UBC Faculty Research and Publications

Paternal psychosocial work conditions and mental health outcomes: A case-control study Maggi, Stefania; Ostry, Aleck; Tansey, James; Dunn, James; Hershler, Ruth; Chen, Lisa; Hertzman, Clyde Mar 31, 2008

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12889_2007_Article_1074.pdf [ 537.12kB ]
JSON: 52383-1.0223092.json
JSON-LD: 52383-1.0223092-ld.json
RDF/XML (Pretty): 52383-1.0223092-rdf.xml
RDF/JSON: 52383-1.0223092-rdf.json
Turtle: 52383-1.0223092-turtle.txt
N-Triples: 52383-1.0223092-rdf-ntriples.txt
Original Record: 52383-1.0223092-source.json
Full Text

Full Text

ralssBioMed CentBMC Public HealthOpen AcceResearch articlePaternal psychosocial work conditions and mental health outcomes: A case-control studyStefania Maggi*1, Aleck Ostry2, James Tansey3, James Dunn4, Ruth Hershler3, Lisa Chen3 and Clyde Hertzman3Address: 1Department of Psychology and Institute of Interdisciplinary Studies, Carleton University, Colonel By Drive, Ottawa, Canada, 2Department of Geography, University of Victoria, Finnerty Road, Victoria, Canada, 3Department of Health Care and Epidemiology, University of British Columbia, Fairview Avenue, Vancouver, Canada and 4Department of Geography, University of Toronto, St. George Street, Toronto, CanadaEmail: Stefania Maggi* - stefania_maggi@carleton.ca; Aleck Ostry - ostry@uvic.ca; James Tansey - james.tansey@ubc.ca; James Dunn - jim.dunn@utoronto.ca; Ruth Hershler - hershler@interchange.ubc.ca; Lisa Chen - lisachen@interchange.ubc.ca; Clyde Hertzman - hertzman@interchange.ubc.ca* Corresponding author    AbstractBackground: The role of social and family environments in the development of mental healthproblems among children and youth has been widely investigated. However, the degree to whichparental working conditions may impact on developmental psychopathology has not beenthoroughly studied.Methods: We conducted a case-control study of several mental health outcomes of 19,833children of sawmill workers and their association with parental work stress, parental socio-demographic characteristics, and paternal mental health.Results: Multivariate analysis conducted with four distinct age groups (children, adolescents, youngadults, and adults) revealed that anxiety based and depressive disorders were associated withpaternal work stress in all age groups and that work stress was more strongly associated withalcohol and drug related disorders in adulthood than it was in adolescence and young adulthood.Conclusion: This study provides support to the tenet that being exposed to paternal work stressduring childhood can have long lasting effects on the mental health of individuals.BackgroundThe etiology and contributing factors of mental healthproblems are complex and multifaceted. Researchers haveinvestigated a broad range of factors that may contributeto the development of mental health problems such as afamily history of mental health problems and socio-demographic factors (e.g., ethnicity and marital status).There is growing evidence that ethnicity and other paren-tal characteristics may impact on child mental health out-comes through socialization practices (for a review seeHughes et al. [1]). For example, increasingly research isdocumenting that positive ethnic identity among minor-ity children and adolescents functions as a protective fac-tor for mental health and school outcomes, and it hasPublished: 31 March 2008BMC Public Health 2008, 8:104 doi:10.1186/1471-2458-8-104Received: 26 January 2007Accepted: 31 March 2008This article is available from: http://www.biomedcentral.com/1471-2458/8/104© 2008 Maggi et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.Page 1 of 10(page number not for citation purposes)been hypothesized that such protective effect is a functionBMC Public Health 2008, 8:104 http://www.biomedcentral.com/1471-2458/8/104of socialization processes occurring between parents andchildren [2-4].While research on how specific ethnic socialization proc-esses impact on child and adolescent outcomes is stillneeded, the role of the social and family environments inthe development of mental health problems among thenon-minority children and youth has been widely investi-gated. Theoretical models of developmental psychopa-thology have highlighted the importance ofenvironmental stressors in the etiology of different mentalhealth outcomes [5-8]. Some of these stressors includeacute traumatic events, chronic strain and adversity, andaccumulation of stressful life events and daily challenges[7]. Some of the most notable environmental stressorsthat have been found to profoundly impact children andyouth mental health are exposure to neighbourhood vio-lence [9]; parental chronic illness [10-13]; and povertyand economic hardship [12]. In addition, temporary orprolonged parental unemployment may add further stressin the form of increased parental alcohol intake, homeviolence and child abuse [13].Research on the impact of unemployment on children'sdevelopment has identified changes in parenting andmarital relationships as the primary factors linked to chil-dren's socio-emotional functioning [14-16]. Several stud-ies originating in the Great Depression of the 1930sindicated that fathers were more susceptible than mothersof emotional instability as a consequence to majorincome family losses [17,18]. Specifically, fathers tendedto have increased anger and be more punitive in disciplin-ing their children [15,19,20].Parental socioeconomic difficulties due to prolongedunemployment or loss of income are also associated withincreased addictive behaviours such as alcohol depend-ence and alcoholism, and family violence [13] which inturn increase the risk of children living in such familyenvironments to develop poor mental and physical healthin adulthood [21-23]. Several studies postulated that theway in which socioeconomic hardship affects parentingstyles, especially in fathers, functions as an importantdeterminant of children's mental health outcomes andschool related problems [24-28].As the effect of socioeconomic hardship and unemploy-ment can be mediated through changes in parentingstyles, adverse psychosocial work conditions experiencedby employed fathers can also play a critical role in chil-dren's outcomes.In addition to the specific type of employment, the degreeenvironment which is acknowledged as a critical determi-nant of children's outcomes.Two studies have been conducted to date on the impact ofpsychosocial work conditions of employed fathers onchildren's outcomes. Steward and Barling [25] conducteda study with 189 grade 4 and 5 Canadian students andfound that paternal psychosocial work conditions wereassociated with parenting behaviours and children'sbehaviour. Ostry and colleagues [29] conducted a studywith 19,833 Canadian children of sawmill workers andfound significant associations between suicidal behav-iours and several psychosocial work conditions of fatherswhen their children were young or in their early adoles-cence.These findings, in conjunction with epidemiologicaltrends indicating that the onset of many mental healthdisorders occurs in childhood and adolescence [30-35],emphasize the importance of understanding how expo-sure to paternal psychosocial work conditions in child-hood influence mental health outcomes throughout thelifespan.This study focuses on the investigation of the componentsof the job strain model [36], also known as the demand/control model, as determinants of mental health out-comes across four broad age groups spanning from child-hood to adulthood. The demand/control modelpostulates that job strain occurs when workers are over-loaded psychologically and at the same time they have nocontrol over their work environment. This combinationof high psychological demand and low control is hypoth-esized to increase the risk of stress related illnessesamongst the workers [37].Here we hypothesize that paternal psychosocial work con-ditions play an important role in the onset of mentalhealth problems and that, consistent with the literatureon developmental psychopathology, paternal mentalhealth and socio-demographic characteristics can be sig-nificant determinants of mental health condition fromchildhood to adulthood.Therefore, the present study investigates the impact ofpaternal psychosocial work conditions on a broad rangeof mental health conditions in a cohort of children whosefathers were employed in a selected group of sawmills inBritish Columbia, Canada. In recognition of the impor-tance of addressing mental health outcomes from a devel-opmental perspective, we investigate the onset of mentalhealth conditions in four different age groups: childhoodand early adolescence, adolescence, young adulthood,Page 2 of 10(page number not for citation purposes)to which fathers experience stress at the workplace canalso contribute significantly to the quality of the familyand adulthood. Because the influence of stressors on men-tal health can change considerably from childhood toBMC Public Health 2008, 8:104 http://www.biomedcentral.com/1471-2458/8/104adolescence and throughout adulthood [38-40] conduct-ing separate analysis for these four groups has allowed usto investigate the specific effects of paternal psychosocialwork conditions on mental health outcomes from a life-course perspective.MethodsThis study is based on a cohort of male sawmill workers(N = 28,794) for whom data on employment history, jobmobility, and physical and psychosocial work conditionswere obtained. The original cohort of sawmill workerswas gathered in the mid 1990s in order to conduct anoccupational study on the effects of chlorophenol anti-sapstain exposure among British Columbian sawmillworkers. Fourteen sawmills located in British Columbiawere identified and personnel records for workers whohad worked in one of these mills for at least one year inthe period between1950 and 1998 were accessed andreviewed. Personal identifying information for eligibleworkers and complete job history records were abstractedfrom personnel records (see Hertzman et al. [41] for acomplete description of methods used to assemble thesawmill workers' cohort).Using birth files from the British Columbia provincialvital statistics registry and BC Linked Health Data Base weidentified the children born between 1952 and 2000 tothe cohort of sawmill workers. Through these two link-ages 37,827 children of sawmill workers were identified,forming an offspring cohort. This study focuses on thedevelopmental psychopathology among the offspringcohort and its association with paternal psychosocialwork conditions.Study participantsThe cohort of adult sawmill workers (i.e., fathers) waslinked to the British Columbia birth file in order to iden-tify all of the children of these workers born in BritishColumbia between 1952 and 2000. There were 37, 827children in the cohort. Ages of the children in the cohortranged from less than one year old to 49 years old in2000. In order to meet the eligibility criteria for this study,the fathers must have worked at least one year in one ofthe study sawmills while their children were between theage of 0 and 16 years. This criterion was set to ensure thatthe psychosocial work conditions had been measured forthe period during childhood and early adolescence. Atotal of 19,833 children of sawmill workers satisfied theeligibility criterion for inclusion.In the International Classification of Diseases, 9th Revi-sion (ICD9), children and adults are defined based on theage cut-off of 14. There are different diagnostic codes forand some may say erroneous, to consider all individuals15 years and older as adults, given the well documenteddevelopmental differences between adolescence andadulthood. Therefore, we used general knowledge ofbroad differentiations between developmental stages toclassify 15 years old and older individuals in three agegroups: adolescents, young adults and adults. Thus, thefour groups identified were: children who had been diag-nosed with mental health problems for the first time (i.e.,onset) by a health professional at 14 years of age oryounger (children and early adolescent group); adoles-cents whose onset of a mental health problem had beendiagnosed between 15 and 19 years of age (adolescentgroup); young adults whose onset of a mental healthproblem had been diagnosed between 20 and 30 years ofage (young adult group); and adults whose onset of amental health problem had been diagnosed between 31and 49 years of age (adult group). Table 1 indicates thenumber of cases identified for each of the mental healthconditions analyzed in the present study among the fourage groups.Paternal psychosocial work conditionsExposure to job strain (i.e., work stress) can be measuredin different ways. Typically work stress is assessed fromself-reports via a questionnaire, inferred from occupa-tional titles, or 'externally' assessed by expert job evalua-tors that evaluate the degree to which certain jobs arestressful on the basis of the job characteristics [42].This study used both occupational titles and job evalua-tors to assess work stress. Specifically we obtained histori-cal estimates of job control, psychological demand,physical demand, social support, and noise among thesawmill workers (i.e., the fathers) in the following way: 4experienced job evaluators (two union and two manage-ment) in the British Columbia sawmill industry filled outthe demand/control questionnaire to obtain a retrospec-tive estimation for all basic job titles prior to 1975 (seeOstry, Marion, Green et al. [43], for these methods); 2) apanel of senior workers was selected in each participatingmill and completed the demand/control questionnairefor basic job titles in their mill for two time periods (1975to 1985), (1985 to 1998) (see Ostry, Marion, Demers etal. [44] for these methods). In the present study theseexpert estimates of fathers' psychosocial work conditionsare treated as exposures to their offspring. Because esti-mates were provided for each sub-dimension of thedemand/control model, we were able to investigate theassociation between specific aspects of fathers' work con-dition (i.e., control, psychological demand, physicaldemand, social support, and noise) and mental healthoutcomes in their offspring.Page 3 of 10(page number not for citation purposes)the paediatric (14 years and younger) and adult (15 andolder) populations. However, it is arguably reductionist,BMC Public Health 2008, 8:104 http://www.biomedcentral.com/1471-2458/8/104Employment history, mental health, and socio-demographicsWhile the study focuses on the effect of paternal psycho-social working conditions on the development of psycho-pathology, there are some potential confoundingvariables that need to be accounted for in the analysis.Such variables are paternal socio-demographic character-istics, paternal mental health and paternal employmenthistory.From the job history records of the sawmill workerscohort we obtained the number of episodes of unemploy-ment, job mobility (classified as upward, downward orMental health and alcohol dependence of the fathers wereobtained from a provincial administrative database.Using probabilistic linkage techniques the sawmill work-ers' cohort was linked to the British Columbia LinkedHealth Database (BCLHDB), consisting of person-spe-cific, longitudinal records on all British Columbians. TheBCLHDB contains files with data on deaths, hospital dis-charges, and all physician encounters for the years 1985through to 2001. The records are stored separately buthave been indexed with an individual service-recipientspecific code so that individual's records can be linkedacross files for specific research projects.Table 1: Number of cases per mental health diagnosisChildren (N = 19,833) Cases Controls TotalICD9 300 Neurotic Disorders 119 357 476ICD9 301 Personality Disorder 32 96 128*ICD9 308 Acute Reaction to Stress 48 144 192ICD9 309 Adjustment Reaction 118 354 472ICD9 311 Depression 187 561 748Adolescents Cases Controls TotalICD9 300 Neurotic Disorders 344 1032 1376ICD9 301 Personality Disorder 81 243 324ICD9 308 Acute Reaction to Stress 181 543 724ICD9 309 Adjustment Reaction 187 561 748ICD9 311 Depression 643 1929 2572ICD9 303 Alcohol Dependence 36 108 144ICD9 304 Drug Dependence 57 171 228ICD9 305 Non-Dependent Drug Abuse 74 222 296Young Adults Cases Controls TotalICD9 300 Neurotic Disorders 1058 3174 4232ICD9 301 Personality Disorder 163 489 652ICD9 308 Acute Reaction to Stress 651 1953 2604ICD9 309 Adjustment Reaction 404 1212 1616ICD9 311 Depression 1682 5046 6728ICD9 303 Alcohol Dependence 147 441 588ICD9 304 Drug Dependence 243 726 969ICD9 305 Non-Dependent Drug Abuse 195 585 780Adults Cases Controls TotalICD9 300 Neurotic Disorders 725 2175 2900ICD9 301 Personality Disorder 151 453 604ICD9 308 Acute Reaction to Stress 468 1404 1872ICD9 309 Adjustment Reaction 324 972 1296ICD9 311 Depression 1118 3354 4472ICD9 303 Alcohol Dependence 125 375 500ICD9 304 Drug Dependence 173 519 692ICD9 305 Non-Dependent Drug Abuse 117 351 468* These diagnostic codes were eliminated from the analysis because the ratio of participants to independent variables was not sufficient.Page 4 of 10(page number not for citation purposes)stable), occupation (manager, tradesman, skilled worker,and unskilled worker) ethnicity, and marital status.Ethical approval was obtained from the University of Brit-ish Columbia (UBC) and the British Columbia MinistryBMC Public Health 2008, 8:104 http://www.biomedcentral.com/1471-2458/8/104of Health to conduct this study. A Data Access Subcom-mittee consisting of health ministry personnel, staff fromthe British Columbia Ministry of Information and Privacy,and the UBC Centre for Health Services and PolicyResearch has been established to handle requests for link-age to the BCLHDB and to ensure that such requests meetscientific and ethical standards, are in the public interest,and conform with the Freedom on Information and Pro-tection of Privacy Act.Number of episodes of unemployment, job mobility,occupation, ethnicity, marital status, alcohol dependenceand mental health of the father were used in the analysisto control for the potential confounding effect of thesevariables on the association between psychosocial workconditions and developmental psychopathology.Mental Health Outcomes of the Children's CohortIn British Columbia children who experience mentalhealth problems can be evaluated by mental health pro-fessional at hospitals or medical clinics. The reason formedical visit or hospitalization (which can include a diag-nosis if one is provided) is recorded on administrativeforms that are sent and stored at the British ColumbiaMinistry of Health. In addition, complete hospital dis-charge and physician visit records for the paediatric andadult population are available through the BCLHDB thatalso provides the ICD9 codes for mental health condi-tions. The BCLHDB was accessed to identify all cases witha diagnosis of mental health occurring between Jan 1st,1991 and March, 31st, 2001. We designated 1991 as thestart year because it was the first year that ICD diagnoseswere obtained on physician billing records in BritishColumbia. The same probabilistic linkage techniques out-lined above were used here for the identification of men-tal health cases.A mental health case was defined as an individual whohad been assigned for the first time a mental health diag-nosis after the start of the father's employment at one ofthe study sawmills. Furthermore, only mental health con-ditions that allowed comparisons to be made between atleast three of the four age groups are reported in this study.A mental health case for children 14 years old and youngerwas defined as a child's first diagnosis labelled with ICD9codes 300 (neurotic disorder), 301 (personality disor-ders), 308 (acute reaction to stress), 309 (adjustment reac-tion), and 311 (depressive disorder).A mental health case for individuals 15 years old and olderwas defined as an individual's first diagnosis labelled withICD9 codes 300 (neurotic disorder), 301 (personality dis-abuse), 308 (acute reaction to stress), 309 (adjustmentreaction), and 311 (depressive disorder).Each participant was assigned a code of 1 (i.e., had beendiagnosed) or 0 (i.e., had not been diagnosed) for each ofthe above ICD9 codes. For discussion purposes, thesemental health diagnoses were grouped into two catego-ries: non-psychotic disorders and drug and alcohol relateddisorders. Non-psychotic disorders include those condi-tions that do not have a psychotic symptomatology andare primarily characterized by anxiety and depressivesymptoms. These include neurotic disorder, personalitydisorder, acute reaction to stress, adjustment reaction anddepression. Drug and alcohol related disorders includethose conditions associate with drug and alcohol use andabuse such as alcohol dependence, drug dependence, andnon-dependent drug abuse.Note that because we identified age at first diagnosis asour criterion for inclusion, the age groups are 'exclusive' asfar as specific diagnoses are concerned. For example, a par-ticipant can only be assigned a diagnosis of depression forthe first time once. Thus each participant will only be rep-resented in the age group reflecting his or her age at timeof diagnosis for that specific diagnosis. On the other hand,the same participants can be diagnosed with differentmental health conditions at different times in their life orwithin the same age period. Because of the high co-mor-bidity in mental health, it is possible that the same partic-ipants appear twice or more in the analysis within anyspecific age group or across different age groups. Figure 1shows the age distribution of first mental health diag-noses among the four groups of participants.AnalysisUsing survival-time to case-control on STATA 8.0, threecontrols were selected for each mental health casematched on age and gender. Controls were chosen ran-domly with replacement from the set at risk. The set at riskwere all the offspring of the sawmill worker's cohort, bornbetween 1952 and 2001, whose father had worked in astudy sawmill for at least one year during the first 16 yearsof the child's life. These could be anyone at risk who alsosatisfied the matching criteria who had not been diag-nosed with a mental health condition at the time of diag-nosis of the case.Statistical analyses were conducted using conditionallogistic regression on STATA 8.0. A total of 28 multivariatemodels were run (4 models for the children's cohort; 8models for the adolescent cohort; 8 models for the youngadult cohort; and 8 models for the adult cohort). Theindependent variables were the same for each of the 28Page 5 of 10(page number not for citation purposes)order), 302 (sexual deviations), 303 (alcohol depend-ence), 304 (drug dependence), 305 (non-dependent drugmodels tested and included: control (ordinal – lowestvalue as the referent); psychological demand (ordinal –BMC Public Health 2008, 8:104 http://www.biomedcentral.com/1471-2458/8/104lowest value as the referent); physical demand (ordinal –lowest value as the referent); social support (ordinal –lowest value as the referent); and noise (ordinal – lowestvalue as the referent).Control variables included in the models were: durationof employment (continuous variable); paternal ethnicity(one dummy variable for Chinese and one dummy varia-ble for Sikh – Caucasian was as referent); marital status(one dummy variable – married as the referent); paternalalcohol dependence (one dummy variable – no diagnosisas the referent); mental health of the father (one dummyvariable – no diagnosis as the referent); suicidal behaviourof the father (one dummy variable – no diagnosis as thereferent); type of employment (one dummy variable fortrades, one dummy variable for skilled, and one dummyvariable for unskilled – management was as referent); Allindependent variables and control variables were enteredat once and no stepwise procedure was used.ResultsFigure 1 indicates that 67% of cases were identified by theage of 30, with young adulthood representing the devel-opmental period with the greatest number of first timemental health diagnoses. The most prevalent diagnoses33%; adults – 31%), and neurotic disorders (children/early adolescents – 21%; adolescents – 20%; young adults– 20%; adults – 21%). Adjustment reaction was the thirdmost prevalent diagnosis among children/early adoles-cents (20%) and adolescents (11%); and acute reaction tostress was the third most prevalent diagnosis amongyoung adults (13%) and adults (13%).Table 2 and Additional file 1, 2 and 3 show the results ofthe multivariate analyses for children/early adolescents,adolescents, young adults, and adults respectively.Ethnicity of the father was often associated with mentalhealth conditions from childhood through adulthood.Children of fathers of Sikh or Chinese origin were consist-ently at lower risk of being diagnosed with mental healthconditions such as depression, neurotic disorders, adjust-ment reaction, acute reaction to stress, and non-depend-ent drug abuse. Ethnic origin was associated with differentage groups depending on the specific mental health con-dition. For example, being of Sikh origin functioned as aprotective factor against depression and neurotic disor-ders among children (neurotic disorder: OR = .193, p =.001, 95% IC = .075–.495; depression: OR = .184, p <.001, 95%CI = .090–.379) and adolescents (neurotic dis-order: OR = .288, p < .001, 95% IC = .185–.447; depres-sion: OR = .439, p < .001, 95%CI = .331–.581), but notamong young adults (neurotic disorder: OR = .830, p >.05, 95% IC = .663–1.04; depression: OR = .870, p > .05,95%CI = .731–1.04) and adults (neurotic disorder: OR =.1.05, p > .05, 95% IC = .657–1.68; depression: OR = .953,p > .05, 95%CI = .652–1.39); and being of Chinese originfunctioned as a protective factor for depression (OR =.211, p < .001, 95%CI = .114–.393) and neurotic disor-ders among young adults (OR = .357, p = .001, 95%CI =.193–.661), and for adjustment reaction among adoles-cents (OR = .187, p < .05, 95%CI = .038–.930) and youngadults (OR = .186, p < .05, 95%CI = .043–.807).Worth of noting is that while the proportion of Chineseparticipants in our sample (approximately 1% or less ofthe general population) is representative of the BritishColumbia population the Sikh participants are overrepre-sented in our sample with approximately 13% of the par-ticipants versus the expected 3% in the general BCpopulation [45].Duration of employment was associated with depressionamong adolescents (OR = .964, p < .05, 95%CI = .937–.991), alcohol dependence among adults (OR = .931, p <.05, 95%CI = .883–.982), and non-dependent drug abuseamong adults (OR = .924, p < .05, 95%CI = .871–.979):the longer the fathers were employed at one sawmill theNumber of cases per age at diagnosisFigur  1Number of cases per age at diagnosis. Figure 1 reveals that 67% of cases were identified by the age of 30, with young adulthood containing the greatest number of mental health diagnoses. The most prevalent diagnoses assigned to all age groups are neurotic disorders (children/early adoles-cents – 21%; adolescents – 20%; young adults – 20%; adults – 21%) and depression (children/early adolescents – 33%; ado-lescents – 37%; young adults – 33%; adults – 31%). The third most prevalent diagnosis among children/early adolescents (20%) and adolescents (11%) was adjustment reaction whereas the third most prevalent diagnosis among young adults (13%) and adults (13%) was acute reaction to stress.0100020003000400050006000<15 15 to 19 20 to 30 >30Page 6 of 10(page number not for citation purposes)assigned to all age groups are depression (children/earlyadolescents – 33%; adolescents – 37%; young adults –less the risk of being diagnosed with each of these condi-tions.BMC Public Health 2008, 8:104 http://www.biomedcentral.com/1471-2458/8/104Page 7 of 10(page number not for citation purposes)Table 2: Results of the multivariate analysis among the children and early adolescents' cohortPredictor Odds Ratio SE z P > |z| 95% CINeurotic DisordersDuration of Employment 0.995 0.047 -0.11 0.912 0.907–1.09Control 0.948 0.067 -0.75 0.451 0.826–1.09Psychological Demand 1.05 0.109 0.46 0.644 0.856–1.29Physical Demand 1.62 0.606 1.29 0.196 0.779–3.37Social Support 1.06 0.278 0.20 0.839 0.629–1.77Noise 0.899 0.363 -0.26 0.794 0.408–1.99Trades Worker 4.31 4.80 1.31 0.189 0.486–38.2Skilled Worker 3.79 4.19 1.21 0.228 0.435–33.1Unskilled Worker 4.03 4.43 1.27 0.204 0.468–34.8Marital Status 1.05 0.081 0.71 0.479 0.908–1.23Chinese 0.459 0.517 -0.69 0.489 0.051–4.17Sikh 0.193 0.093 -3.42 0.001* 0.075-.495Paternal Alcoholism 0.308 0.365 -0.99 0.320 0.030–3.13Paternal Mental Health 3.08 0.920 3.77 0.000* 1.72–5.53Paternal Suicidal Behaviours 0.590 0.597 -0.52 0.602 0.081–4.28Acute Reaction to StressDuration of Employment 1.04 0.079 0.46 0.646 0.892–1.20Control 0.955 0.122 -0.36 0.721 0.744–1.23Psychological Demand 0.963 0.165 -0.22 0.827 0.688–1.35Physical Demand 0.698 0.420 -0.60 0.550 0.214–2.27Social Support 0.620 0.269 -1.10 0.271 0.265–1.45Noise 1.26 0.865 0.34 0.735 0.329–4.84Trades Worker 0.747 0.609 -0.36 0.721 0.151–3.69Skilled Worker 0.436 0.368 -0.98 0.325 0.083–2.28Unskilled Worker 1.20 0.964 0.23 0.818 0.250–5.79Marital Status 1.06 0.109 0.59 0.555 0.869–1.30Chinese 3.93e-16 2.33e-8 -0.00 1.00 0Sikh 0.188 0.154 -2.04 0.041* 0.038-.935Paternal Alcoholism 0.504 0.333 -1.04 0.299 0.138–1.84Paternal Mental Health 1.60 0.679 1.10 0.271 0.694–3.68Paternal Suicidal Behaviours 1.53e-15 6.32e-8 -0.00 1.00 0Adjustment ReactionDuration of Employment 0.960 0.048 -0.80 0.421 0.869–1.06Control 0.955 0.066 -0.65 0.515 0.834–1.09Psychological Demand 0.838 0.100 -1.47 0.140 0.663–1.05Physical Demand 1.28 0.520 0.62 0.536 0.580–2.84Social Support 0.820 0.215 -0.75 0.452 0.489–1.37Noise 0.706 0.262 -0.94 0.349 0.340–1.46Trades Worker 0.755 0.614 -0.35 0.730 0.153–3.71Skilled Worker 0.821 0.694 -0.23 0.816 0.156–4.30Unskilled Worker 0.909 0.743 -0.12 0.908 0.183–4.51Marital Status 1.00 0.072 0.08 0.934 0.873–1.15Chinese 0.514 0.570 -0.60 0.549 0.058–4.52Sikh 0.154 0.077 -3.71 0.000* 0.057-.414Paternal Alcoholism 0.873 0.700 -0.17 0.866 0.181–4.20Paternal Mental Health 2.04 0.568 2.57 0.010* 1.18–3.52Paternal Suicidal Behaviours 0.874 1.19 -0.10 0.922 0.060–12.6DepressionDuration of Employment 0.991 0.037 -0.23 0.821 0.921–1.07Control 0.934 0.050 -1.25 0.211 0.841–1.03Psychological Demand 0.949 0.076 -0.64 0.522 0.811–1.11Physical Demand 0.750 0.204 -1.05 0.293 0.440–1.28Social Support 0.787 0.153 -1.22 0.222 0.537–1.15Noise 0.685 0.200 -1.29 0.197 0.386–1.21Trades Worker 0.672 0.359 -0.74 0.458 0.236–1.91Skilled Worker 0.482 0.264 -1.33 0.184 0.164–1.41Unskilled Worker 0.480 0.262 -1.34 0.179 0.165–1.39Marital Status 1.00 0.077 0.07 0.943 0.865–1.16Chinese 0.565 0.455 -0.71 0.479 0.116–2.74Sikh 0.184 0.067 -4.60 0.000* 0.090-.379Paternal Alcoholism 1.08 0.746 0.12 0.904 0.282–4.17Paternal Mental Health 2.17 0.484 3.49 0.000* 1.40–3.36Paternal Suicidal Behaviours 1.12 1.14 0.12 0.907 0.154–8.21*p < .05BMC Public Health 2008, 8:104 http://www.biomedcentral.com/1471-2458/8/104Paternal psychosocial work conditions were also signifi-cant factors associated with all mental health conditionsinvestigated. However, different sub-dimensions of psy-chosocial work conditions were associated with specificdiagnoses across the different age groups (see Table 2 andAdditional files 1, 2, 3): control was associated withadjustment reaction among adolescents (OR = .866, p <.05, 95%CI = .775–.967) and alcohol dependence amongadolescents (OR = .770, p < .05, 95%CI = .536–.994); psy-chological demand was associated with neurotic disordersamong young adult (OR = .702, p < .001, 95%CI = .568–.868); noise was associated with acute reaction to stressamong adults (OR = 1.42, p < .05, 95%CI = 1.06–1.91);and social support was associated with non-dependentdrug abuse among adults (OR = 1.95, p < .05, 95%CI =1.12–3.42).DiscussionIn this study we investigated whether paternal psychoso-cial work conditions played a role in developmental psy-chopathology among the children of a cohort of sawmillworkers from British Columbia, Canada. The presentstudy has generated some important findings that contrib-ute to the literature on mental health and developmentalpsychopathology in different ways.First, we found that ethnicity of the father was a significantpredictor where the risk of being diagnosed with one ofseveral mental health outcomes from childhood to adult-hood was lower for children of Chinese of Sikh fathers.This finding is consistent with what we found amongadult sawmill workers where mental health diagnoseswere less likely to be observed among Sikh and Chineseworkers [46]. This finding is also consistent with findingsfrom studies on the role of ethnicity on child develop-ment indicating that positive ethnic identity amongminority children and adolescents functions as a protec-tive factor for mental health and school outcomes [2-4].Second, we found that the longer the fathers wereemployed at one sawmill the less the risk of being diag-nosed with depression, alcohol dependence, and non-dependent drug abuse. This finding is consistent with theliterature on unemployment and health in that is suggeststhat children may be affected by paternal unemploymentand loss of income in important ways [24-28].Third, we found that the association between paternalpsychosocial work conditions (that in most cases weremeasured years prior to the mental health diagnosis) andmental health outcomes could be different depending onthe age cohort (e.g., whether it was the children's or theadult's cohort). For example, with the exception of adjust-hood than it did during childhood. In some cases, pater-nal psychosocial work conditions were associated withmental health diagnoses in adulthood and not in otherage groups such as in acute reaction to stress and non-dependent drug abuse. These findings suggest that whilepaternal psychosocial work conditions are significantlyassociated with a number of different mental health diag-noses, their effects are generally more noticeable over thelong term than in the immediate future.To a certain extent the results reported in this study areconsistent with findings of the developmental literature inthat they provide further evidence of the significantimpact of early influences on health and developmentfrom childhood throughout adulthood. However, thesefindings have some important elements of novelty, mostnotably the original evidence of the critical role that expo-sure to adverse paternal psychosocial work conditionsplay in the first sixteen years of life contributing to theonset of mental health outcomes in different age cohorts.While this study used rigorous design and analytical strat-egies, there are some limitations that need to be men-tioned. One limitation is related to the use of anadministrative database instead of complete clinicalrecords to infer diagnosis which can lead to misclassifica-tion bias. However, this potential bias may be minimizedbecause we used a large number of participants. Anotherimportant limitation of the study is the possible biaslinked to the expected co-morbidity in mental health andthe fact that the same participants may have appearedtwice or more in the analysis.At the conceptual level, a greater issue that constitutes apotentially important limitation is linked to the historicalchallenge of defining mental health and the ways inwhich mental health manifests in minority cultures. In thepresent study mental health was implicitly definedaccording to the parameters of psychiatry and clinical psy-chology traditions embodied in the DSM IV (and reflectedin the ICD9 codes) where 'mental disease' rather than'mental health' is identified. This is a very important issueespecially in light of fact that important associationsbetween ethnic origin and several mental health diag-noses were here identified. Vega and Rumbaut [47] elo-quently describe the important issues that influence theway minority mental health is investigated in NorthAmerican societies. Specifically, they argue that one of thegreatest challenges of contemporary researchers lies intheir ability to disentangle cultural influences from moreaccurate measurement and understanding of psychiatricproblems in minority cultures.Page 8 of 10(page number not for citation purposes)ment reaction, parental psychosocial work conditionsplayed a more important role in adolescence and adult-Despite the potential limitations of this study, the find-ings reported here are overall important in many respects.BMC Public Health 2008, 8:104 http://www.biomedcentral.com/1471-2458/8/104First, they contribute to the advancement of theories ofdevelopmental psychopathology by incorporating con-cepts from the occupational health literature (i.e., psycho-social work conditions) and generating discussion onways in which these can be 'brought home' and turnedinto 'adverse childhood experiences'. Second, results ofthis study also add to the literature of the effect of parentalunemployment on child development and specificallyidentify what sub-dimensions of psychosocial work con-ditions are the most probable to increase the risk fordevelopmental psychopathology from childhood toadulthood. Third, findings from this study may also haveimportant clinical implications by informing age specificintervention strategies and family therapy approacheswhere parents are offered support to cope with work stressand related occupational challenges.ConclusionBy investigating several mental health outcomes of 19,833children of sawmill workers from British Columbia, Can-ada we found that the degree to which children areexposed to adverse paternal psychosocial work conditionsduring childhood is associated with their mental healthoutcomes throughout adulthood. his study provides sup-port to the tenet that being exposed to paternal work stressduring childhood can have long lasting effects on themental health of individuals.Competing interestsThe author(s) declare that they have no competing inter-ests.Authors' contributionsSM designed the analytical plan and conceived the con-ceptual framework of the manuscript; AO participated inthe development of the analysis and conceptual frame-work. JT and JD reviewed drafts of the manuscript. RH andLC carried out the analysis. CH established the sawmillcohort, obtained linkages with health administrative data,and took part in drafting the manuscript. All authors readand approved the final manuscript.Additional materialAcknowledgementsWe thank Jim van Os and Johannes Siegrist for their insightful comments during the review process of this manuscript. This work was funded by the Canadian Population Health Initiative. Drs. Maggi and Ostry held Scholar Awards from the Michael Foundation for Health Research in British Colum-bia. Dr. Hertzman holds a Canada Research Chair in Population Health.References1. Hughes D, Rodriguez J, Smith EP, Johnson DJ, Stevenson HC, SpicerP: Parents' ethnic-racial socialization practices: a review ofresearch and directions for future study.  Developmental Psychol-ogy 2006, 42:747-770.2. Chatman CM, Eccles JS, Malanchuk O: Identity negotiations ineveryday settings.  In Navigating the future: social identity, coping andlife tasks Edited by: Downey G, Eccles JS, Chatman CM. New York:Russell Sage Foundation; 2005:116-140. 3. Oyserman D, Harrison K, Bybee D: Can racial identity be promo-tive of academic efficacy?  International Journal of Behavioral Devel-opment 2001, 25:379-385.4. Shelton JN, Yip T, Eccles JS, Chatman CM, Fuligni A, Wong C: Ethnicidentity as a buffer of psychological adjustment to stress.  InNavigating the future: social identity, coping, and life tasks Edited by:Downey G, Eccles JS, Chatman CM. New York: Russell Sage Founda-tion; 2005:96-115. 5. Cicchetti D, Toth SL: A developmental perspective on internal-izing and externalizing disorders.  In Internalizing and ExternalizingExpression of Dysfunction Edited by: Cicchetti D, Toth SL. New York:Erlbaum; 1991:1-19. 6. Cicchetti D, Toth SL, Eds: Developmental perspectives on trauma: theory,research and intervention Rochester, NY: Rochester University Press;1997. 7. Haggerty RJ, Sherrod LR, Garmezy N, Rutter M, Eds: Stress, risk, andresilience in children and adolescents: processes, mechanisms, and interven-tions New York: Cambridge University Press; 1994. 8. Rutter M: Pathways from childhood to adult life.  Journal of ChildPsychology and Psychiatry 1989, 30(1):23-51.9. Attar B, Guerra N, Tolan P: Neighborhood disadvantage, stress-ful life events, and adjustment in urban elementary-schoolchildren.  Journal of Clinical Child Psychology 1994, 23:391-400.10. Kliewer W: Children's coping with chronic illness.  In Handbookof children's coping: linking theory and intervention. Issues in clinical childpsychology Edited by: Wolchik S, Sandler IN. New York: Plenum Press;1997:275-300. 11. Worsham N, Compas B, Ey S: Children's coping with parental ill-ness.  In Handbook of children's coping: linking theory and intervention.Issues in clinical child psychology Edited by: Wolchik S, Sandler IN. NewYork: Plenum Press; 1997:195-213. 12. McLoyd VC, Wilson L: The strain of living poor: parenting,social support, and child mental health.  In Children in poverty:Additional file 1Results of the multivariate analysis among the adolescent cohort. The data provided represent the multivariate analysis for the adolescent cohort.Click here for file[http://www.biomedcentral.com/content/supplementary/1471-2458-8-104-S1.doc]Additional file 2Results of the multivariate analysis among the young adult cohort. The data provided represent the multivariate analysis for the young adult cohort.Click here for file[http://www.biomedcentral.com/content/supplementary/1471-2458-8-104-S2.doc]Additional file 3Results of the multivariate analysis among the adult cohort. The data pro-vided represent the multivariate analysis for the adult cohort.Click here for file[http://www.biomedcentral.com/content/supplementary/1471-2458-8-104-S3.doc]Page 9 of 10(page number not for citation purposes)child development and public policy Edited by: Huston AC. New York:Cambridge University Press; 1991:105-135. Publish with BioMed Central   and  every scientist can read your work free of charge"BioMed Central will be the most significant development for disseminating the results of biomedical research in our lifetime."Sir Paul Nurse, Cancer Research UKYour research papers will be:available free of charge to the entire biomedical communitypeer reviewed and published immediately upon acceptancecited in PubMed and archived on PubMed Central BMC Public Health 2008, 8:104 http://www.biomedcentral.com/1471-2458/8/10413. Garmezy N: Stressors of childhood.  In Stress, coping, & develop-ment in children Edited by: Garmezy N, Rutter M. New York: McGraw-Hill Book Company; 1983. 14. Conger RD, Conger K, Elder G, Lorenz F, Simons R, Whitbeck L: Afamily process model of economic hardship and adjustmentof early adolescent boys.  Child Development 1992, 63:526-541.15. Elder G, Nguyen T, Caspi A: Linking family hardship to chil-dren's lives.  Child Development 1985, 56:361-375.16. Lempers J, Clark-Lempers D, Simons R: Economic hardship,parenting, and distress in adolescence.  Child Development 1989,60:25-39.17. Elder G: Children of the Great Depression Chicago, University of Chi-cago Press; 1974:1-444. 18. Eichorn DH, Clausen JA, Haan N, Honzik MMP: PHM: Present and pastin middle life New York, Academic Press; 1981:1-500. 19. Elder G, Liker J, Cross C: Parent-child behavior in the Great-Depression: life course and intergenerational influences.  InLife Span development and behavior Edited by: Baltes P, Brim Jr O. NewYork, Academic Press; 1984:109-158. 20. Liker JK, Elder GHJ: Economic hardship and marital relations inthe 1930s.  American Sociological Review 1983, 4:343-359.21. Edwards VJ, Holden GW, Felitti VJ, Anda RF: Relationshipbetween multiple forms of childhood maltreatment andadult mental health in community respondents: results fromthe adverse childhood experiences study.  American Journal ofPsychiatry 2003, 160:1453-1460.22. Anda RF, Croft JB, Felitti VJ, Nordenberg D, Giles WH, WilliamsonDF, Giovino GA: Adverse childhood experiences and smokingduring adolescence and adulthood.  JAMA 1999, 282:1652-1658.23. Felitti VJ, Anda RF, Nordenberg D, Williamson DF, Spitz AM, EdwardsV, Koss MP, Marks JS: Relationship of childhood abuse andhousehold dysfunction to many of the leading causes ofdeath in adults: the adverse childhood experiences (ACE)Study.  American Journal of Preventive Medicine 1998, 14:245-258.24. Flanagan CA, Eccles JS: Changes in parents' work status andadolescents' adjustment at school.  Child Development 1993,64:246-257.25. Stewart W, Barling J: Fathers' work experiences effect chil-dren's behaviors via job-related affect and parenting behav-iors.  Journal of Organizational Behavior 1996, 17:221-232.26. Barling J, Zacharatos A, Hepburn CG: Parents' job insecurityaffects children's academic performance through cognitivedifficulties.  Journal of Applied Psychology 1999, 84:437-444.27. Barling J, Dupre KE, Hepburn CG: Effects of parents' job insecu-rity on children's work beliefs and attitudes.  Journal of AppliedPsychology 1998, 83:112-118.28. MacEwan KE, Barling J: Effects of maternal employment experi-ences on children's behavior via mood, cognitive difficulties,and parenting behavior.  Journal of Marriage and Family 1991,53:635-644.29. Ostry A, Maggi S, Tansey J, Dunn J, Hershler R, Chen L, Louie AM,Hertzman C: The impact of fathers' physical and psychosocialwork conditions on attempted and completed suicideamong their children.  BMC Public Health 2006, 6:77-85.30. Kessler RC, Berglund P, Demler O, Jin R, Walters EE: Lifetimeprevalence and age-of-onset distribution of DSM-IV disor-ders in the National Comorbidity Survey Replication.  Archivesof General Psychiatry 2005, 62:593-602.31. Kilpatrick DG, Acierno R, Saunders B, Resnick HS, Best CL, SchnurrPP: Risk factors for adolescent substance abuse and depend-ence: data from a national sample.  Journal of Consulting and Clin-ical Psychology 2000, 68:19-30.32. Kilpatrick DG, Ruggiero KJ, Acierno R, Saunders BE, Resnick HS, BestCL: Violence and risk of PTSD, major depression, substanceabuse/dependence, and co-morbidity: results from theNational Survey of Adolescents.  Journal of Consulting and ClinicalPsychology 2003, 71(4):692-700.33. Biederman J, Hirshfeld-Becker DR, Rosenbaum JF, Hérot C, FriedmanD, Snidman N, Kagan J, Faraone SV: Further evidence of associa-tion between behavioural inhibition and social anxiety inchildren.  American Journal of Psychiatry 2001, 158:1673-1679.34. U. S. Department of Health and Human Services, Administration onChildren, Youth and Families: In focus: Understanding the effects of mal-treatment on early brain development Washington, DC: Government35. Rutter M, Yule B, Quinton D, Rowlands O, Yule W, Berger M:Attainment and adjustment in two geographical areas: IIIsome factors accounting for area differences.  British Journal ofPsychiatry 1975, 126:520-533.36. Karasek R, Theorell T: Healthy work: stress, productivity, and the recon-struction of working life New York (NY): Basic Books; 1990. 37. Karasek R, Theorell T: The demand-control-support model andCVD.  In The workplace and cardiovascular disease Edited by: SchnallPL, Belkic K, Landsbergis P, Baker D. Philadelphia (PA): Hanley & Bel-fus Inc; 2000:78-83. 38. Brooks-Gunn J, Auth JJ, Petersen AC, Compas BE: Physiologicalprocesses and the development of childhood and adolescentdepression.  In The depressed child and adolescent 2nd edition. Editedby: Goodyer IM. Cambridge, England: Cambridge University Press;2001:79-118. 39. Gunnar MR: Quality of early care and buffering of neuroendo-crine stress reactions: potential effects on the developingbrain.  Preventive Medicine: an International Journal Devoted to Practiceand Theory 1998, 27:208-211.40. Leffert N, Petersen AC: Biology, challenge, and coping in ado-lescence: effects on physical and mental health.  In Child devel-opment and behavioral pediatrics: crosscurrents in contemporarypsychology Edited by: Bornstein MH, Genevro JL. Hillsdale, NJ: Erl-baum; 1996:129-154. 41. Hertzman C, Teschke K, Ostry A, Hershler R, Dimich-Ward H, KellyS: Mortality and cancer incidence among sawmill workersexposed to chlorophenate wood preservatives.  American Jour-nal of Public Health 1997, 87:71-79.42. Belkic KL, Landsbergis PA, Schnall PL, Baker D: Is job strain a majorsource of cardiovascular disease risk?  Scandinavian Journal ofWork Environment and Health 2004, 30:85-128.43. Ostry A, Marion SA, Green L, Demers PA, Hershler R, Kelly S: Com-parison of expert-rater methods for assessing psychosocialjob strain.  Scandinavian Journal of Work Environment and Health 2001,27:1-6.44. Ostry A, Marion SA, Demers PA, Hershler R, Kelly S, Teschke K:Measuring psychosocial job strain with the job content ques-tionnaire using experienced job evaluators.  American Journal ofIndustrial Medicine 2001, 39:397-401.45. Statistics Canada. n.d: Population by selected ethnic origins, byprovince and territory (Table) British Columbia.  Versionupdated January 25, 2005.  accessed August 29, 200746. Ostry A, Maggi S, Tansey J, Dunn J, Hershler R, Chen L, Louie AM,Hertzman C: The impact of psychosocial work conditions onattempted and completed suicide among Western Canadiansawmill workers.  Scandinavian Journal of Public Health 35(3):265-71.47. Vega WA, Rumbaut RG: Ethnic minorities and mental health.Annual Review of Sociology 1991, 17:351-383.Pre-publication historyThe pre-publication history for this paper can be accessedhere:http://www.biomedcentral.com/1471-2458/8/104/prepubyours — you keep the copyrightSubmit your manuscript here:http://www.biomedcentral.com/info/publishing_adv.aspBioMedcentralPage 10 of 10(page number not for citation purposes)Printing Office; 2001. 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items