1.11 0.83 1.37 1.13 0.18 X1869 34 0.95 1.25 0.60 0.98 0.62 X1652 166 1.17 1.26 1.25 1.01 0.28 X1872 410 1.25** 1.22 1.23* 1.30* <.01 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal X1877 250 1.14 1.15 1.11 1.16 0.13 X2014 131 1.19 1.18 1.11 1.28 0.09 XI893 44 1.06 0.95 1.24 0.96 0.77 X2015 12 0.91 0.49 1.15 1.06 1.00 XI894 48 1.86** 1.63 1.73 2.20** <.01 X2016 226 1.06 1.18 0.93 1.08 0.67 X1895 13 1.06 2.10 0.62 0.69 0.66 X2017 9 1.16 0.90 1.91 0.54 0.75 X1897 356 1.26** 1.18 1.18 1.41** <.01 X2019 49 1.07 0.92 1.22 1.07 0.61 XI899 60 1.14 1.43 0.98 1.02 0.67 X2020 153 1.06 1.00 1.26 0.92 0.66 X1902 24 1.72* 2.45* 1.48 1.27 0.12 X2022 90 1.01 0.90 1.10 1.03 0.80 X1906 11 1.00 0.62 1.03 1.29 0.77 X2023 20 1.60 1.84 1.94 1.16 0.17 X1909 122 1.13 1.44* 0.86 1.08 0.61 X2027 50 1.22 0.98 1.35 1.29 0.16 X1910 13 0.90 0.56 0.22 1.93 0.68 X2028 10 1.72 1.39 1.11 2.75 0.07 X1918 24 1.72* 2.81** 1.03 1.27 0.17 X2029 127 1.30* 1.35 1.19 1.37 0.02 X1922 11 1.04 1.91 0.71 0.59 0.64 X2031 38 1.44* 1.61 0.97 1.71 0.06 X1923 17 0.96 1.00 1.69 0.32 0.57 X2035 18 1.49 1.89 1.78 0.98 0.35 X1925 55 1.25 1.17 1.50 1.16 0.19 X2062 103 1.06 0.79 1.40 1.00 0.44 X1930 20 1.60 2.03 2.05 0.95 0.24 X2063 27 1.08 0.88 1.39 0.95 0.71 X1936 16 0.85 0.71 0.50 1.53 0.92 X2065 76 1.22 1.33 1.30 1.03 0.29 XI948 21 1.25 1.57 1.13 1.09 0.55 X2066 105 1.08 0.99 1.10 1.16 0.37 X1957 17 1.23 0.57 1.46 1.88 0.17 X2083 10 1.52 1.11 2.51 1.30 0.24 XI966 17 1.13 1.82 1.02 0.43 0.70 X2143 80 1.26 1.04 1.57* 1.24 0.06 X1970 18 1.15 1.05 0.96 1.40 0.50 X2145 40 1.42 1.04 1.80* 1.45 0.04 X1977 9 1.19 0.90 2.31 0.48 0.72 X2180 27 1.27 0.95 1.06 1.77 0.12 X1980 38 1.35 1.03 1.80* 1.27 0.09 X2192 32 1.22 1.01 1.30 1.35 0.25 X1981 12 1.14 1.36 1.75 0.29 0.91 X2202 12 1.11 1.04 0.75 1.61 0.59 XI984 25 1.37 0.71 1.96* 1.36 0.10 X2204 94 1.28* 1.59* 1.08 1.16 0.18 X1986 32 1.25 1.02 1.76 0.97 0.33 X2283 11 1.08 0.79 1.78 0.66 0.86 XI988 12 1.42 1.83 1.89 0.63 0.58 X2293 375 1.30** 1.26* 1.34** 1.30* <.01 X1992 29 1.30 0.97 1.20 1.74 0.11 X2295 450 1.30** 1.25* 1.35** 1.31** <.01 X1994 22 1.05 .1.10 0.60 1.41 0.71 X2297 262 1.25** 1.14 1.25 1.37** <.01 X1995 13 1.60 1.60 1.90 1.26 0.21 X2298 390 1.34** 1.33** 1.41** 1.27* <.01 X1998 22 1.21 0.95 1.00 1.61 0.26 X2303 361 1.39** 1.42** 1.42** 1.33* <.01 XI999 19 1.01 0.74 1.05 1.19 0.78 X2305 206 1.38** 1.22 1.32 1.62** <.01 X2001 23 1.21 0.54 1.12 1.92 0.14 X2306 335 1.38** 1.40** 1.41** 1.34* <.01 X2003 9 1.70 1.28 0.48 3.61* 0.06 X2307 301 1.38** 1.40** 1.39** 1.36* <.01 X2006 65 1.06 1.05 0.98 1.16 0.59 X2308 208 1.40** 1.25 1.34* 1.61** <.01 X2007 59 1.29 1.45 1.55 0.90 0.29 X2309 31 1.23 1.06 1.46 1.17 0.30 X2011 33 0.77 0.80 1.11 0.37* 0.10 X2310 154 1.04 1.01 0.92 1.18 0.50 X2013 14 1.34 1.28 1.44 1.29 0.37 X2311 155 1.08 1.12 1.05 1.08 0.49 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal X2312 18 0.96 0.38 1.08 1.30 0.71 X2396 12 0.84 1.14 0.46 0.91 0.54 X2314 24 0.92 0.86 0.95 0.95 0.77 X2398 98 1.31* 1.54* 1.31 1.08 0.12 X2315 9 1.88 2.19 3.34* 0.00 0.33 X2400 10 1.02 2.15 0.93 0.24 0.42 X2316 14 0.94 0.45 1.04 1.28 0.80 X2401 . 20 1.06 0.46 1.30 1.49 0.39 X2317 33 1.19 1.01 1.34 1.21 0.34 X2403 68 1.02 0.97 0.96 1.13 0.73 X2318 23 1.02 0.73 1.12 1.18 0.71 X2404 10 1.02 2.15 0.93 0.24 0.42 X2319 23 1.20 0.73 1.30 1.48 0.26 X2405 48 1.07 1.65* 1.09 0.52 0.42 X2320 33 0.95 0.92 1.20 0.72 0.68 X2417 77 1.32* 1.06 1.67* 1.31 0.03 X2325 18 1.06 0.93 0.73 1.63 0.57 X2418 18 1.03 0.39 1.13 1.49 0.49 X2327 10 0.67 0.63 0.83 0.55 0.25 X2423 64 1.06 0.88 0.96 1.30 0.40 X2328 20 1.09 0.52 1.16 1.55 0.36 X2424 31 1.21 0.67 1.58 1.45 0.15 X2329 15 1.20 1.08 0.46 2.22 0.27 X2436 13 0.77 1.91 0.16 0.46 0.11 X2330 15 1.20 1.08 0.46 2.22 0.27 X2441 11 0.88 0.48 1.61 0.51 0.72 X2331 15 1.20 1.08 0.46 2.22 0.27 X2449 179 1.05 1.17 1.11 0.90 0.98 X2332 15 1.20 1.08 0.46 2.22 0.27 X2463 18 1.06 1.57 0.98 0.64 0.71 X2333 15 1.20 1.08 0.46 2.22 0.27 X2467 13 1.02 0.90 1.12 1.05 0.88 X2335 86 0.99 1.07 0.73 1.16 0.97 X2468 10 1.02 2.15 0.93 0.24 0.42 X2336 48 1.06 0.96 1.37 0.82 0.82 X2470 96 1.31* 1.46 1.25 1.23 0.06 X2342 12 0.80 1.32 0.63 0.40 0.23 X2475 111 1.06 1.00 0.76 1.42* 0.27 X2346 15 0.93 0.60 1.02 1.16 0.95 X2480 118 1.11 1.06 1.26 1.03 0.38 X2351 16 1.13 0.98 0.60 1.94 0.39 X2482 14 1.32 1.57 1.38 0.95 0.56 X2352 16 1.19 0.89 1.33 1.35 0.42 X2496 53 0.91 1.00 1.00 0.72 0.34 X2354 12 0.81 1.29 0.78 0.39 0.25 X2501 24 1.37 0.94 1.43 1.80 0.09 X2361 100 1.22 1.28 1.11 1.29 0.11 X2513 118 1.11 1.28 1.07 0.98 0.66 X2363 98 1.29* 1.48* 1.16 1.24 0.09 X2514 118 1.11 1.28 1.07 0.98 0.66 X2365 23 1.27 1.73 0.34 1.72 0.39 X2517 56 1.30 1.54 1.13 1.24 0.19 X2373 18 1.01 0.38 1.11 1.49 0.52 X2518 218 1.23* 1.16 1.27 1.26 0.01 X2377 64 1.11 1.07 1.09 1.15 0.46 X2521 9 1.05 1.24 0.78 1.08 0.96 X2378 9 0.91 0.55 0.85 1.41 0.89 X2522 9 1.04 1.13 0.85 1.18 0.91 X2379 12 1.17 1.72 0.30 1.39 0.81 X2523 136 1.15 1.25 1.29 0.89 0.46 X2380 85 1.29* 1.56* 1.23 1.08 0.19 X2529 46 0.97 0.88 1.24 0.79 0.79 X2381 30 1.07 1.02 1.35 0.86 0.84 X2532 10 0.94 0.58 0.52 1.76 0.71 X2384 175 1.14 1.23 1.04 1.13 0.28 X2533 13 1.55 2.96* 1.78 0.31 0.69 X2386 11 0.83 0.84 0.63 1.05 0.66 X2534 97 1.24 1.08 1.34 1.33 0.05 X2393 55 1.26 1.53 1.00 1.25 0.25 X2537 21 0.86 1.01 0.85 0.76 O.46 X2394 21 1.06 0.87 1.17 1.11 0.73 X2538 62 1.24 1.31 0.97 1.47 0.13 X2395 11 1.11 0.31 1.52 1.50 0.42 X2541 22 1.18 0.70 1.56 1.24 0.34 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal X2548 71 1.18 1.31 1.06 1.15 0.35 X2812 62 1.05 0.99 1.06 1.10 0.67 X2549 54 0.92 0.98 1.07 0.73 0.39 X2821 68 1.36* 0.95 1.49 1.62* <.01 X2555 52 0.94 1.00 0.93 0.87 0.60 X2822 68 1.00 0.96 0.86 1.17 0.79 X2560 31 1.30 1.66 1.02 1.16 0.41 X2823 14 0.94 0.62 0.83 1.36 0.83 X2562 23 1.08 0.67 2.19* 0.45 0.83 X2826 12 1.13 0.51 0.71 2.16 0.26 X2570 21 1.15 0.68 1.62 1.12 0.44 X2836 9 1.39 0.00 1.71 3.27* 0.07 X2572 19 1.31 0.57 1.52 1.98 0.09 X2838 91 1.37** 1.63* 0.98 1.51* 0.02 X2582 9 1.05 0.97 1.04 1.16 0.84 X2839 28 1.28 1.32 1.30 1.22 0.32 X2598 45 1.20 1.54 1.09 0.96 0.61 X2842 25 1.40 0.68 1.97 1.57 0.06 X2604 26 1.08 1.45 1.41 0.37 0.65 X2844 11 1.67 3.08* 0.81 1.30 0.40 X2608 17 1.15 0.56 1.10 1.76 0.25 X2847 9 1.16 0.90 1.91 0.54 0.75 X2629 65 1.24 1.32 1.19 1.23 0.18 X2851 9 0.85 0.81 0.33 1.33 0.87 X2644 9 1.07 0.85 1.11 1.35 0.70 X2852 173 1.27** 1.55** 1.01 1.23 0.07 X2652 44 1.10 1.02 0.90 1.38 0.42 X2861 23 1.19 0.83 1.30 1.40 0.31 X2656 123 1.17 1.04 1.39* 1.08 0.16 X2862 18 1.08 1.59 0.81 0.72 0.78 X2657 25 1.45 1.59 2.02* 0.83 0.28 X2863 177 1.26* 1.50** 1.11 1.15 0.09 X2661 84 1.11 0.86 1.26 1.21 0.23 X2864 56 1.25 1.44 1.55 0.79 0.45 X2672 18 1.39 1.28 0.91 1.89 0.15 X2865 59 0.99 0.98 0.97 1.02 1.00 X2674 50 1.17 1.00 1.30 1.20 0.28 X2866 64 1.10 1.07 1.08 1.15 0.49 X2676 146 1.13 1.16 0.99 1.23 0.21 X2867 39 1.31 1.16 1.76 1.10 0.19 X2686 45 1.23 1.59 1.21 0.90 0.59 X2869 21 1.12 0.67 ' 1.58 1.10 0.50 X2689 25 2.89** 1.93 2.95** 4.00** <.01 X2871 131 1.10 1.18 1.04 1.09 0.46 X2701 37 1.29 1.10 1.47 1.29 0.16 X2873 9 0.95 0.83 1.14 0.85 0.92 X2703 14 1.17 0.65 1.04 2.07 0.27 X2874 9 0.89 0.57 0.98 1.14 0.98 X2712 23 1.37 1.44 1.27 1.43 0.21 X2876 18 1.61 1.92 0.87 1.99 0.10 X2730 68 1.20 1.41 1.18 1.01 0.44 X2878 25 1.28 1.39 0.91 1.56 0.26 X2732 45 1.29 1.67* 1.24 0.95 0.44 X2881 56 1.19 1.16 1.04 1.36 0.22 X2744 67 1.00 1.19 0.81 0.93 0.70 X2887 85 1.27 1.63* 1.14 1.03 0.29 X2754 14 1.04 0.96 0.95 1.24 0.79 X2897 11 1.12 0.50 1.54 1.46 0.45 X2769 15 1.10 1.28 1.58 0.49 0.89 X2900 15 0.95 0.69 0.66 1.54 0.78 X2777 106 1.03 0.93 0.80 1.33 0.41 X2905 12 1.13 0.51 0.71 2.16 0.26 X2789 32 1.22 1.01 1.30 1.35 0.25 X2925 64 1.31 1.50 1.29 1.10 0.18 X2793 364 1.20* 1.06 1.36** 1.18 <.01 X2938 91 1.23 1.28 1.37 1.01 0.21 X2797 47 1.30 1.14 1.72* 1.04 0.16 X2951 23 1.08 1.70 0.72 0.81 0.7 7 X2803 19 1.13 1.03 1.45 1.02 0.66 X2954 103 1.39** 1.60** 1.54* 1.02 0.06 X2805 21 0.91 0.80 1.13 0.82 0.72 X2955 53 0.96 0.96 1.22 0.78 0.64 X2807 9 0.85 0.57 0.88 1.08 0.85 X2963 39 1.00 0.67 1.21 1.15 0.65 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal X2965 11 0.80 0.42 1.13 0.86 0.69 X3100 39 1.71** 1.52 1.86 1.78 <.01 X2966 10 1.03 0.36 1.07 1.63 0.57 X3103 180 1.02 0.96 1.18 0.92 0.93 X2973 57 0.97 0.99 1.08 0.87 0.74 X3104 45 1.26 1.38 0.74 1.70* 0.13 X2974 89 1.32* 1.55* 1.21 • 1.17 0.11 X3105 130 1.08 0.98 1.17 1.12 0.34 X2981 61 0.96 1.00 0.93 0.94 0.72 X3106 10 0.99 1.84 0.63 0.33 0.44 X2982 80 1.25 1.44 1.34 0.98 0.27 X3108 57 1.12 1.45 0.63 1.27 0.63 X2984 68 1.37* 1.25 1.85** 0.96 0.07 X3111 11 1.04 1.49 1.25 0.48 0.69 X2986 9 2.02 4.34** 1.78 0.00 0.57 X3117 47 1.00 0.92 1.02 1.06 0.90 X2987 16 0.90 0.53 1.32 0.84 0.87 X3120 69 1.03 1.00 0.99 1.10 0.74 X2988 13 1.84 2.34 2.77* 0.38 0.25 X3130 28 0.99 1.60 0.64 0.77 0.50 X3000 61 1.16 0.90 1.44 1.16 0.23 X3142 41 1.14 1.07 1.39 0.96 0.55 X3003 124 1.12 0.93 1.30 1.13 0.19 X3148 11 1.91 1.99 3.03* 0.58 0.17 X3007 125 0.99 0.98 1.01 0.97 0.89 X3152 14 0.87 1.30 0.88 0.38 0.33 X3008 12 1.13 0.56 0.60 2.16 0.29 X3160 20 1.05 1.26 0.98 0.91 0.97 X3009 140 1.10 1.20 1.04 1.05 0.53 X3167 64 1.24 1.17 1.33 1.21 0.16 X3012 52 0.96 0.99 1.17 0.78 0.61 X3204 28 1.67* 2.30* 1.76 1.04 0.13 X3014 539 1.16* 1.14 1.15 1.20* 0.03 X3205 16 0.96 1.06 1.62 0.49 0.59 X3017 296 1.25** 1.27* 1.37** 1.11 0.02 X3211 140 1.34** 1.68** 1.26 1.12 0.06 X3019 252 1.18* 1.26 1.23 1.07 0.13 X3220 20 1.01 1.47 0.82 0.77 0.69 X3020 169 1.09 1.29 0.87 1.10 0.61 X3231 184 1.17 0.85 1.53** 1.12 0.04 X3025 53 0.96 0.99 1.17 0.80 0.65 X3232 15 1.06 1.22 1.39 0.48 0.85 X3027 262 1.04 1.03 1.10 1.01 0.66 X3235 64 1.03 0.72 1.37 0.99 0.64 X3032 186 1.16 0.99 1.24 1.25 0.05 X3239 16 1.17 1.04 2.26 0.83 0.68 X3033 45 0.98 0.85 0.90 1.18 0.84 X3241 100 1.06 0.85 0.87 1.44* 0.21 X3038 230 1.20* 1.32* 1.10 1.19 0.07 X3243 11 0.99 1.00 0.94 1.01 0.97 X3039 39 1.07 1.49 0.45 1.28 0.92 X3256 96 1.05 1.09 1.21 0.85 0.96 X3045 81 1.19 1.27 0.98 1.35 0.18 X3262 104 1.14 1.18 1.03 1.21 0.26 X3048 21 0.99 0.75 1.17 1.04 0.89 X3264 21 1.07 0.94 0.97 1.31 0.64 X3051 20 1.04 0.98 1.40 0.82 0.96 X3265 290 1.21* 1.24 1.10 1.28* 0.02 X3054 42 0.97 1.01 0.78 1.14 0.98 X3269 12 1.38 1.31 1.26 1.54 0.31 X3064 304 1.27** 1.19 1.31* 1.32* <.01 X3272 14 0.85 1.02 0.53 1.01 0.61 X3069 103 1.18 1.56** 0.84 1.12 0.46 X3275 38 1.35 1.13 1.72 1.23 0.11 X3077 13 1.00 0.71 1.25 1.06 0.86 X3278 15 0.94 1.12 0.76 0.95 0.77 X3082 178 1.08 1.24 0.85 1.11 0.62 X3286 167 1.19 1.13 1.15 1.28 0.05 X3088 37 1.19 1.13 1.26 1.17 0.37 X3287 14 1.18 0.96 0.46 2.30* 0.27 X3089 53 1.12 1.46 0.96 0.94 0.88 X3289 18 0.86 0.81 0.64 1.12 0.75 X3093 13 0.63 1.09 0.16 0.59 0.08 X3290 11 1.07 0.80 0.91 1.56 0.59 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal X3373 11 1.55 0.90 1.60 2.16 0.11 X3989 29 1.35 1.16 1.44 1.44 0.14 X3396 15 0.73 1.02 0.76 0.43 0.15 X3991 18 1.07 1.69 1.39 0.30 0.54 X3401 21 1.08 1.13 0.94 1.15 0.77 X3992 33 1.06 1.37 0.81 1.01 1.00 X3542 20 0.63 1.21 0.42 0.37 0.02 X4004 31 0.93 0.87 0.72 1.22 0.95 X3558 12 0.87 0.45 1.80 0.42 0.66 X4016 23 1.12 0.41 1.42 1.59 0.23 X3559 64 0.91 0.91 1.22 0.63 0.33 X4021 32 0.94 1.61 0.79 0.36* 0.15 X3569 12 1.68 0.36 2.15 2.82* 0.02 X4025 21 1.18 1.39 1.27 0.95 0.72 X3570 11 1.51 0.36 1.61 2.81* 0.05 X4033 11 1.99* 0.49 3.50** 2.01 0.02 X3635 104 1.33* 1.57* 1.22 1.18 0.08 X4041 17 1.14 1.36 0.94 1.10 0.77 X3644 46 1.46* 1.63 1.55 1.21 0.08 X4042 111 1.25* 1.38 1.34 1.03 0.17 X3647 11 2.00* 1.79 1.20 2.82* 0.03 X4045 64 1.43* 1.45 1.62* 1.22 0.04 X3672 373 1.21** 1.11 1.21 1.31** <.01 X4046 18 0.79 0.14 0.92 1.25 0.90 X3691 245 1.12 1.12 1.06 1.18 0.16 X4057 34 1.37 1.12 2.14** 0.89 0.17 X3693 25 1.55 0.57 1.84 2.25* 0.01 X4063 18 1.33 1.82 1.18 0.92 0.60 X3695 42 1.39 1.76* 1.06 1.32 0.17 X4067 15 1.37 2.05 0.45 1.79 0.38 X3700 10 0.61 0.80 0.59 0.39 0.10 X4073 11 1.12 0.35 1.82 1.11 0.55 X3712 251 1.15 1.18 1.12 1.15 0.13 X4096 30 1.08 0.81 1.01 1.43 0.43 X3716 11 0.89 1.36 1.27 0.00 0.30 X4097 75 1.05 1.07 1.16 0.92 0.88 X3720 - 20 1.04 0.78 1.00 1.34 0.62 X4101 96 0.99 1.03 1.23 0.73 0.60 X3722 10 0.88 0.95 1.09 0.58 0.60 X4103 13 1.72 1.73 1.60 1.81 0.11 X3743 16 1.14 1.78 0.00 1.92 0.58 X4104 16 1.50 1.54 1.24 1.77 0.15 X3755 11 1.50 2.03 1.45 0.88 0.51 X4107 9 0.62 0.47 0.61 0.75 0.27 X3761 245 1.22* 1.14 1.12 1.40** <.01 X4113 54 0.93 0.98 1.06 0.77 0.49 X3764 18 0.82 0.57 1.48 0.48 0.41 X4119 13 0.99 0.64 2.19* 0.23 0.81 X3765 9 1.19 1.60 1.12 0.85 0.88 X4131 68 1.31 1.25 1.62* 1.06 0.11 X3767 191 1.19* 1.19 1.28 1.11 0.10 X4134 9 1.07 0.55 2.22 0.47 0.80 X3782 357 1.09 1.19 1.27* 0.81 0.92 X4140 9 1.96 3.19* 1.64 1.25 0.25 X3812 40 1.11 1.01 0.90 1.41 0.39 X4178 15 1.08 1.43 0.82 0.99 0.99 X3835 16 1.10 1.11 0.82 1.35 0.66 X4207 16 1.67 1.09 0.97 3.09** 0.02 X3842 176 1.09 1.05 1.07 1.16 0.27 X4230 12 3.52** 2.71 4.22** 3.50 <.01 X3845 9 1.50 1.40 2.35 0.60 0.44 X4231 30 0.96 1.00 1.05 0.83 0.73 X3870 142 1.10 1.32 0.98 0.94 0.89 X4233 9 0.75 0.85 0.79 0.65 0.40 X3881 109 1.20 1.42* 1.08 1.11 0.27 X4237 79 0.99 1.06 0.92 0.98 0.84 X3912 12 1.59 1.63 1.58 1.57 0.20 X4242 158 1.00 0.89 0.92 1.19 0.55 X3941 145 1.08 1.26 0.93 1.04 0.73 X4255 31 0.97 0.85 1.30 0.72 0.83 X3950 16 0.80 0.17 1.14 1.02 0.81 X4263 27 1.10 1.24 1.24 0.78 0.92 X3955 36 0.86 1.09 0.96 0.54 0.19 X4267 32 1.96** 1.19 2.40* 2.56** <.01 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal X4268 10 1.88 0.53 3.24* 1.56 0.05 X5075 9 0.94 0.61 0.97 1.24 0.91 X4282 29 1.40 1.88* 1.17 0.98 0.42 X5085 11 1.99 1.12 2.37 2.50 0.03 X4297 11 1.45 0.86 1.25. 2.16 0.14 X5093 12 2.97** 3.00* 4.35* 1.67 0.01 X4330 11 1.81 1.13 3.21* 0.96 0.13 X5094 16 1.21 1.08 1.13 1.42 0.43 X4393 91 1.01 1.05 1.03 0.95 0.95 X5115 38 1.28 0.89 1.70 1.22 0.15 X4425 9 1.58 0.69 3.74* 1.28 0.15 X5131 25 1.06 1.00 1.01 1.18 0.72 X4471 14 0.99 0.44 2.15* 0.41 0.96 X5135 12 1.43 2.39 0.69 1.34 0.50 X4521 11 0.92 1.54 0.97 0.27 0.41 X5145 46 1.18 1.26 1.50 0.84 0.58 X4540 24 1.11 1.32 0.60 1.49 0.58 X5161 49 1.30 1.66* 1.18 1.03 0.36 X4542 83 0.98 0.97 0.94 1.01 0.89 X5184 10 3.00** 5.62** 0.90 2.86 0.03 X4548 35 1.25 1.56 1.67 0.52 0.79 X5185 90 1.07 1.32 1.07 0.76 0.80 X4599 17 1.02 1.43 0.51 1.15 0.90 X5189 9 1.67 0.53 2.66 1.83 0.10 X4622 9 1.27 0.91 1.99 0.85 0.57 X5192 13 0.85 0.58 1.03 0.93 0.75 X4668 35 1.27 1.66 0.76 1.37 0.36 X5206 9 0.93 0.64 1.31 0.85 0.91 X4697 67 0.99 1.05 1.14 0.82 0.71 X5213 28 1.16 0.63 2.32** 0.65 0.53 X4699 9 3.02** 2.53 2.91 3.66 <.01 X5235 11 1.07 2.06 0.71 0.59 0.66 X4731 78 1.09 1.15 1.18 0.95 0.75 X5258 11 0.97 1.77 0.51 0.60 0.51 X4753 20 1.41 0.70 1.34 2.13* 0.06 X5262 631 1.21** 1.15 1.11 1.36** <.01 X4779 109 1.18 1.40 1.21 0.91 0.52 X5263 557 1.28** 1.24* 1.24* 1.37** <.01 X4785 36 1.25 1.65 0.70 1.41 0.37 X5299 255 0.97 0.90 1.06 0.95 0.81 X4794 9 2.65* 3.29 2.39 2.39 0.04 X5311 58 1.20 1.46 1.06 1.08 0.44 X4849 16 1.34 2.19* 0.52 1.26 0.61 X5318 11 1.26 1.80 1.75 0.00 0.96 X4891 9 1.52 0.48 3.14* 0.64 0.25 X5404 149 1.24* 1.12 1.32 1.27 0.03 X4906 10 1.08 1.02 1.61 0.63 0.97 X5408 31 1.11 1.16 1.15 1.04 0.69 X4918 95 1.09 0.94 1.10 1.23 0.30 X5417 397 1.14 1.07 1.30** 1.07 0.08 X4922 12 1.08 2.06 0.66 0.78 0.77 X5421 9 0.96 1.38 1.11 0.35 0.59 X4987 75 1.05 0.87 0.90 1.40 0.32 X5427 9 1.19 1.29 0.82 1.42 0.63 X5001 9 1.65 1.20 1.84 1.83 0.16 X5447 164 1.18 1.14 1.28 1.11 0.12 X5029 289 1.14 1.27* 1.02 1.09 0.30 X5459 22 1.08 0.71 1.32 1.25 0.51 X5034 21 1.01 1.19 0.53 1.32 0.91 X5467 29 1.04 0.72 1.00 1.40 0.51 X5037 21 0.92 1.27 0.55 0.96 0.61 X5495 34 1.40 1.63 1.40 1.18 0.21 X5044 36 0.95 1.06 0.70 1.08 0.80 X5502 182 1.24* 1.21 1.22 1.30 0.02 X5054 33 1.23 1.13 1.08 1.45 0.23 X5503 250 1.07 1.10 1.09 1.02 0.58 X5065 9 1.82 2.16 1.03 2.47 0.14 X5504 20 1.03 0.58 0.90 1.60 0.44 X5067 62 1.38* 1.36 1.66* 1.13 0.06 X5505 32 0.84 0.81 0.96 0.75 0.36 X5068 10 1.60 1.93 1.52 1.44 0.27 X5507 10 1.03 0.67 1.46 0.91 0.86 X5071 54 1.37* 1.27 1.67* 1.15 0.07 X5509 52 1.27 1.19 1.45 1.16 0.18 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal X5515 89 1.21 1.30 1.21 1.12 0.23 X5892 9 0.97 0.24 1.41 1.63 0.56 X5516 171 1.24* 1.32 1.16 1.23 0.05 X5898 16 0.85 0.97 0.47 1.12 0.64 X5518 12 1.72 2.09 1.79 1.15 0.21 X5899 11 1.76 1.70 2.94* 0.55 0.24 X5520 137 1.10 1.44* 1.14 0.76 0.81 X5906 233 1.20* 1.16 1.22 1.22 0.03 X5521 21 0.92 0.83 0.45 1.60 0.86 X5907 21 1.10 0.87 1.29 1.17 0.58 X5582 9 1.00 0.74 2.60 0.00 0.80 X5911 158 1.07 1.07 0.99 1.15 0.44 X5629 11 1.01 1.65 1.17 0.27 0.55 X5915 25 1.29 1.04 1.85 0.89 0.32 X5638 98 1.12 0.99 1.15 1.22 0.23 X5918 30 1.33 1.35 1.39 1.22 0.24 X5654 14 0.87 0.38 0.85 1.44 0.88 X5928 30 1.11 1.35 0.84 1.19 0.73 X5658 24 1.06 0.76 1.10 1.27 0.59 X5930 25 1.33 1.09 1.63 1.17 0.23 X5673 38 1.13 0.80 0.95 1.63 0.20 X5935 16 1.10 0.65 0.95 1.76 0.39 X5683 60 0.99 1.00 1.52* 0.44* 0.47 X5940 13 0.92 0.40 1.07 1.33 0.81 X5686 501 1.12 1.21* 1.01 1.15 0.20 X5941 55 0.90 0.98 1.09 0.67 0.30 X5689 19 0.96 1.29 0.59 0.98 0.72 X5942 17 1.16 0.65 1.12 1.72 0.31 X5690 126 1.13 0.89 1.37 1.13 0.16 X5943 16 1.34 1.01 1.38 1.78 0.19 X5691 9 1.09 0.59 2.53 0.39 0.86 X5945 10 1.15 2.03 0.00 1.29 0.96 X5697 386 1.20** 1.12 1.31** 1.17 0.01 X5946 22 1.09 0.73 1.00 1.60 0.39 X5701 9 1.16 0.69 1.47 1.42 0.51 X5950 424 1.24** 1.34** 1.19 1.19 0.02 X5704 13 1.22 1.21 1.40 1.01 0.61 X5974 18 1.02 0.88 0.85 1.41 0.72 X5710 14 0.85 0.97 0.71 0.89 0.58 X5976 182 1.14 0.93 1.25 1.23 0.06 X5711 88 0.93 0.98 0.78 1.04 0.66 X5983 138 1.17 1.09 1.29 1.13 0.14 X5714 18 1.22 0.38 1.77 1.66 0.17 X5986 55 1.00 0.66 0.78 1.57* 0.32 X5752 12 1.53 0.36 3.12** 1.20 0.12 X5991 12 1.00 1.22 0.76 1.01 0.92 X5757 79 1.16 0.95 1.32 1.19 0.19 X5992 50 1.02 0.69 1.10 1.29 0.47 X5758 239 1.17 0.95 1.34* 1.21 0.02 X5997 39 1.03 1.40 0.99 0.72 0.68 X5808 17 1.11 1.28 0.53 1.60 0.62 X6034 14 1.08 0.26 1.32 1.61 0.38 X5811 12 1.07 1.85 0.49 0.84 0.76 X6115 83 1.29* 1.56* 1.20 1.10 0.18 X5822 80 1.31* 1.40 1.12 1.45 0.05 X6125 9 1.45 0.45 2.39 1.55 0.20 X5824 15 0.98 1.48 0.63 0.76 0.61 X6186 360 1.17* 1.15 1.17 1.18 0.04 X5825 12 1.07 1.77 0.95 0.54 0.71 X6191 139 1.10 1.20 1.04 1.06 0.53 X5849 470 1.17* 1.23* 1.16 1.13 0.08 X6205 34 0.85 0.78 0.94 0.82 0.44 . X5864 83 1.12 1.01 0.87 1.47* 0.17 X6222 141 1.04 0.92 1.07 1.12 0.49 X5873 88 1.03 1.12 0.77 1.16 0.81 X6238 182 1.19 1.00 1.39* 1.17 0.04 X5876 288 1.18* 1.33* 1.11 1.12 0.12 X6246 133 1.10 1.39* 0.94 0.97 0.84 X5886 204 0.96 1.05 0.93 0.91 O.46 X6265 10 0.95 0.45 2.26 0.33 0.91 X5890 11 0.79 0.73 0.76 0.88 0.55 X6289 66 1.37* 1.61* 1.00 1.46 0.06 X5891 15 1.03 1.12 0.88 1.07 0.95 X6293 293 1.17* 1.12 1.23 1.16 0.05 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal X6352 28 1.24 0.41 1.50 1.82 0.07 X6502 14 1.14 1.04 1.43 1.00 0.69 X6353 81 1.01 0.99 0.80 1.22 0.72 X6505 9 1.07 0.58 2.45 0.40 0.87 X6358 9 1.57 2.03 0.52 2.23 0.25 X6507 28 1.09 0.94 1.21 1.13 0.61 X6360 46 1.04 1.06 1.09 0.95 0.92 X6517 16 1.29 1.14 0.96 1.82 0.25 X6362 20 1.08 0.50 1.15 1.60 0.35 X6518 11 1.05 1.83 0.26 0.95 0.74 X6363 40 1.10 0.82 1.27 1.26 0.37 X6523 16 1.09 1.40 0.68 1.25 0.86 X6366 29 1.20 0.90 1.13 1.59 0.21 X6524 13 1.24 1.21 1.48 1.01 0.57 X6367 40 1.21 1.17 1.36 1.08 0.36 X6525 16 1.09 1.40 0.68 1.25 0.86 X6368 51 1.26 1.22 1.18 1.40 0.14 X6526 13 1.25 0.95 1.93 0.98 0.46 X6371 9 2.06 1.72 1.20 3.82* 0.03 X6527 9 1.09 0.68 1.00 1.74 0.53 X6372 17 1.05 0.79 1.74 0.65 0.89 X6528 18 1.12 1.54 0.48 1.45 0.74 X6373 15 1.11 0.65 0.85 1.96 0.35 X6529 24 1.11 1.22 1.08 0.99 0.79 X6374 14 1.03 0.65 0.85 1.71 0.54 X6536 13 1.14 0.79 1.46 1.16 0.56 X6376 24 1.19 0.85 1.45 1.19 0.37 X6537 17 1.25 0.58 1.62 1.71 0.17 X6377 19 1.19 0.59 1.31 1.66 0.24 X6538 11 1.39 0.76 1.10 2.37 0.14 X6379 9 1.81 2.04 0.75 2.43 0.13 X6539 25 0.81 1.08 1.15 0.27* 0.11 X6382 188 1.09 1.09 1.18 1.00 0.50 X6540 14 1.17 1.07 1.34 1.13 0.60 X6386 18 1.12 0.58 1.26 1.53 0.37 X6542 16 1.09 1.33 0.91 1.00 0.93 X6388 18 1.15 0.54 1.08 1.90 0.22 X6543 13 1.24 1.18 0.70 2.10 0.32 X6391 22 1.29 0.71 1.27 1.89 0.11 X6551 18 1.29 0.40 1.99 1.61 0.13 X6393 86 1.26 1.51* 1.14 1.11 0.21 X6557 15 1.30 1.05 1.69 1.20 0.36 X6396 22 1.19 0.68 1.11 1.81 0.20 X6559 9 1.07 0.53 2.55 0.42 0.81 X6407 60 0.95 0.98 0.90 0.94 0.67 X6564 14 1.21 0.85 1.76 1.15 0.44 X6423 17 0.67 0.88 0.52 0.63 0.10 X6567 12 1.09 1.11 1.46 0.73 0.95 X6430 399 1.25** 1.33** 1.22 1.19 0.02 X6569 11 1.45 0.86 1.25 2.16 0.14 X6432 44 1.22 1.11 1.19 1.36 0.20 X6572 68 1.27 1.24 1.32 1.25 0.11 X6434 13 1.03 1.58 0.25 1.22 0.91 X6582 14 1.18 0.93 1.26 1.40 0.45 X6449 16 1.22 1.78 0.58 1.42 0.62 X6583 9 1.49 1.28 3.37* 0.58 0.42 X6456 38 0.86 0.98 0.88 0.72 0.30 X6586 12 0.81 0.87 0.86 0.70 0.46 X6463 124 1.05 0.98 0.92 1.22 0.43 X6588 26 1.14 0.86 1.51 1.02 0.51 X6468 12 0.81 0.23 1.07 1.02 0.81 X6596 •88 1.07 0.93 0.87 1.40 0.27 X6472 12 0.82 1.13 0.60 0.71 0.41 X6597 34 1.43 1.51 1.56 1.20 0.14 X6475 106 1.11 1.52** 0.96 0.80 0.81 X6599 45 1.12 1.38 1.45 0.56 0.94 X6492 12 1.13 0.77 1.78 0.90 0.66 X6602 45 1.12 1.28 0.92 1.19 0.57 X6493 448 1.20** 1.25* 1.20 1.15 0.04 X6603 353 1.16* 1.20 1.13 1.17 0.07 X6494 34 1.15. 0.70 2.22** 0.61 0.53 X6619 388 1.10 1.06 1.14 1.10 0.19 X6499 623 1.21** 1.15 1.18 1.29** <.01 X6620 235 1.11 1.10 1.05 1.17 0.19 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal X6624 179 1.13 1.09 0.99 1.29 0.12 X6903 13 0.93 0.56 0.67 1.62 0.74 X6628 106 1.21 1.48* 1.11 1.03 0.34 X6904 9 1.33 1.78 1.34 0.87 0.70 X6629 470 1.11 1.11 1.06 1.14 0.15 X6905 14 1.03 0.59 1.12 1.45 0.57 X6648 41 1.30 1.06 1.77* 1.09 - 0.15 X6911 177 1.14 1.22 1.09 1.10 0.28 X6652 15 0.99 1.28 0.53 1.21 0.95 X6912 83 1.29* 1.56* 1.20 1.10 0.18 X6658 10 0.75 0.75 0.23 1.26 0.62 X6913 33 0.90 0.81 0.96 0.90 0.64 X6663 28 1.06 1.05 1.25 0.86 0.89 X6921 56 1.18 1.15 1.04 1.36 0.22 X6678 99 1.29* 1.23 1.29 1.35 0.03 X6953 25 1.19 1.76 0.46 1.32 0.66 X6693 551 1.21** 1.34** 1.19 1.10 0.12 X6966 19 1.29 0.82 2.51** 0.58 0.44 X6695 12 0.98 0.90 1.00 1.05 0.98 X6967 32 0.83 1.07 0.42 0.71 0.19 X6732 15 1.09 0.59 1.04 1.87 - 0.37 X6970 366 1.12 1.11 1.23 1.01 0.27 X6749 14 1.03 0.59 1.12 1.45 0.57 X6990 51 1.18 0.74 1.53 1.23 0.17 X6756 39 1.13 1.12 1.44 0.84 0.71 X6998 12 1.18 1.61 1.37 0.36 0.99 X6757 10 1.01 2.18 0.46 0.65 0.55 X7020 149 1.21 1.17 1.19 1.27 0.05 X6765 12 1.00 0.54 0.55 2.36 0.46 X7030 13 2.03* 2.86* 2.75* 0.48 0.18 X6766 17 0.94 1.54 0.55 0.61 0.39 X7040 13 0.76 1.13 0.58 0.52 0.21 X6767 14 1.26 1.40 0.98 1.45 0.46 X7044 9 2.02 1.35 3.46* 1.32 0.09 X6769 58 1.07 1.00 1.11 1.11 0.58 X7080 329 1.20* 1.35** 1.12 1.13 0.09 X6773 24 1.08 0.65 1.50 1.14 0.52 X7092 10 1.05 0.70 1.50 0.91 0.83 X6777 14 0.93 1.01 0.90 0.88 0.76 X7100 12 1.07 0.51 1.57 1.16 0.61 X6785 38 1.35 1.76 1.34 0.97 0.34 X7108 34 1.20 1.05 1.07 1.46 0.24 X6786 80 1.13 1.25 1.06 1.06 0.53 X7123 22 0.92 0.74 0.93 1.09 0.93 X6791 22 1.40 1.90 0.96 1.29 0.33 X7128 58 1.03 0.99 1.16 0.99 0.86 X6800 432 1.17* 1.14 1.07 1.31** <.01 X7153 10 1.99 1.89 2.78 1.26 0.11 X6801 9 0.87 2.34 0.56 0.00 0.15 X7157 25 1.39 1.60 1.22 1.33 0.24 X6804 288 1.12 1.11 0.99 1.25* 0.09 X7161 9 1.25 2.46 0.75 0.75 0.98 X6806 13 1.30 1.40 1.02 1.58 0.40 X7163 12 1.08 1.72 0.60 0.79 0.80 X6819 52 1.34 1.13 1.27 1.62 0.03 X7203 316 1.21** 1.23 1.24 1.18 0.02 X6858 9 1.31 1.68 1.46 0.81 0.74 X7204 156 1.12 0.92 1.26 1.19 0.12 X6859 26 1.03 0.80 1.58 0.72 0.99 X7211 47 0.89 0.76 0.97 0.94 0.64 X6864 11 1.67 1.84 1.51 1.65 0.18 X7233 12 1.12 1.09 1.17 1.08 0.76 X6874 24 0.68 0.94 0.60 0.51 0.05 X7255 234 1.14 1.15 1.11 1.16 0.15 X6880 20 0.88 0.64 0.52 1.50 0.93 X7312 139 1.21 1.31 1.17 1.14 0.15 X6881 89 1.12 1.09 0.91 1.35 0.24 X7313 543 1.21** 1.18 1.20 1.24* <.01 X6891 15 0.73 1.27 0.49 0.48 0.12 X7314 41 0.97 0.72 0.92 1.25 0.75 X6895 9 1.33 1.78 1.34 0.87 0.70 X7315 52 1.23 1.67* 1.08 0.96 0.54 X6898 52 0.96 0.99 1.17 0.78 0.61 X7316 179 1.10 1.17 1.08 1.04 0.50 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal X7320 48 1.09 1.19 1.07 1.03 0.72 X7648 18 1.12 0.76 0.86 1.83 0.35 X7327 214 1.13 1.34* 1.17 0.89 0.66 X7651 50 1.23 1.31 0.99 1.40 0.20 X7328 147 1.22* 1.22 1.27 1.17 0.08 X7669 459 1.10 1.06 1.09 1.15 0.13 X7348 58 1.16 1.09 1.41 0.97 0.44 X7885 56 1.18 0.90 1.38 1.25 0.18 X7385 39 1.30 1.06 1.37 1.45 0.11 X7911 19 1.07 0.77 1.10 1.37 0.52 X7390 9 0.84 0.39 2.74 0.52 0.78 X7917 25 1.22 1.61 1.45 0.60 0.82 X7393 9 0.84 0.39 2.74 0.52 0.78 X7937 47 1.14 1.53 1.03 0.88 0.87 X7397 147 1.09 0.99 1.35* 0.94 0.50 X7942 181 1.15 1.20 1.14 1.11 0.21 X7407 19 1.07 1.97 0.83 0.57 0.58 X7949 16 1.31 0.78 1.45 1.66 0.20 X7411 11 2.34* 2.55 2.29 2.24 0.03 X7972 10 1.38 1.45 1.96 0.76 0.55 X7422 16 0.89 0.78 1.02 0.88 0.74 X7988 24 0.80 0.72 0.58 1.12 0.54 X7432 99 1.10 1.07 1.14 1.08 0.44 X7989 38 0.95 0.53 0.79 1.56 0.52 X7433 29 1.06 1.54 0.11* 1.46 0.88 X7990 11 1.07 1.01 0.35 1.78 0.60 X7440 9 0.87 0.77 1.34 0.56 0.64 X8158 127 1.07 1.26 0.89 1.06 0.76 X7442 318 1.16* 1.22 1.19 1.09 0.13 X8176 13 0.72 1.19 0.36 0.51 0.13 X7443 220 1.20* 1.17 1.35* 1.09 0.07 X8178 21 1.22 0.99 1.81 0.95 0.46 X7445 308 1.14 1.15 1.01 1.27* 0.05 X8187 10 1.01 2.18 0.46 0.65 0.55 X7456 13 0.74 0.55 1.04 0.62 0.36 X8191 21 0.85 1.34 0.60 0.54 0.20 X7465 230 1.04 1.16 0.99 0.96 0.95 X8193 19 1.15 1.50 0.72 1.17 0.77 X7480 37 1.33 0.84 1.71 1.37 0.08 X8195 13 0.91 1.73 0.66 0.44 0.34 X7486 52 0.96 0.99 1.17 0.78 0.61 X8196 13 0.92 1.85 0.64 0.45 0.36 X7507 411 1.23** 1.13 1.33** 1.22* <.01 X8203 57 1.03 0.86 1.22 1.00 0.75 X7514 18 1.06 1.01 1.58 0.62 1.00 X8206 11 1.34 0.88 1.67 1.37 0.34 X7515 41 1.12 1.10 1.25 0.99 0.62 X8232 13 0.95 1.47 0.40 0.98 0.66 X7516 252 1.17 1.22 1.05 1.22 0.08 X8237 15 0.97 0.90 0.81 1.19 0.96 X7540 18 0.74 1.62 0.13* 0.47 0.05 X8241 62 1.06 1.33 0.86 0.98 0.97 X7544 14 1.00 0.95 1.23 0.85 0.96 X8247 26 0.78 1.11 0.81 0.40 0.09 X7556 16 0.95 0.91 0.49 1.48 0.88 X8248 21 0.85 0.60 1.07 0.88 0.65 X7578 11 1.41 2.32 0.78 1.14 0.62 X8251 14 1.12 1.94 0.80 0.72 0.82 X7579 15 1.38 1.02 2.54* 0.93 0.34 X8256 83 1.04 1.05 1.15 0.94 0.86 X7592 14 1.03 0.59 1.12 1.45 0.57 X8257 26 0.80 0.86 1.08 0.41 0.18 X7610 380 1.11 1.21 1.13 1.00 0.47 X8258 68 1.00 1.11 1.08 0.82 0.70 X7611 168 1.27* 1.67** 1.33 0.83 0.34 X8261 12 1.08 2.26 0.43 0.82 0.71 X7620 336 1.24** 1.12 1.34** 1.26* <.01 X8264 35 0.95 1.42 0.71 0.73 0.40 X7621 172 1.03 0.96 1.09 1.02 0.72 X8267 13 1.02 1.89 0.46 0.69 0.57 X7639 85 1.24 1.49* 1.15 1.06 0.30 X8272 76 1.31* 1.14 1.30 1.52 0.02 X7643 361 1.22** 1.45** 1.08 1.14 0.08 X8286 201 1.13 1.25 1.00 1.13 0.31 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal X8287 151 1.22* 1.50** 1.21 0.93 0.36 X8654 73 1.00 0.94 1.26 0.80 0.84 X8299 73 1.03 1.17 0.94 0.96 0.95 X8656 28 1.24 1.25 1.00 1.48 0.28 X8309 184 1.09 1.16 1.25 0.88 0.70 X8658 283 1.20* 1.06 1.41** 1.13 0.02 X8317 95 0.96 1.15 1.09 0.66 0.30 X8708 14 1.32 2.03 1.11 0.93 0.67 X8319 ' 22 1.26 1.39 1.29 1.10 0.45 X8711 9 0.45* 0.67 0.40 0.19 0.02 X8321 9 1.25 0.76 1.74 1.34 0.43 X8712 59 0.99 0.98 1.11 0.92 0.90 X8322 10 0.76 0.26 0.60 1.43 0.86 X8713 60 0.95 0.96 0.96 0.91 0.67 X8335 75 1.37* 1.26 1.26 1.60* 0.01 X8716 11 1.31 1.16 1.15 1.70 0.35 X8345 134 1.08 1.02 1.03 1.20 0.32 X8741 11 1.10 1.20 0.98 1.10 0.83 X8350 81 1.04 0.93 1.12 1.06 0.65 X8742 90 1.22 1.30 1.49* 0.88 0.33 X8532 309 1.18* 1.13 1.24 1.18 0.03 X8745 12 1.13 0.51 0.71 2.16 0.26 X8533 388 1.09 1.17 1.12 1.00 0.55 X8749 82 1.00 0.99 1.05 0.96 0.95 X8536 164 1.11 1.19 0.88 1.26 0.25 X8750 15 1.59 1.51 2.13 1.20 0.18 X8545 9 1.35 1.89 0.95 1.23 0.60 X8752 11 1.05 1.20 0.85 1.11 0.92 X8553 397 1.11 1.12 0.96 1.23* 0.10 X8793 52 0.92 0.99 1.13 0.71 0.40 X8559 62 0.91 0.96 1.07 0.72 0.34 X8798 14 1.13 1.63 0.92 0.91 0.96 X8563 9 1.26 1.65 2.17 0.00 0.96 X8833 11 1.12 0.93 1.32 1.12 0.69 X8564 20 1.46 1.45 1.30 1.62 0.15 X8843 11 1.07 0.92 1.19 1.11 0.78 X8569 72 1.04 0.98 0.79 1.34 0.48 X8859 10 0.75 0.26 0.62 1.28 0.77 X8570 87 1.17 1.03 1.33 1.17 0.18 X8862 230 1.21* 1.31* 1.13 1.17 0.08 X8571 423 1.17* 1.05 1.32** 1.13 0.03 X8864 156 1.11 1.35* 0.95 1.02 0.69 X8573 89 1.17 1.16 1.35 0.98 0.32 X8865 41 1.13 1.19 1.03 1.16 0.55 X8574 146 1.13 1.02 1.32 1.06 0.23 X8866 24 0.84 0.71 0.93 0.90 0.57 X8577 9 1.10 1.26 0.73 1.36 0.80 X8867 149 1.29* 1.47* 1.10 1.30 0.03 X8582 44 0.79 1.03 0.81 0.56 0.07 X8868 372 1.25** 1.37** 1.10 1.28* <.01 X8583 213 1.06 0.98 1.10 1.10 0.39 X8923 14 1.38 0.83 1.56 1.80 0.16 X8585 18 2.03** 3.59** 0.70 1.98 0.06 X8962 25 1.37 1.35 1.74 0.95 0.28 X8586 12 1.13 0.51 0.71 2.16 0.26 X8963 84 1.09 1.18 1.00 1.08 0.61 X8593 60 0.95 1.01 0.87 0.94 0.65 X8964 73 1.04 1.13 1.14 0.84 0.88 X8594 13 1.14 1.06 1.30 1.05 0.70 X8975 24 0.78 0.20* 1.04 1.09 0.75 X8596 14 1.12 1.00 1.40 0.93 0.76 X8978 .93 0.97 0.90 0.89 1.10 0.98 X8645 173 1.17 1.39* 1.08 1.03 0.36 X8981 175 1.06 0.84 1.23 1.10 0.29 X8646 71 0.98 0.96 0.95 1.04 0.98 X8982 27 1.00 1.29 0.97 0.78 0.72 X8647 21 1.07 0.83 1.18 1.21 0.62 X8987 334 1.25** 1.20 1.20 1.35** <.01 X8648 20 1.00 0.93 0.95 1.13 0.91 X8988 44 1.20 0.84 1.83* 0.90 0.31 X8651 50 1.13 0.81 1.18 1.44 0.19 X8993 15 1.63 0.89 2.75* 1.34 0.08 X8652 45 0.91 1.04 0.66 1.07 0.63 X8996 294 1.19* 1.25 1.08 1.23 0.04 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal X9001 52 1.19 1.24 1.23 1.09 0.38 X9100 11 1.05 1.83 0.26 0.95 0.74 X9011 13 1.49 1.02 1.55 1.86 0.14 X9101 9 1.32 0.75 4.12** 0.00 0.68 X9012 24 1.22 1.52 0.69 1.54 0.40 X9102 12 1.07 1.80 0.22 1.30 0.93 X9014 13 0.99 0.87 1.02 1.04 0.98 X9106 11 1.05 1.83 0.26 0.95 0.74 X9015 20 1.28 0.63 1.91 1.26 0.22 X9107 12 1.07 1.87 0.91 0.54 0.69 X9018 224 0.94 1.10 0.97 0.76* 0.12 X9126 358 1.11 0.92 1.26* 1.15 0.05 X9019 16 1.20 0.71 2.15* 0.68 0.56 X9133 9 0.86 1.14 0.79 0.63 0.53 X9021 87 1.21 1.22 1.05 1.34 0.12 X9143 10 2.30* 2.11 3.91* 0.73 0.08 X9022 99 1.05 0.96 1.03 1.17 0.47 X9153 132 1.04 0.90 1.12 1.11 0.46 X9023 199 1.10 1.12 1.05 1.13 0.30 X9181 274 1.16 1.23 1.20 1.06 0.17 X9024 374 1.11 1.19 1.24* 0.92 0.58 X9182 12 1.17 1.08 1.62 0.72 0.73 X9025 61 0.93 1.06 1.00 0.71 0.33 X9188 153 1.11 1.01 1.17 1.13 0.24 X9027 72 0.98 1.05 1.01 0.86 0.68 X9191 9 2.96** 1.22 5.07** 1.33 0.01 X9028 152 1.11 1.33* 0.95 1.01 0.70 X9196 12 1.13 0.51 0.71 2.16 0.26 X9029 160 1.06 0.82 1.31 1.06 0.31 X9197 15 1.00 0.63 1.17 1.20 0.76 X9030 191 1.21* 1.23 1.17 1.22 0.05 X9199 15 1.13 0.47 1.32 1.58 0.35 X9031 50 1.36 0.90 1.95** 1.26 0.04 X9200 84 1.11 1.37 0.84 1.16 0.60 X9032 68 0.98 1.05 1.08 0.82 0.65 X9201 12 0.98 0.27 0.81 1.69 0.54 X9035 71 0.99 1.20 0.78 0.93 0.64 X9202 33 0.98 1.40 0.77 0.81 0.56 X9036 16 1.15 1.73 0.66 1.05 0.91 X9203 25 1.41 1.57 1.58 1.00 0.27 X9045 20 1.10 1.86 0.52 1.05 0.97 X9216 25 1.32 1.00 1.48 1.47 0.16 X9052 71 0.99 1.22 0.78 0.93 0.65 X9217 351 1.20** 1.30* 1.18 1.13 0.06 X9053 71 0.98 1.22 0.78 0.90 0.58 X9218 38 1.06 0.82 1.04 1.35 0.45 X9059 69 1.01 1.19 0.86 0.92 0.74 X9244 23 1.35 2.00 1.35 0.80 0.58 X9062 26- 1.22 0.66 2.27** 0.92 0.31 X9248 132 1.21 1.37 1.25 1.02 0.25 X9071 38 1.15 1.05 1.26 1.13 0.44 X9249 276 1.15 1.33* 0.92 1.20 0.18 X9073 9 0.66 0.66 0.29 0.96 0.36 X9252 13 0.90 0.47 1.29 0.87 0.91 X9075 14 1.25 1.33 1.08 1.33 0.49 X9256 239 1.24** 1.20 1.16 1.34* <.01 X9076 69 1.00 0.97 1.03 0.99 1.00 X9258 371 1.13 1.18 1.03 1.19 0.11 X9078 397 1.27** 1.11 1.30* 1.39** <.01 X9259 A X9260 158 1.13 1.09 1.01 1.28 0.14 X9079 291 1.22** 1.20 1.28* 1.17 0.02 249 1.02 1.10 0.98 0.99 0.98 X9081 43 1.07 1.16 1.06 0.99 0.85 X9266 18 1.03 1.64 1.28 0.29 0.46 X9082 31 1.43 0.75 1.67 1.84 0.03 X9270 39 1.07 1.00 1.03 1.20 0.59 X9085 9 1.06 1.69 1.00 0.61 0.77 X9271 153 1.15 1.19 1.23 1.03 0.29 X9088 148 1.02 0.92 1.11 1.04 0.67 X9273 137 1.13 1.22 1.14 1.05 0.39 X9092 235 1.14 1.27 1.03 1.12 0.25 X9275 118 1.11 1.04 1.29 1.02 0.37 X9093 34 1.11 0.83 1.21 1.29 0.40 X9277 9 1.70 1.28 0.48 3.61* 0.06 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal X9279 9 1.70 1.28 0.48 3.61* 0.06 X9391 128 1.10 1.01 1.35 0.95 0.47 X9284 182 1.17 1.17 1.22 1.11 0.14 X9394 13 0.99 1.75 0.00 0.94 0.58 X9285 22 0.81 1.10 0.56 0.80 0.30 X9397 72 1.04 0.91 1.02 1.21 0.50 X9286 29 1.11 0.96 1.55 0.81 0.72 X9398 15 1.44 1.73 0.33 2.10 0.21 X9287 43 0.93 1.05 0.89 0.86 0.58 X9399 21 0.68 0.89 0.58 0.60 0.08 X9288 286 1.12 1.22 1.14 1.00 0.42 X9401 14 1.22 1.80 1.44 0.46 0.95 X9289 20 1.26 0.87 1.34 1.59 0.21 X9402 13 1.11 1.60 1.07 0.56 0.83 X9290 55 1.35 1.72* 1.30 1.02 0.23 X9403 50 1.20 0.96 1.34 1.27 0.18 X9291 21 1.06 1.29 0.98 0.86 0.93 X9407 9 1.26 1.65 2.17 0.00 0.96 X9293 164 1.13 0.96 1.34* 1.09 0.16 X9411 106 1.14 1.24 1.16 1.02 0.44 X9296 573 1.26** 1.25* 1.31** 1.21* <.01 X9412 171 1.17 1.40* 1.18 0.95 0.42 X9298 20 0.77 0.89 0.85 0.61 0.24 X9414 18 1.23 0.75 1.34 1.68 0.23 X9300 122 1.16 1.26 1.05 1.17 0.25 X9417 44 1.31 1.36 1.14 1.44 0.13 X9305 146 1.23* 1.45* 1.11 1.15 0.14 X9423 63 0.88 0.99 0.95 0.71 0.22 X9316 248 1.09 1.01 1.21 1.06 0.25 X9424 12 1.13 0.51 0.71 2.16 0.26 X9321 64 1.02 1.12 0.90 1.04 0.98 X9435 13 1.12 0.52 1.57 1.23 0.50 X9322 9 0.94 0.90 1.20 0.68 0.80 X9437 388 1.21** 1.18 1.27* 1.17 0.02 X9323 72 0.99 0.85 1.25 0.86 0.93 X9444 20 1.28 0.81 1.46 1.70 0.16 X9326 72 0.99 0.85 1.25 0.86 0.93 X9446 48 1.17 0.86 1.66* 0.96 0.32 X9327 18 1.01 1.53 1.07 0.47 0.52 X9447 17 0.81 0.74 0.93 0.76 0.46 X9328 226 1.11 1.11 1.13 1.09 0.25 X9448 17 0.81 0.74 0.93 0.76 0.46 X9332 18 1.01 0.52 1.41 1.11 0.71 X9454 136 1.07 0.83 1.27 1.09 0.32 X9333 262 1.31** 1.19 1.50** 1.25 <.01 X9455 118 1.17 1.11 1.19 1.21 0.14 X9336 95 1.04 1.22 1.03 0.88 0.86 X9462 74 1.08 1.10 1.04 1.10 0.59 X9343 60 1.17 1.52 1.09 0.94 0.64 X9464 170 1.12 1.06 1.06 1.23 0.16 X9347 46 1.13 1.15 1.34 0.90 0.67 X9465 17 0.89 0.90 1.04 0.79 0.65 X9350 24 1.24 1.08 1.18 1.43 0.28 X9466 29 1.26 0.96 1.40 1.46 0.17 X9352 217 1.12 1.28 1.04 1.05 0.45 X9472 70 1.27 1.24 1.05 1.54* 0.06 X9363 53 1.15 1.00 1.36 1.07 0.37 X9474 18 0.66 0.60 1.04 0.27 0.08 X9364 157 1.08 0.93 1.26 1.03 0.39 X9475 12 1.10 0.29 1.14 2.00 0.34 X9365 110 1.11 1.28 0.97 1.07 0.60 X9481 15 1.24 1.55 1.32 0.95 0.68 X9374 11 1.37 0.90 1.64 1.74 0.23 X9489 14 1.15 0.49 0.93 2.11 0.23 X9375 179 1.10 0.96 1.19 1.14 0.19 X9502 23 1.06 0.96 0.93 1.33 0.63 X9380 45 1.38 1.18 1.36 1.64 0.04 X9503 9 1.17 0.88 2.23 0.48 0.75 X9383 33 1.22 1.01 1.28 1.37 0.23 X9504 60 1.27 1.52 1.50 0.80 0.40 X9385 24 1.41 1.46 1.12 1.57 0.16 X9515 68 1.20 1.41 1.18 1.01 0.44 X9390 9 0.75 0.52 0.68 1.09 0.65 X9516 50 1.23 1.50 1.23 0.95 0.50 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal X9518 45 1.21 0.98 1.51 1.23 0.20 X9686 34 1.50* 2.11* 0.89 1.44 0.14 X9521 15 0.60 0.66 0.78 0.30 0.05 X9687 30 1.10 1.00 1.57 0.75 0.81 X9531 149 1.06 1.49** 1.02 0.67* 0.36 X9689 18 1.10 0.54 1.42 1.37 0.42 X9536 12 0.74 0.24 0.82 1.05 0.63 X9692 348 1.15 1.10 1.20 1.14 0.07 X9541 31 1.39 1.40 1.42 1.36 0.14 X9701 9 0.94 0.88 0.33 1.63 0.88 X9542 22 1.17 0.80 1.57 1.13 0.42 X9719 214 1.07 1.03 1.09 1.10 0.37 X9547 10 0.84 1.12 0.55 0.79 0.50 X9722 201 1.31** 1.35* 1.46** 1.14 0.01 X9564 10 1.33 0.00 1.78 2.64 0.11 X9730 28 0.97 0.94 1.34 0.65 0.70 X9566 38 1.08 1.11 1.15 0.97 0.80 X9731 10 1.01 2.18 0.46 0.65 0.55 X9571 12 1.33 1.06 1.44 1.49 0.33 X9773 12 1.13 0.51 0.71 2.16 0.26 X9572 14 1.16 1.03 1.24 1.21 0.57 X9777 12 1.13 0.51 0.71 2.16 0.26 X9574 14 1.11 0.25 0.71 2.39* 0.19 X9786 73 0.99 1.05 0.90 1.02 0.91 X9576 39 0.92 0.97 0.87 0.93 0.63 X9788 12 0.87 1.40 0.66 0.46 0.34 X9578 74 1.04 0.95 0.77 1.38 0.43 X9791 65 1.12 1.07 1.08 1.21 0.38 X9582 18 1.13 0.63 1.11 1.58 0.37 X9792 32 1.23 1.60 1.05 1.00 0.59 X9588 11 1.08 0.58 1.13 1.54 0.53 X9795 18 1.49 1.89 1.78 0.98 0.35 X9593 77 0.93 0.94 0.78 1.05 0.69 X9797 35 0.88 0.79 1.11 0.76 0.51 X9597 49 1.23 1.58 1.54 0.63 0.74 X9800 32 1.07 1.60 0.82 0.79 0.76 X9601 14 1.17 0.65 1.04 2.07 0.27 X9801 251 1.15 1.23 1.03 1.19 0.14 X9602 45 1.12 1.15 1.01 1.21 0.51 X9845 55 1.31 1.58 1.07 1.29 0.17 X9604 16 1.35 1.83 1.09 1.15 0.50 X9854 39 0.92 0.52 1.08 1.22 0.78 X9605 25 0.88 1.01 0.64 0.99 0.58 X9857 10 0.60 0.39 0.61 0.81 0.24 X9606 109 1.14 1.04 1.17 1.21 0.19 X9875 15 1.12 0.51 1.53 1.24 0.49 X9607 86 1.01 0.86 1.04 1.12 0.69 X9878 179 1.05 1.04 0.94 1.17 0.43 X9610 16 0.81 0.36 1.69 0.71 0.67 X9879 100 1.01 1.01 0.93 1.07 0.90 X9620 48 1.33 1.50 1.30 1.19 0.17 X9880 170 1.27* 1.08 1.59** 1.11 0.02 X9623 21 1.16 1.13 1.23 1.12 0.58 X9881 156 1.14 0.98 1.15 1.28 0.07 X9628 138 1.08 0.92 1.16 1.17 0.26 X9891 10 0.75 0.75 1.07 0.43 0.34 X9638 99 1.20 1.08 1.13 1.39 0.07 X9893 193 1.33** 1.29 1.58** 1.13 <.01 X9641 16 1.14 2.41* 0.44 0.72 0.69 X9894 175 1.38** 1.47** 1.62** 1.06 0.01 X9643 20 1.26 0.59 1.99 1.60 0.14 X9895 160 1.40** 1.42* 1.61** 1.18 <.01 X9646 60 1.32 1.41 1.62* 0.91 0.20 X9898 70 1.54** 1.97** 1.57* 1.09 0.03 X9649 9 0.97 0.69 0.65 1.51 0.79 X9902 73 1.13 1.28 0.85 1.27 0.41 X9652 81 1.17 0.86 1.63** 0.98 0.22 X9903 43 0.77 0.84 0.95 0.53 0.07 X9656 47 1.27 1.41 1.16 1.25 0.24 X9912 25 1.41 1.57 1.58 1.00 0.27 X9666 14 1.09 1.81 0.63 1.03 0.97 X9914 26 1.05 1.02 0.86 1.26 0.71 X9685 96 1.35* 1.48 1.19 1.39 0.03 X9916 146 1.14 1.07 1.27 1.10 0.19 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal X9918 93 1.10 1.26 1.04 1.02 0.65 Y1006 25 3.11** 2.70** 2.46 4.12** <.01 X9920 173 1.31** 1.06 1.26 1.61** <.01 Y1012 124 1.03 0.90 1.18 1.02 0.63 X9921 12 1.51 2.29 1.32 1.09 0.43 Y1013 236 1.27** 1.21 1.52** 1.09 0.01 X9922 78 1.24 1.19 1.23 1.30 0.10 Y1014 489 1.22** 1.20 1.24* 1.23* <.01 X9923 11 2.41* 3.20* 0.73 3.04 0.03 Y1016 479 1.19* 1.37** 1.13 1.08 0.20 X9925 20 1.25 2.14* 1.06 0.69 0.95 Y1018 66 1.05 1.13 1.09 0.93 0.92 X9926 11 1.13 0.97 1.28 1.15 0.67 Y1019 237 1.26** 1.20 1.56** 1.05 0.02 X9927 14 1.13 1.01 1.34 1.00 0.71 Y1020 609 1.23** 1.20 1.21* 1.28** <.01 X9928 12 1.07 1.36 0.84 1.00 0.99 Y1022 215 1.27** 1.12 1.12 1.58** <.01 X9929 13 1.11 0.76 1.42 1.18 0.60 Y1023 63 0.94 0.96 1.07 0.82 0.57 X9930 11 1.14 0.95 1.33 1.15 0.66 Y1024 280 1.15 1.38** 1.01 1.05 0.42 X9933 10 1.11 0.36 1.74 1.12 0.56 Y1026 234 1.28** 1.31* 1.39** 1.15 0.01 X9934 84 1.07 1.22 1.11 0.89 0.93 Y1028 22 0.92 1.41 0.82 0.59 0.38 X9936 116 0.99 1.00 0.82 1.14 0.91 Y1030 317 1.23** 1.25 1.16 1.28* 0.01 X9937 43 1.04 0.56 0.85 1.76* 0.20 Y1032 72 0.98 1.19 0.79 0.92 0.60 X9940 17 0.64 0.72 0.54 0.68 0.11 Y1034 275 1.18* 1.25 1.18 1.11 0.10 X9941 157 1.08 1.02 1.22 1.00 0.48 Y1036 15 1.15 0.64 1.47 1.41 0.40 X9944 19 0.99 0.99 0.98 0.99 0.96 Y1037 300 1.17* 1.22 1.20 1.08 0.13 X9945 9 1.63 2.64 1.06 1.39 0.38 Y1038 191 1.16 1.23 1.22 1.04 0.23 X9946 26 0.88 1.19 0.64 0.84 0.41 Y1040 510 1.17* 1.05 1.15 1.32** <.01 X9947 261 1.10 1.03 1.14 1.13 0.17 Y1041 29 1.37 0.70 1.79 1.63 0.05 X9948 99 1.02 1.04 0.98 1.04 0.87 Y1042 461 1.26** 1.33** 1.15 1.29* <.01 X9949 203 1.07 1.01 1.02 1.18 0.28 Y1043 203 1.28** 1.66** 0.99 1.19 0.08 X9950 27 1.24 1.01 1.24 1.45 0.24 Y1044 157 1.12 1.12 1.24 1.00 0.39 X9952 36 1.29 1.11 1.83* 0.94 0.25 Y1045 26 1.01 1.45 1.02 0.58 0.56 X9953 23 1.04 0.65 1.23 1.28 0.56 Y1046 234 1.27** 1.19 1.12 1.51** <.01 X9956 19 1.04 0.47 1.14 1.56 0.41 Y1047 260 1.21* 1.18 1.23 1.21 0.02 X9958 92 1.31* 1.29 1.11 1.52* 0.02 Y1049 17 1.01 0.53 1.07 1.47 0.57 X9962 9 0.97 1.37 1.52 0.00 0.49 Y1050 269 1.16 0.92 1.39** 1.17 0.02 X9971 49 1.26 1.62* 1.15 0.99 0.44 Y1051 444 1.21** 1.21 1.29** 1.14 0.02 X9975 135 1.09 1.23 1.04 0.98 0.74 Y1053 149 1.12 1.01 1.20 1.15 0.19 X9977 102 1.31* 1.48* 1.30 1.14 0.08 Y1054 179 1.09 0.90 1.04 1.34* 0.09 X9978 12 1.08 2.09 0.53 0.72 0.70 Y1055 52 0.96 0.99 1.17 0.78 0.61 X9984 74 1.01 1.24 0.83 0.97 0.82 Y1056 154 1.16 0.91 1.32 1.24 0.05 xxxxx 15 0.68 0.67 1.03 0.36 0.13 Y1057 27 1.45 0.80 1.70 1.89 0.03 Y0005 28 1.57* 2.53** 0.85 1.39 0.18 Y1058 26 1.06 0.91 1.44 0.81 0.88 Y1000 407 1.15* 1.12 1.27* 1.08 0.09 Y1059 475 1.18* 1.26* 1.11 1.17 0.06 NIOSH Cases Ever Low Medium High Ordinal NIOSH Cases Ever Low Medium High Ordinal Y1061 119 1.03 1.02 1.21 0.86 0.97 Z0267 11 0.69 0.38 0.66 1.08 0.52 Y1062 127 1.15 1.39* 1.08 0.97 0.57 Z0475 33 1.39 2.03* 1.02 1.06 0.39 Y1064 47 1.08 0.94 0.99 1.31 0.43 Z0477 26 1.68* 2.20* 1.71 1.05 0.12 Y1066 148 1.12 0.86 1.39* 1.10 0.15 Z0482 138 1.09 1.24 1.12 0.91 0.82 Y1067 146 1.09 1.09 1.22 0.95 0.57 Z0483 218 1.25** 1.37* 1.21 1.16 0.05 Y1068 11 1.47 0.36 1.53 2.74* 0.06 Z0495 183 1.15 1.20 1.00 1.25 0.12 Y1069 232 1.21* 1.15 1.25 1.21 0.03 Z0496 94 1.16 0.85 1.41 1.24 0.10 Y1070 195 1.32** 1.34* 1.28 1.36* <.01 Z0547 42 1.01 1.09 0.81 1.20 0.87 Y1071 202 1.28** 1.36* 1.25 1.22 0.02 Z0583 60 1.03 1.14 1.07 0.88 0.91 Y1072 232 1.02 1.06 1.04 0.97 1.00 Z0599 16 1.11 1.37 0.80 1.16 0.82 Y1074 169 1.07 0.92 1.18 1.09 0.33 Z0660 13 0.71 1.41 0.50 0.17 0.06 Y1079 34 0.70 0.78 0.77 0.55 0.04 Z0673 12 0.60 0.89 0.39 0.58 0.09 Y1080 124 1.11 1.25 1.01 1.07 0.51 Z0701 53 0.95 0.99 1.17 0.78 0.57 Y1081 171 1.17 1.26 1.02 1.24 0.12 Z0920 19 1.19 0.59 1.31 1.66 0.24 Y1083 96 1.34* 1.53* 1.23 1.26 0.05 Z0927 424 1.21** 1.16 1.28* 1.19 0.01 Y1085 62 1.00 1.22 1.13 0.68 0.55 Z0947 53 0.91 0.99 1.04 0.71 0.34 Y1086 75 1.21 1.45 1.12 1.04 0.39 Z1037 14 0.90 0.52 1.01 1.23 0.90 Y1087 193 1.16 1.24 1.06 1.17 0.16 Z1043 13 1.05 1.26 0.49 1.40 0.82 Y1090 12 1.32 1.65 2.29 0.28 0.87 Z1061 16 1.22 0.65 1.06 2.17 0.18 Y1092 9 1.05 1.30 0.67 1.20 0.95 Z1121 24 1.30 1.99* 0.86 0.92 0.73 Y1096 15 1.39 1.54 1.20 1.39 0.33 Z1122 17 1.23 0.57 1.46 1.88 0.17 Y1098 118 1.17 0.99 1.46* 1.14 0.11 Z3004 23 1.39 0.83 1.79 1.66 0.08 ZOOOO 30 1.07 1.40 1.25 0.61 0.78 Z3140 9 0.85 0.83 0.97 0.77 0.66 a Column Headings: NIOSH = NIOSH agent code. See NIOSH NOES website www.cdc.gov/noes/srch-noes.html to search for  NIOSH agent names and NIOSH agent codes. Cases = number of  bladder cancer cases exposed. Ever = odds ratio for  ever exposure. Low/Medium/High = odds ratio for  low/medium/high cumulative exposure relative to the non-exposed where groups are divided by tertiles of  the controls. Ordinal = p-value for  fitting  a line through the non-exposed, low, medium, and high exposure groups by assigning scores of  0, 1, 2, and 3 respectively. Odds Ratios in italics  represent a decreasing linear fit. * Significant  at a 5% alpha level ** Significant  at a 1% alpha level Appendix B Table B.l: NIOSH Agent Name Abbreviations Agent Name Abbreviation NIOSH Agent Name 1, 2-ETHANEDIAMINE, RP W C IB HP 2,5-PYRROLIDINEDIONE, 12AE MPIB D RP 2,5-PYRROLIDINEDIONE, 12AE MPIB D 2-BUTENEDIOIC ACID (E)-, PW 1,3-B EB 2-PROPENOIC ACID, 2M CEPWC2 ALANINE, 3-(P-(BIS(2-CE)A)P-, L- ALKENES, C15-18 ALPHA-, RPW SDP CS S BUTYRIC ACID, 4-(P-(B(2-CE)A)P)- ETHANOL, 2-(2-(2-BE)E)- ETHYLAMINE, 2-(P-(l, 2-D-l-B)P)-N,N-D-,(Z)- NICKEL CHLORIDE (NICL2) , HH N,N-BIS(2-CE)-2-NL (CHLORNAPHAZINE) NONYLPHENOL ETHYLENE OA PHENOL, DODECYL-, SULFURIZED, CCSO PHOSPHORODITHIOIC ACID, MOOB E ZS PHOSPHORODITHIOIC ACID, OOB(2E)E ZS PHOSPHORODITHIOIC ACID, OOZS PLUTONIUM, RADIOACTIVE E (NO) POC - GASOLINE (LEADED) POC - JET FUEL & GASOLINE, ULD PURINE, 6-((1-M-4-N-5-YL)THIO)- SOLVENT RD HVY PF DIST (PETROLEUM) SULFONIC ACIDS, PETROLEUM, CSO SULFONIC ACIDS, PETROLEUM, MS SULFURIC ACID, AMMONIUM N(2+) S(2:2:l) SULFURIC ACID, NICKEL(2+) SALT(1:1) , HH UREA, N-(2-CE)-N'-(4-MC)-N-NITROSO- 1, 2-ETHANEDI AMINE, REACTION PRODUCTS WITH CHLORINATED ISOBUTY- LENE HOMOPOLYMER 2,5-PYRROLIDINEDIONE, l-(2-((2-((2-((2-AMINOETHYL)AMINO)ETHYL)AMINO) ETHYL)AMINO)ETHYL)-, MONOPOLYISOBUTENYL DERIVS., REACTION PR 2,5-PYRROLIDINEDIONE, l-(2-((2-((2-((2-AMINOETHYL)AMINO)ETHYL)AMINO) ETHYL)AMINO)ETHYL)-, MONOPOLYISOBUTENYL DERIVS. 2-BUTENEDIOIC ACID (E)-, POLYMER WITH 1,3-BUTADIENE AND ETHENYLBEN- ZENE 2-PROPENOIC ACID, 2-ME-, C12 ESTER, POLY W/ C16 2ME2PROPENOATE, ISO- C10 2ME2PROPENOATE, ME 2ME2PROPENOATE, C18 2ME2PROPENOATE, C14 2ME2PROPENOATE ALANINE, 3-(P-(BIS(2-CHLOROETHYL)AMINO)PHENYL-, L- ' ALKENES, C15-18 ALPHA-, REACTION PRODUCTS WITH SULFURIZED DODE- CYLPHENOL CALCIUM SALT, SULFURIZED BUTYRIC ACID, 4-(P-(BIS(2-CHLOROETHYL)AMINO)PHENYL)- ETHANOL, 2-(2-(2-BUTOXYETHOXY) ETHOXY)- ETHYLAMINE, 2-(P-(l, 2-DIPHENYL-l-BUTENYL)PHENOXY)-N,N-DIMETHYL-,(Z)- NICKEL CHLORIDE (NICL2) , HEXAHYDRATE N,N-BIS(2-CHLOROETHYL)-2-NAPTHYLAMINE (CHLORNAPHAZINE) NONYLPHENOL ETHYLENE OXIDE ADDUCT PHENOL, DODECYL-, SULFURIZED, CARBONATES, CALCIUM SALTS, OVERBASED PHOSPHORODITHIOIC ACID, MIXED 0, 0-BIS(SEC-BU AND 1,3-DIMETHYLBUTYL) ESTERS, ZINC SALTS PHOSPHORODITHIOIC ACID, 0, O-BIS(2-ETHYLHEXYL) ESTER, ZINC SALT PHOSPHORODITHIOIC ACID, O-(2-ETHYLHEXYL) O-ISOBUTYL ESTER, ZINC SALT PLUTONIUM, RADIOACTIVE ELEMENT (NATURALLY OCCURING) PRODUCTS OF COMBUSTION - GASOLINE (LEADED) PRODUCTS OF COMBUSTION - JET FUEL AND GASOLINE, UNLEADED PURINE, 6-((l-METHYL-4-NITROIMIDAZOL-5-YL)THIO)- SOLVENT REFINED HEAVY PARAFFINIC DISTILLATE (PETROLEUM) SULFONIC ACIDS, PETROLEUM, CALCIUM SALTS, OVERBASED SULFONIC ACIDS, PETROLEUM, MAGNESIUM SALTS SULFURIC ACID, AMMONIUM NICKEL(2+) SALT (2:2:1) SULFURIC ACID, NICKEL(2+) SALT (1:1) , HEXAHYDRATE UREA, N-(2-CHLOROETHYL)-N'-(4-METHYLCYCLOHEXYL)-N-NITROSO-

Cite

Citation Scheme:

    

Usage Statistics

Country Views Downloads
China 22 27
United States 2 0
City Views Downloads
Beijing 22 0
Sunnyvale 1 0
Ashburn 1 0

{[{ mDataHeader[type] }]} {[{ month[type] }]} {[{ tData[type] }]}

Share

Share to:

Comment

Related Items