UBC Faculty Research and Publications

Nurse requirements in British Columbia : an analysis of the 1979-82 trends, part 1 : methods and preliminary… Kazanjian, Arminée 1947-; Chan, Susan Sep 30, 1984

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

Item Metadata


52383-Kazanjian_A_et_al_Nurse_requirements.pdf [ 40.46MB ]
JSON: 52383-1.0075921.json
JSON-LD: 52383-1.0075921-ld.json
RDF/XML (Pretty): 52383-1.0075921-rdf.xml
RDF/JSON: 52383-1.0075921-rdf.json
Turtle: 52383-1.0075921-turtle.txt
N-Triples: 52383-1.0075921-rdf-ntriples.txt
Original Record: 52383-1.0075921-source.json
Full Text

Full Text

Report S:l6 NURSE REQUIREMENTS IN BRITISH COLUMBIA: AN ANALYSIS OF THE 1979-82 TRENDS Part I: Methods and Preliminary Findings - Institutional Analysis Division of Health Services Research and Development Office of the Coordinator of Health Sciences The J.F. McCreary Health Sciences Centre The University of British Columbia Vancouver, B.C., Canada V6T 1Z6 September, 1984 REPORTS OF THE DIVISION OF HEALTH SERVICES RESEARCH AND DEVELOPMENT The following reports have been issued and are available to the public from the Division. SERIES HHRU (statistical tables only) HMRU 75:8 HHRU 75:9 HNRU 78:1 HMRU 78:2 HMRU 78:3 HMRU 83:1 NEW HSRD SERIES General Manpower Stock Simulator (GHSS) - Demonstration Example. (E.R. Sh:til,ington) The Number of Active Physicians in British Columbia, 1974-1984 - A Situation Report Analysis of Changes in Non-Post-Graduate Physicians from September 1975 to September 1977. (G. Muir, K. Kerluke) Selected Tables of Directory Active Physicians by RHD by Co111111Unity, as at September 1978. (G. Muir) Manpower Implications Derived from Payments by the Medical Services Colllllission to Physicians in British Columbia, 1976-1978. (G. Muir, K. Kerluke) Fee Practice Medical Service Expenditures Per Capita, and Full-Time-Equivalent Physicians in B.C. 1981-1982. (M.L. Barer, P. Wong Fung), July, 1983 PRODUCTION REPORTS: P;2 Training of Dietary Aides and 2'1tchnologists in Canada, Winter, 1975. (D.O. Anderson) P:3 PRODUCTION 77. A Status Report on the Production of Health Personnel in the Province of British Columbia. March, 1978 P:4 PRODUCTION 79. A Status Report on the Production of Health and Human Services Personnel in the Province of British Columbia. March, 1980 P:S PRODUCTION Bl. A Status Report on the Production of Health and Human Services Personnel in the Province of British Columbia, March, 1982 P:6 PRODUCTION 83. A Status Report on the Production of Health and Human Services Personnel in the Province of British Columbia, March, 1984. ROLLCALL REPORTS: R:7 Medical Laboratory 'l'echnologists in British Columbia, 1978 - A Survey of the Membership of the Societies. January, 1979. (A.J. Stark) R:S Supplement to ROLLCALL 77 for Selected Health Groups - A Status Report of Selected Health Personnel in the Province of British Columbia. March, 1979 R:9 Place of Graduation for Selected Health Occupations - 1977. July, 1979. (B. Mccashin, W.G. Manning) R:lO Dental Technicians in British Columbia, 1979 -A Survey of Licentiates of the Dental Technicians' Board of British Columbia, 1979. November, 1979. (C.W. Kinnis) R:ll Audiologists and Speech/Language Pathologists in British Columbia - 1979. The Results of a Survey of Members of the B.C. Speech and Hearing Association. February, 1980 --------------------------------------------------------Series HSRD R:l (1973) - R:6 (1978) and early reports in other series not listed. (Some are available on request). Report S:l6 NURSE REQUIREMENTS IN BRITISH COLUMBIA: AN ANALYSIS OF THE 1979-82 TRENDS Part I: Methods and Preliminary Findings - Institutional Analysis Division of Health Services Research and Development Office of the Co-ordinator of Health Sciences The J.F. McCreary Health Sciences Centre The University of British Columbia Vancouver, B.C., Canada, V6T 1Z6 A. Kazanjian, Dr.Soc. S. Chan, B.Sc. September, 1984 This report is one of a series describing the distribution of health manpower and health care resources. These reports, prepared for the Health Manpower Working Group of the Ministry of Health, are working documents and comments or suggestions are welcome. HEAL TH MANPOWER RESEARCH UNIT r:o ornn Of THI C'OORDISATOR Hl AL TH SC'lf.NC'ES C'ENTRF. PHONE: t6041 228~810 Mr. Chris Lovelace, Chairman, Health Manpower Working Group, Ministry of Health, 1515 Blanshard Street, Victoria, B .C., vaw JCS Dear Mr. Lovelace: 41h FLOOR l.R.r. Bl'ILDING THE l 1Nl\IERSIT\' or BRITISH C'OLnlBIA VANCOlVER. B.C .• C'ANADA V6T 1\\'5 September 26, 1984 It is with pleasure that I transmit to you and to the members of the Health Manpower Working Group the completed report "Nurse Requirements in British Columbia: An Analysis of the 1979-1982 Trends, Part I - Methods and Preliminary Findings". This report is the first in a series on RN requirements. It provides, we believe, the first in-depth analysis of deployment patterns in the Institutional Section for this province. Future work will include analyses of the Conununity Health and Education Sectors and on Net Requirements for RNs in B.C. While this is only one of a series, and as such does not present the complete picture, the information contained herein is, nevertheless, accurate and comprehensive. We trust the members of the Health Manpower Working Group and others will find it a useful planning tool. /slm. Sincerely yourt;"7 1t-Xv2~1 ,.., - -Arminee Kazanjian, Dr.Soc., Research Associate, Division of Health Services Research and Development. A Research Unil for the Health Manpower Working Group. Ministry of Health. British Columbia ACKNOWLEDGEMENTS Special appreciation is extended to Mrs. P. Wadsworth, Executive Director of the British Columbia Health Association, for granting permission to access the data and to the Data Services staff who took time from their busy schedules to provide us the numerous computer tapes. We are greatly indebted, of course, to the hospital Administrators and Presidents who agreed to release their payroll data through the B.C.H.A. We thank also the staff of the other hospitals who provided us data on an individual basis. The staff of the Research Division and Hospital Financial Services of Hospital Programs, Ministry of Health provided data on various occasions and we thank them for their assistance. Several members of the Research Unit have provided input to the study during its various stages. Important input was provided by Dr. A. Stark and Dr. M. Barer, both during the course of the study and in the writing of the report. Cheryl Jackson, Sharon Jansen and Gordon Wong have provided technical assistance throughout the course of the study. Dr. Kent Brothers provided statistical assistance. Mary Brunold and Susan Moloney provided secretarial support and mastered the art of word processing. TABLE OF CONTENTS List of Figures List of Tables I. Introduction 1. Background 2. Scope and Objectives II. Methodological Approach III. 3. Conceptual Framework 4. Discussion of Nurse Requirements Models 5. Research Design and Data Components Analysis of Findings 6. Descriptive Analysis of Data i ii 1 2 4 8 15 21 7. Requirements for RNs in Institutional Care_ 46 8. Requirements for PNs in Institutional Care 64 IV. Discussion 9. Sunmary of Findings 10. Concluding Remarks Appendices Appendix A Appendix B Appendix C 67 70 Figure 1 2 3 4 5 6 7 8 - 1 -LIST OF FIGURES Conceptual Framework for Nurse Requirements Estimation ••••• Annual Percent Changes for Nurse Paid Hours in B.C. Regular, Part-Time, Casual, 1979-1982 ••••••••••••••••••••••••••••••• Scattergram: Per Capita Registered Nurses Regular Paid Hours by Per Capita Practical Nurses Regular Paid Hours for Regional Hospital Districts, 1979 •••••••••••••••••••••••••••••••••••• Scattergram: Per Capita Registered Nurses Regular Paid Hours by Per Capita Practical Nurses Regular Paid Hours for Regional Hospital Districts, 1980 ••••••••••••••••••••••••••••••••••••• Scattergram: Per Capita Registered Nurses Regular Paid Hours by Per Capita Practical Nurses Regular Paid Hours for Regional Hospital Districts, 1981 ••••••••••••••••••••••••••••••••••••• Scattergram: Per Capita Registered Nurses Regular Paid Hours by Per Capita Practical Nurses Regular Paid Hours for Regional Hospital Districts, 1982 ••••••••••••••••••••••••••••••••••••• Scattergram: General Practitioners per 10,000 Population by Specialists per 10,000 Population for Regional Hospital Districts, 1979 .. ............................................. . Scattergram: General Practitioners per 10.000 Population by Specialists per 10,000 Populationfor Regional Hospital Districts, 1981 ................................................. . 9 39 47 48 49 50 51 52 Table 1 2 3 4 5 6 7 8 9 10 11 12 13 - 1 i -LIST OF TABLES Proportion of Salaries to Total Operating Expenses for B.C. Hospitals, by Regional Hospital District, 1978-1981 ••••••••• Representative Annual Starting Salaries of Selected Personnel in B.C., 1976-1982 •••••••••••••••••••••••••••••••••••••••••• Representative Annual Salaries of Selected Personnel as a Percentage of Registered Nurses' Salaries, 1976-1982 •••••••• Labour Force Activity Rates for RNs, B.C., 1975-1983 •••••••• Labour Force Activity Rates for LPNs, B.C., 1979-1983 •••••••• Average Relative Salaries of Registered Nurses to Practical Nurses in B.C., by Regional Hospital District, 1979-1982 ••••• Nurse Paid Hours, Regular, Part-Time, Casual, (percent totals) by Regional Hospital District, B.C., 1979. (Registered Nurses) ••••••••••••••••••••••••••••••••••••••••••••••••••••••• Nurse Paid Hours, Regular, Part-Time, Casual, (percent totals) by Regional Hospital District, B.C., 1980. (Registered Nurses) •••••••••••••••••••••••••••••••••.••••••••••••••••••••• Nurse Paid Hours, Regular, Part-Time, Casual, (percent totals) by Regional Hospital District, B.C., 1981. (Registered Nurses) •••....••..•..•..•....•..••..••..•••.•.••••••.••••.•••• Nurse Paid Hours, Regular, Part-Time, Casual, (percent totals) by Regional Hospital District, B.C., 1982. (Registered Nurses) ••••••••..••...•.•..••••.•...••....•.••••••••....•.•.•• Nurse Paid Hours, Regular, Part-Time, Casual, (percent totals) by Regional Hospital District, B.C., 1979. (Practical Nurses) • •••...••.....••••...•••..••...•••••••.•..••••......... Nurse Paid Hours, Regular, Part-Time, Casual, (percent totals) by Regional Hospital District, B.C., 1980. (Practical Nurses) .•••.....•••......••••••••••.••..•.•.•••••••.........•. Nurse Paid Hours, Regular, Part-Time, Casual, (percent totals) by Regional Hospital District, B.C., 1981. (Practical Nurses) ••••••••••••••.••.••.•.•••••.•••.•.........•..•...••••. 14 Nurse Paid Hours, Regular, Part-Time, Casual, (percent totals) 22 23 24 25 25 28 30 31 32 33 34 35 36 - 111 -LIST OF TABLES (continued) Tab1 e Page 15 16 17 18 19 20 21 22 23 24 25 26 27 28 by Regional Hospital District, B.C., 1982. (Practical Nurses). • • • • • • • • • . • • • • • • • • • • • • • • • • • • • • • • • • • . • • • • • • • • . . . • • • • • • • .37 Annual Percent Changes for Nurse Paid Hours, in B.C., Regular, Part-Time, Casual, 1979-1982 •••••••••••••••••••••••••••••••••• Total Paid Hours (percent change), Ratio Registered Nurse/Practical Nurse for B.C., 1979-1982 ••••••••••••••••••••• Rated Hospital Bed Capacity, by Regional Hospital District, B.C., 1979-1982 ••••••••••••••••••••••••••••••••••••••••••••••• Number of Employed Practical Nurses in B.C. by Regional Hospital District, 1979-1982 •••••••••••••••••••••••••••••••••• Per Patient-Day Costs for B.C. Hospitals, by Regional Hospital District, 1978-1981 ••••••••••••••••••••••••••••••••••••••••••• Average Length of Stay by Regional Hospital District, B.C., 197g_1982 ••••••••••••••••••••••••••••••••••••••••••••••••••••• Matrix of Correlation Coefficients of Variables in Requirements Model, B.C., 1979 Data ................ · •....................... Matrix of Correlation Coefficients of Variables in Requirements Model, B.C., 1980 Data ................ -....................... . Matrix of Correlation Coefficients of Variables in Requirements Mode 1, B .C., 1981 Data •••••••••••••••• · ••••••••••••••••.••••••• Matrix of Correlation Coefficients of Variables in Requirements Model, B.C., 1982 Data •.•.•...•••.••• : ..•......•..•••.••••.... Selected Statistics from the Multiple Regression on Registered Nurse Requirements, B.C., 1979-1982 ••••••••••••••••••••••••••• More Selected Statistics from the Multiple Regression on Registered Nurse Requirements, B.C., 1979-1982 •••••••••••••••• Registered Nurses Paid Hours, Observed and Predicted, B.C., by Regional Hospital District, 1979-1982 •••••••••••••••••••••••••• Selected Statistics from the Multiple Regression on Practical Nurse Requirements, B.C., 1979-1982 •••••••••••••••••••••••••••• 38 38 40 43 44 45 57 57 58 58 60 62 63 65 - 1 -I. INTRODUCTION 1. Background Fiscal restraint and cost containment in the past few years have put manpower research in the forefront of planning, especially in the health sector where the disbursement of limited funds is based on a complex allocation process (1,2). Health economists argue, and the evidence suggests, that the health manpower market is markedly different from other labour markets (3,4,5) and, that the traditional factors that affect supply/demand in the general labour market do not apply as readily to health manpower. In addition, there is considerable evidence to indicate that the market for nurse manpower is appreciably different from that for other health occupations (6,7,8). The nurse supply situation in British Columbia and elsewhere, in the past five years has been variously presented as one of shortage, then of surplus, reflecting not only oscillatory experience, but also a variety and diversity of viewpoints on what nurse requirements for this province should be. In order to describe and analyze the particular situation in B.C. a comprehensive study of nurse manpower was undertaken that would include analyses of requirements as well as of supply. A careful assessment of the relationships between supply and requirements data, and judicious assumptions regarding future requirements would, thus, provide the necessary input for making informed manpower planning decisions. An initial analysis of the current supply of Registered Nurse (RN) manpower in the province (9), reported separately, was conducted to provide some of the data necessary for effective human resource planning in the long term. The purpose of this part of the study is, firstly, to examine factors that have historically affected requirements for nursing services and secondly, in so doing to develop a model appropriate for estimating future requirements. - 2 -2. Scope and Objectives The demand for health care personnel, and in particular nursing personnel, is a "derived demand" prompted by "demand" for the products of this sector -- health care services. Traditional analyses of input markets examine such factors as (1) price (wage) elasticities of supply of input, (ii) demand elasticities for the primary product and (iii) input substitutes. However, in the Canadian context, the interrelationships between the aforementioned factors are somewhat different than in the U.S. where price plays a cardinal role in determining primary product demand (10). Other factors which are relevant to the Canadian setting include financial allocation decisions and availability of complementary capital input such as health care facilities, hospital beds, etc. These and others will be discussed in detail in a subsequent chapter. The objective of this part of the study is twofold: (i) to examine the response of the hospital sector (as the major employer of nurses) to the differences in wages of various nursing personnel and (ii) to develop a simple requirements model that incorporates a number of institutional characteristics (for exogenous variables) which influence the "demand" for nursing services (dependent variable). Requirements is defined as the amount of services, or manpower, required to satisfy an explicit or implicit set of assumptions about how a particular health sector functions (11). Demand is generally used in this report as a synonym for requirements. We will not, at this stage, engage in any discussion of the appropriateness of this concept versus others such as "need". Our choice of terminology and conceptual framework was guided solely by practical considerations regarding a feasible research design and data collection. A major drawback in this field of research has been researchers' reluctance to progress from studies of supply only, to those which combine supply and demand in an analytic framework that does not set A priori assumptions of "shortages" or "surpluses". Usually, the deceptively simple-appearing task of enumerating supply proves to be a major undertaking which comes to an abrupt end after counting stock. Another drawback has been a general reluctance to conceptualize beyond the traditional "economic demand" model which, albeit appropriate in a number of health manpower situations, does not appear to be of high value in the case of nurse manpower in British Columbia due to a number of structural factors. The most salient among these are production factors, demographic factors, fiscal factors and organizational factors. The relative importance of each set of factors varies appreciably in the - 3 -different methodological approaches as well as for the particular health system and manpower category under study. In contrast to studies which do not progress beyond counting stock, a number of studies have examined both supply of and demand for nurses using widely accepted theoretical frameworks (6). Despite the generally-acknowledged contributions made by such studies, temporal and geographic limitations detract from their generalizability. It has been suggested that the ideal way to undertake supply/demand research is to study both sides simultaneously (12). This approach guarantees the equal attention in design necessary for each component, and data generated from the supply study can be adequately linked to that generated from the one on requirements. We have endeavoured to follow this principle in the design of the research framework; data collection, however, has been carried out sequentially, starting with supply. It has thus been possible to move freely between the two components and further develop/refine supply-side analyses while progressing steadily towards the preliminary analysis of requirements data. This report attempts to place nurse requirements issues in a broader context in order to improve our understanding of some of the underlying factors. More particularly, an attempt is made to address the foll.owing questions: (i) What is the level of utilization (effective demand) of nursing services? Does it vary appreciably over time? (ii) Is there any manpower substitution among nurse categories by hospitals? Does this vary over time or across geographic areas? (iii) What are the major factors that affect nurse requirements? (iv) What are the interrelationships among the major quantifiable factors that affect nurse requirements? In attempting to answer these questions the report is divided into three parts. Section II, chapters 3-5, provides a detailed presentation of the methodological approach used in the development of a conceptual framework, research design and data analysis. Some background information on nurses' wages, labour force activity rates and unemployment patterns are presented in Section III, Chapter 6. Chapters 7 and 8 present a preliminary analysis of findings for each of the two groups under study -- the Registered (and graduate) Nurses (RNs) and the Licensed Practical Nurses (LPNs) and practical nurses (PNs). Section IV delves into a discussion of findings and the short- and long-term -implications of these. - 4 -II. t-'ETHODOLOGICAL APPROACH 3. Conceptual Framework The literature on supply and demand is of appreciable proportion, reflecting the variety of authors' theories as well as the diversity of markets and political philosophies. However, the health sector's share of market studies, although equally diverse in theoretical orientation, has been relatively small. Ranking high among the major reasons for its relative paucity is the difference in type of market. Labour markets, unlike commodities markets, are less likely to conform to the simple equilibrium price-auction view of the economy, the classical economic analysis, which postulates that the market operates strictly according to the principles of competitive supply and demand (13). The human capital framework forced onto labour market behavior ignores a wide variety of characteristics that make human investments different from physical investments, treating labour as just another factor of production. The productivity of human labour, for example, is not only technologically determined. Motivation and personal effort are at least as important as the technology factor. As well, human beings have preferences that are socially determined; that is, each person's likes and dislikes are highly interdependent. Relative deprivation, a sociological concept, and not absolute income, has more explanatory value in analyzing market behavior. In brief, the equilibrium price-auction view of the health market, central to supply-side economics, has not provided as many answers as the questions it has raised. Moreover, in the Canadian context where price elasticity of demand for the primary products is not particularly relevant since a universal medical insurance scheme is available, and the input demand response to a change in input prices (wages of the category under study) are generally nonexistant. Although a change in wage rates does not immediately effect a change in product prices (direct cost to the patient (or insurer?)), employer response under fixed budget conditions results either in the reduction of the particular services, or cost containment in other areas. However, the usefulness of the equilibrium-price auction approach is limited. Keynesian macro-economic models, however, take a different approach to explain demand for labour. Econometric models use aggregate output, not wage rates, and supply equations are represented by long-term demographic trends and job opportunities. Another theoretical perspective relevant to our study framework is that of institutional labour economics which focuses on inter-skill wage differentials. Totally inconsistent with standard economic theory, the former uses interdependent preferences, relative deprivation, norms of social justice and wage contours to determine the structure of wages. The inconsistent approaches and empiric.al evidence refuting some of the tenets of the standard economic model have not, however, resulted in - 5 -any reformulation or reexamination of that approach. The problem has been addressed, rather, by hypothesized "market imperfections". It is not our intent here to examine this phenomenon of inconsistent and sometimes contradictory approaches. Nor is it our intent to provide a comprehensive presentation of all theories. The purpose of the brief overview of selected approaches has been to demonstrate the variety and diversity of opinion in this area and provide a rationale for the model developed in this study. Our model is developed with the objective of contributing to -our understanding of the nurse labour market in this province. We make no presumptions about accruing possible theoretical knowledge to the field of economic modelling. Our inclination has been to use general health manpower requirements models as a backdrop, developing our specific model that is suited for a study of nursing requirements in B.C. In general, requirements for any category of health manpower have been examined as a function of at least the following broadly defined variables; the availability of: (i} complementary capital input; these are inputs that generally are used in combination with the input of interest. The latter - in this case - refers to nursing services, and the former to hospital beds, nursing homes, community clinics, as appropriate for the particular type of nursing services under study. An increase in complementary capital input availability will be accompanied by a similar increase in the demand for the input of interest; (ii} substitute capital input; these are inputs which could be considered substitutes such as ambulatory care, to the complementary capital input. Theoretically, the latter will tend to display opposite demand responses to a change in the former; for example, an increase in day care surgical facilities might be accompanied by a decrease in demand for hospital beds; (iii} complementary manpower; this is viewed as those manpower categories associated with the group under study, in the sense of complementary utilization. For example, an increase in supply of physician manpower tends to be accompanied by an increase in demand for nurse manpower, given the current roles and functions of nursing practice; (iv} substitute manpower; this is viewed as those manpower categories, usually less highly skilled and conmanding commensurately lower wages, that could possibly provide the same or similar services, such as practical nurses providing some of the services now provided by registered nurses. Relative wages and wage elasticity of supply of substitutes ' - 6 -are factors that might be expected to determine the use of substitute manpower; (v) financial resources; this factor obviously influences the structural characteristics of the delivery system such as personnel mix at the agency-level as well as type of facility mix (alternative delivery modes) at the system-level. Such structural characteristics greatly influence health services utilization rates and, ultimately, have impact on the demand for health manpower; (vi) a set of "patient" variables that would be associated with the deployment of nursing stock, such as a demographic profile of the population, utilization rates and health status. More formally, DX= f(CC1·····ccg; CS1 ••••• csh; MC1 •••• MCk;MS1 ••••• MS1; F1 ••••• Fm; P], ••••• Pn) where, Ox is the demand for health manpower category x CC1 •••• CSg are complementary capital input variables CS1 •••• csh are substitute capital input variables MC1 •••• MCk are complementary manpower variables MS1 •••• MSt are substitute manpower variables Fi •••• Fm denote financial variables and, PJ. .... Pn denote patient variables. Practical considerations and data limitations, as well as the validity of the underlying assumptions and the applicability of the theoretical framework, ultimately define and shape the working version of a model. Among the restrictive factors, data availability and estimation considerations are the crucial ones. Ideally, one would like to have the best estimate of requirements as well as accurate estimates of all the independent variables. Membership and licensure data are generally inadequate as they over or under-estimate demand (Ox) and supply, especially where unlicensed substitutes (MS) and/or complementary (MC) groups are concerned. For the capital variables (CS and CC), decisions to choose the most relevant ones are weighed against the probabilities of having access to accurate data, such as choosing between rated bed capacity and funded beds. The usual problems of patient variable identification and specification of the relevant financial variables have to be dealt with before a model can be operationalized. - 7 -Similarly, estimation considerations such as interactive effects between the independent variables have to be resolved, by addressing such questions as: Is the impact of MC variables on Dx dependent on the F variables? Finally, identification of appropriate estimation techniques is central to estimating the model, based on the nature of the model, the type of data and the questions being addressed. The data have to provide sufficient ''observa·tions'' to suppor-t the number of explanatory variables essential to the model, one needs to check for multicollinearity, etc. These will be discussed in detail in Chapter 5. We have attempted in this section to briefly discuss those among the more general issues concerning health manpower demand analyses that are particularly relevant to nurse manpower demand analyses. As well, we have presented a very brief discussion of anticipated practical limitations to this type of research. In those instances where data and methodological problems can be sufficiently overcome to affirm the reliability of results, analysis of requirements will provide the appropriate assessment tools for measuring the impact of changes in selected policy and other variables on requirements in the future. In the following section we will specifically discuss issues concerning nurse requirements. - 8 -4. Discussion of Nurse Requirements Models Studies of nurse requirements contain various operational definitions reflecting both variety in analytical approach and in modelling technique. In order to provide a comprehensive yet concise discussion of work in this area, we have provided a schematic conceptual framework (14) to depict what influences the demand for nurses {manpower and services). The schematic presentation of factors that influence demand provides not only a clear delineation of the areas of analysis but also an implicit identification of policy issues {Fig.1). The framework is intentionally simple in areas where nurse requirements are neither a direct input nor a direct output of the processes concerned such as in the area of nurse education. The explanatory value of the schematic framework is limited to those processes and issues that are directly related to the nurse manpower and nursing services markets. Clearly, this discussion of the variety of analytical approaches cannot be based on an exhaustive survey of the literature because of the interdisciplinary breadth of the subject which includes the areas of economics, planning, sociology, epidemiology and psychology. Rather, the discussion material should be viewed as a general compilation of work directly pertaining to major areas of analysis in nurse manpower. The analysis of the demand for nursing services {within a health services context and as separate from the analysis of derived demand) requires a thorough understanding of health consumers' behavior including their consumption of nursing services. To do so requires knowledge of the group's individual cultural beliefs, perceptions of health and illness and attitudes towards health care and perceptions of the effectiveness/efficiency of the particular system, including cost and quality of care. Most of the work in this area, however, focuses mainly on the cost factors of the health services market since beliefs and attitudes are the most difficult to quantify and the least amenable to government intervention except through major educational programs of prevention and health promotion. A number of models of economic behavior describe variations in the type of medical services utilized as a function of service costs, physician availability, patient income, etc. {15,16). Other models describe market interactions between the services, manpower and education markets (17,18). Although an aggregate model of the supply and demand for hospital inpatient care {19) does have some relevance to the subject under study, an econometric model of the allocation of U.S. health care resources under Medicare {20) is of no particular value to the understanding of nurse requirements in this country. Alternatively, studies on patterns of health services utilization are used to describe the influence of consumer discretionary behavior on the demand for health services and, by inference, nursing services. Since data describing the aforementioned consumer parameters are often hard to obtain, proxy measures are frequently used based on general _,.,, Cl-ZZ ow -:ic 1-W ca: o..-:::i:::i UCI uw oa: Figure 1: NURSING EDUCATION [iuRsE STOCK I MANPOWER RESOURCES SPECIALTY & GEOGRAPHIC DISTRIBUTION LABOUR-FORCE BEHAVIOUR - 9 -Conceptual Framework for Nurse Requirements Estimation DEMAND FOR NURSE MANPOWER HEALTH CARE DELIVERY HEALTH CONSUMER MEDICAL REASONS FOR HEALTH SERVICES UTILIZATION ! INCIDENCE ~i AND =~ PREVALENCE ; ~ OF ILLNESS =::111 n "' 0 BEHAVIOUR SOCIO-DEMOGRAPHIC FACTORS OF ILLNESS OCCURRENCE - 10 -information on socioeconomic, demographic and epidemiological conditions. The demand for services could then be translated, and sometimes is, into requirements for manpower, depending on the magnitude and type of services desired and the staffing patterns and size of the facilities used. Work in this area includes early modelling attempts to delineate social, economic, and demographic factors which influence the utilization of health services by family units (21) and Markov modelling to predict proportions of the population in different health service categories, to provide utilization estimates (22). More recent work in this area examines dimensions of access to medical care for the specific purpose of identifying the appropriate social indicators with which planners and policymakers could then assess future demand for health care resources (23,24,25). A large proportion of the work classified as consumer health behavior modelling could have general applications in the area of nurse requirements {26,27,28). Although some could not directly address questions of demand for manpower since the focus is mainly on the demand for facilities and/or services {26,27,29,30), others set out to explicitly describe consumer utilization of various health resources with respect to changes in specific consumer attributes (16,28). Thus research in health services utilization and research in health consumer behavior basically study the same processes from different vantage points. The latter are more interested in the socio-economic implications of human (health) behavior whereas the former are interested in the economic impact on the (health) system, or a particular institution, of human behavior. Studies from both fields could have direct applications to manpower planning. Yet another approach to modelling demand that is consumer service-based is the illness incidence model. Similar to the consumer behavior models and utilization models, it relies on the same data to determine incidence. However, the focus of incidence models is mainly epidemiological in that its outputs purportedly reflect the health status of a population rather than the utilization of health services. The ensuing models attempt to define the "need" for health care resources which, traditionally, had been the long-standing standard for ascertaining the demand for health care {31). Two early studies (32,33) provide models which relate socioeconomic and health parameters such as income, education, medical expenditures, etc. to age-adjusted and diagnosis-specific death rates. Another model, using an econometric production function for health, estimates the relationships between population characteristics and mortality rates, disability days and number of physician visits (34). In contrast, a simulation model developed for the Pan American Health Organization (16) uses life-death Markov processes to describe the life-expectancy of a population, given the pre-specified outcomes of various possible decision variables regarding fertility and mortality rates. Although illness incidence models have limited direct application to manpower requirements, recent sunmaries of research on the indicators - 11 -and correlates of health services utilization (35,36) provide a useful and different perspective on the subject. The largest proportion of models concerned with factors influencing specifically the demand for nurse manpower falls within the category of health care delivery models. However, these models are usually developed using data describing the nursing care functions of a single hospital or specific community and, only a few are from a global perspective of health services delivery using broader-based or hypothetical data. A number of models have been developed in the past twenty years regarding the allocation and scheduling of the various nurse manpower categories and their respective productivity levels. The first systematic study to develop an inpatient classification scheme relating the degree of illness to direct patient care requirements which then is used to examine the relationship between case-load and nursing workload, was done in 1960 (37). This was followed by a Markov model to predict the future number of patients in each of the categories of the aforementioned classification scheme (38). It was further used to develop a linear programming model to optimize the allocation and mix of nursing personnel to meet current and future demands (39). Among other early models in this area (40,41) which relate the number of nursing hours per patient day to personnel mix, patient mix, etc. a model was developed by researchers at the Hopital Sainte-Justine in Montreal (42a,42b) relating the number and type of nursing resources required to the number of patients in various conditions. This simulation model describes the amount of time consumed in the performance of specific tasks with variations in the personnel mix and patient mix. More recent work from the same hospital (43,44,45) provides further development of the model and its successful application since 1975 in 24 pediatric units of that hospital. Similar work has also been done in this province (46,47). These studies describe the classification of pediatric patients in two hospitals according to their care levels and nursing service requirements. Among the more general health care delivery models are those which primarily use theoretical structures to optimize personnel utilization, community benefits, or hospital outputs under hypothetical conditions (48,49,50). In particular, one nurse-specific study employs Monte Carlo simulation techniques to explore fixed staffing versus controlled variable staffing alternatives and to specify the optimum staff allocation policy to minimize costs (50). More recently, as part of a large-scale effort to construct the methodology for evaluating health care systems design, a nurse-staff-activity model was developed (51). Demand for nursing care was calculated as a function of expected patient characteristics and average times required for the performance of nursing tasks. The model was intended for use in the assessment of organizational and policy alternatives regarding nursing operation. - 12 -Finally, a number of studies describing the supply of and demand for nurse manpower develop models that are manpower market-specific. A model minimizing the difference between demand and employment subject to constraints on the percent change in number of workers, total available resources, etc., was developed for 20 health occupations including nurses (52). Another model, using regression analysis, contains 126 simultaneous equations to describe the effects of cnanges in factors influencing nurse supply and demand such as wages, employment, desired employment, vacancies, and retirements (53). The purpose of this work was to generate forecasts of the endogenous variables to provide a basis for choosing among policy alternatives, based on their respective impact on the nurse market. The latter is treated as ten separate fields of nursing: general duty nurses, head nurses, nurse supervisors, and directors of nursing service in hospitals; school nurses; nurse educators; office nurses; private duty nurses; public health nurses; and industrial nurses. The model was used to simulate the nurse market during the historical period 1947 to 1966, and to forecast the dependent variable values. Yet another model characterizing the behavior of the nurse labour market describes nurse turnover rates, number of RNs, LPNs and nursing aides as a function of patient population in various hospital services, distribution and mix of nursing personnel across these services, nurse wages and hospital discretionary policies (54). Only a few of the models examined in this review explicitly consider the behavior of more than one economic market (depicted in Figure 1) at a time. Among those which do, the relationship between the markets for education (of health professionals) and health manpower are the most likely paired choice. The continuing "shortage" of nurses to 1985 in the U.S. has been predicted by a study using data from several National League for Nursing surveys. The study examines admissions to RN schools of nursing, provides projections based on historical trends, and identifies factors influencing the severity of the nurse shortage (55). As such, this work uses a theoretical research-based modelling approach rather than a technique-oriented modelling approach. Another research-based study examines all existing evidence concernin9 market "imbalances" and draws conclusions regarding supply {56). A detailed study of nurse education/manpower market relationships for the specific purpose of assessing the impact of various policy alternatives proposes a systematic integrated health manpower planning technique and demonstrates its application to nursing personnel in one state (8). An iterative process explores the implications of alternative planning policy decision strategies intended to balance manpower supply and requirements. Alternative policy strategies include changing the scale of operations of educational institutions; interstate migration patterns; labour force participation rates; and job design of LPN and RN positions. However, population:personnel ratios are used as the basis for estimating future requirements for nurses, making the study vulnerable to the usual criticism regarding the inherent limitations of th is approach. - 13 -An extensive study, completed recently, estimating the costs and benefits of Baccalaureate education for RNs is one of the very few that examine a three way relationship of nurse education/manpower/services markets (57). The requirements model developed for this study is based on an econometric production function and future demand forecasts are based on three different scenarios comprising stable trends, or restrained growth, or demand shift due to the proposed changes in entry to practice policy. Another massive effort to perform a systematic economic analysis of the market for nurses in the U.S. has culminated in a cl ass ic study on nurse "shortages" (58). This work provides an exhaustive review of the literature and develops an eclectic model of the nurse market. The potential effects of expanding collective bargaining for nurses, the possibility of creating more competitive market conditions through nurse registries, the effects of demand increases due to Medicare or similar programs and, the potential and actual impact of nurse training legislation on the rate of increase in the supply of nurses are all examined in detail. Manpower requirements in general, and nurse requirements in particular are influenced by the complex interaction of so many economic, political and social factors that we cannot hope to quantify and fully measure all the variables concerned. Alternative methodologies for estimating requirements are of course available to the planner and the policy maker, but no one method has been identified as being superior to others. All methodologies are tools for research and the quality of the product is as much a reflection of the insight and good judgement the researcher possesses as of the excellence of the instrument (59). Although current exhaustive inventories of manpower models are not numerous, some work has been done in compiling, reviewing and/or evaluating research in this area (14,56,60). The usefulness of a model ultimately lies in its degree of relevance to the performance measure of interest. As well, the validity, generalizability and operational feasibility of a model are criteria which determine its value to planners. For example, the relevance of cultural factors to health services utilization in British Columbia may be more/less than that estimated in a model developed in another province or country. Furthermore, model validity is considered in terms of the realism and consistency of the basic assumptions and the structural relationships of the factors involved. Model generalizability is a measure of its portability, that is, its potential use for problems that are either similar, but not necessarily identical, to the one underlying the development of the model in one jurisdiction, or identical problems in other jurisdictions. Operational feasibility is technology-dependent, based on the user's computational capabilities and constraints and the model's operational requirements. Therefore, the usefulness of any model, including those discussed in this chapter, is defined by the degree to which each model can: - 14 -(i) provide an accurate representation of the real world as perceived by the user, (ii) manipulate the parameters that bear on the analysis of problems of the user's choice, and (iii) perform the intended analysis within the time and resource constraints of the user. Since most of the discussed models were based on studies done in the U.S. they generally did not meet the conditions set in (i), given the vast differences in health systems. In other cases, the studies did not meet criteria in (iii) and thus could not be adapted for specific use in this province. They do, however, provide a wide variety of parameters some of which may be considered relevant to the situation in B.C. Thus, the usefulness of these models is dependent upon the degree they meet the conditions set in (ii) above. Our model was therefore developed with an intent to overcome such limitations. The underlying assumptions pertain to this province in particular, yet are likely to apply to other Canadian provinces. The performance measure - nursing requirements for institutional care, for community health, and for teaching, is sufficiently general to have system-wide applications. Yet it could be used for estimating the demand for subspecialties within nursing services. Last, but not least, the model developed in this study is based on already available data and/or data that are easily available to planners and policy-makers. The amount of precision in results that may be lost due to expediency and practicality factors is minimal, and should neither affect the validity of the model nor its applicability. - 15 -5. Research Design and Data Components The model developed here is a simple one and does not attempt to explain all factors affecting nurse requirements. After careful consideration of substantial research evidence, a simplified approach was given preference for this initial model. The intent was to m1n1m1ze methodological and data availability problems· that would erode the reader's faith in the results. The ordinary least squares equation model was designed for multiple regression analysis. Based on evidence from other studies, and subjective judgement where quantitative evidence is unavailable, the model estimates nurse requirements using the major manpower demand components described in Fig.l. Multiple regression analysis was chosen because it is a general statistical technique through which one can analyze the relationship between a dependent (endogenous or criterion) variable and a set of independent (exogenous or predictor) variables. In our study this technique is used as a descriptive tool by which the linear dependence of one variable on others is sunmarized and decomposed. Multiple regression as a descriptive tool is most frequently used for three purposes. First, it is used to find the best linear prediction equation and evaluate its predictive accuracy. The analysis produces a prediction equation that indicates how scores on the independent variables could be weighted and sunmed to obtain the best possible prediction of the dependent variable. Statistical measures give an indication of how accurate the prediction equation is and how much of the variation in the dependent variable is accounted for by the joint influence of a set of independent variables. A second use of this technique is to simplify the prediction equation by deleting the independent variables that do not add substantially to prediction accuracy once certain other independent variables are included. A third use of this type of analysis is to control for other confounding factors in order to evaluate the contribution of a specific variable. Emphasis in this case is on the examination of particular relationships within a multivariate context as indicated by partial coefficients. A partial regression coefficient B, in an equation for three predictor variables stands for the expected change in Y with a change of one unit of x1 when x2 and x3 are held constant. If we can assume a "multivariate normal" distribution in which each variable is distributed normally about all of the others, then we can meet the three required underlying assumptions of multiple regression analysis. These comprise assumptions regarding a linear and additive-- 16 -type model, normal distributions of Vs for fixed Xs, and equal standard deviations (homoscedasticity). We shall discuss some of the methodological limitations in the section on data analysis (Chapter 6). Since different factors influence the demand for the various nursing services, we defined the three major employment sectors as the Institutional Care sector, the Community Health sector and the Teaching sector, and developed a separate requirements model for each one. This report pertains only to the analysis for the Institutional Care sector. As is customary to first modelling attempts, the early version of this model underwent several transformations before becoming fully operational. The initial model was: Nurse requirements = f (bedstock, supply of physicians, relative supply of various nurse categories, health expenditures, relative wages of nurses, patient variables). This model includes all the broadly defined factors influencing requirements (discussed in Chapter 3) and is sensitive to practical considerations such as data availability. However, the task of converting such a theoretical framework into a state of estimation readiness is not without some trials and tribulations. Since our ultimate objective was to obtain as broad a perspective as possible, nurse requirements was defined in its broadest sense to comprise the RN/LPN personnel mix. Ideally, the concept should also include all other nursing aides who provide various levels of nursing care. Realistically, in the absence of a licensing or regulatory body to ascertain size of stock, variations in occupational definitions, etc. for this latter level of nursing personnel, it was not feasible to include the aides group without drastically altering the course of the study. Thus, the operationalized model became: Where: RELPOHRS = a0 + a1ACUBEOS + a2DOCSUPPLY + a3LPNSUPPLY + a4B~DGET + asRELWAGE + a6POPAGE + a7AVLNST RELPDHRS is the ratio of RN paid hours to total (RN + LPN) nursing paid hours; ACUBEDS is the ratio of acute rated bed capacity to total rated bed capacity; DOCSUPPLY is the ratio of physicians in General Practice to Specialist physicians; - 17 -LPNSUPPLY is the number of LPNs per 10,000 population; BUDGET is the per capita hospital net income from patient services; RELWAGE is the average hourly RN wages relative to average hourly LPN wages; POPAGE is the age-sex breakdown of the population; AVLNST is the average length of patients' stay in hospital. The unit of analysis or "observation" unit was the Regional Hospital District (RHO). The endogenous variable, RELPDHRS, estimated from data on "consumed" hours of nursing services (effective demand), presented some difficulties in definition, measurement, and availability. Although one would like to be able to estimate the demand for warm bodies, practically, collecting such data posed a number of problems. The most serious among these was the human and other resources required to access payroll information of all personnel employed in a number of tertiary referral hospitals (which did not belong to the centralized payroll system we have utilized) to extract the necessary information on individuals working in the categories of nursing personnel under study. Alternatively, one could use funded positions which would require a survey of all institutions which presents the usual non-response problems of this method of data collection. As well, in some instances funded positions were a misleading measure if effective demand made it necessary for the institutions to employ a large proportion of casual nursing personnel, some of whom may have filled positions beyond those budgeted, and conversely, some may have temporarily filled vacant positions. Estimating full-time equivalents (fte's) presents similar problems of instances where the allocated fte quotient could be surpassed/unfilled or only temporarily filled for certain periods of time. We, therefore, opted for a definition which was sufficiently broad to include nursing services provided by the large casual/relief component of nurse stock. Data availability was a second, albeit important, consideration for our choice of operational definitions. RELPDHRS includes all remuneration including statutory holidays, vacation, and paid leave, and is different from worked hours which is comprised of the time actually spent on the job. ACUBEDS, as the ratio of acute to total rated bed capacity, is a measure of complementary capital input availability. A relative measure of the acute component of rated bed capacity was taken because it is by far the most nurse-intensive component, where approximately 68 percent of RNs are employed (62). As well, we were interested in exploring the impact on requirements, over time, of the variation in acute/other bed distributions. DOCSUPPLY and LPNSUPPLY were the complementary and substitute manpower availability variables, respectively. The general - 18 -practitioner/specialist ratio was proposed again in order to explore variations over time and its influence on requirements. The availability of LPNs as substitute manpower is, of course, central to the study of relative utilization patterns. Financial resource availability measures included the per capita net income of hospitals from patient services, BUDGET, and relative wages of nurses, RELWAGE, as constraining factors that would influence nursing personnel-mix in hospitals. Effective demand (patient) variables included demographic (age-sex) distributions (POPAGE) and average length of stay (l\.VLNST). The latter variable was used as a proxy measure of utilization level for the lack of a better, easily quantifiable measure for estimating volume of service. It includes length of stay in short-term Acute and Psychiatric and Newborn units. The obvious variable to use, of course, would have been per capita patient days, i.e. total patient ~population. However, this variable posed extreme problems of multicolinearity due to its high correlation with several of the independent variables. Although all of the variables were given similar careful consideration, discussion and deliberation, and various anticipated problems of methodology, data availability and theoretical relevance were given due attention, some unanticipated data peculiarities (discussed in Chapter 7 became evident only after the data preparation phase was completed. Source of Data Data were compiled from a variety of sources. We chose to use secondary data (i.e. data collected for other purposes) for several reasons, not the least of which was the evidence that it is the preferred method for modelling analysis and enhances the researcher's objectivity (63). The British Columbia Health Association (BCHA) was the primary data provider. The data transfer was made from their payroll and other personnel systems in the form of computer tapes. Individual identifiers and confidential data had been deleted and each member hospital had consented to the data release. Annual records of each employee in the two nurse categories under study who were employed by these institutions between 1979 and 1982 were obtained. Hospitals may have joined or left the BCHA payroll system at various points in the specified time span. Such inconsistencies were reconciled and the 1982 record indicated the inclusion of 117 hospitals. These comprise almost all the B.C. hospitals listed by the Hospital Programs of the Provincial Health Ministry, excluding those operated by the province. Also excluded from the analysis are a small number of Long-Term Care facilities (Personal/Intermediate Care levels) which belong to the BCHA system. Our decision to exclude them was based on a number of factors such as inconsistencies in data availability for these facilities, structural differences that would invalidate, underlying assumptions and, - 19 -most importantly, the limitations imposed by the unrepresentative nature of this small sample. A small number of tertiary referral hospitals did not, unfortunately for us, participate in the BCHA's centralized payroll system. The non-participants were contacted separately and comparable data obtained from the respective payroll/personnel departments. We attempted to obtain actual financial records as often as possible and, where necessary, used estimates based on either historical records for the specific agency from its Annual Return of Health Care Facilities' submission to Statistics Canada, or a general averaging of comparable data from agencies with similar structural characteristics which were included in the BCHA system. In the rare instance where inconsistencies between sources from the same institution occurred, these were resolved by using that which was identified as the "official" record by management personnel in that institution. The date of record for BCHA data is the end of each calendar year, December 31st. Information on patient days (of separated patients) and separations (needed to estimate average length of stay), rated bed capacity, and hospital budgets were obtained from the Annual Return of Health Care Facilities records in the form of magnetic tapes, HSl and HS2. These tapes, owned by Statistics Canada, are part of the cooperative database made available to us on an on-going basis and by special arrangement. The Annual Return of Health Care Facilities - Hospitals, Parts One and Two (Annual Returns) provides basic information of value to hospitals and provincial hospital authorities and to the two national agencies, Statistics Canada and Health and Welfare Canada. In accordance with the Statistics Act/Section .£!_, the Annual Returns are completed by all public, proprietary and federal hospitals in Canada, regardless of the hospital's status under the federal-provincial hospital insurance program. Since the Annual Returns are first sent to provincial authorities for verification who then forward them to the appropriate federal authorities, the ensuing computer tapes -- after much editing and data transformation -- are available for research purposes approximately 18-24 months subsequent to their submission. This time lag and current data availability problem was overcome by combining sources. From Hospital Programs of the Provincial Ministry of Health we obtained "recent" data contained in the pre-specified "cells" relevant to the study for the fiscal years 1980-81 and 1981-82. Other information from this same source included rated bed capacity, and patient days and separations for 1982-83, from other reporting documents available to the Ministry. During the course of the study we received from Statistics Canada HSl and HS2 tapes for 1980-81. In addition, we used already available Statistics Canada tapes for the equivalent 1979-80 information. Although sources for the different years varied, we did not have major data standardization problems since these were compiled in the same manner (same format) and for the same purpose (Annual Returns), regardless of source (Provincial or Federal level). In general, data were not released by provincial authorities unless accuracy was verified. - 20 -Population figures for the study period were interpolated, using a straight line, from 1976 and 1981 Census data. Data on supply of Physicians and Licensed Practical Nurses for each of the years under study were those published in the respective ROLLCALL editions (64) and are part of the cooperative database available to the researchers. - 21 -III. ANALYSIS OF FINDINGS 6. Descriptive Analysis of Data Our preliminary examination of hospital expend1tures affirmed the generally intuitive impression that gross salaries and wages comprise a considerably large proportion, ranging between 65-75 percent, of total hospital operating expenses (including supplies). Table l shows a detailed breakdown of such figures over the four-year study period. Representative Salaries of Selected Personnel The logical next questions, therefore, were posed: What are salary distributions for the various nurse manpower categories over time? How do these compare to other occupational categories in the health sector, or to categories with similar training and/or to traditionally female occupations? Table 2 provides a summary of the period 1976-1982. While there appears to be appreciable variation in salary among health workers, in terms of relative salary increases these variations are much less pronounced. In the period examined, the RN appears to have done not much better (69.3%) than, for example, the Health Record Administrator (66.1%) and only slightly better than the Medical Social Worker (64.7%) or the Practical Nurse (65.3%). The CPI increase for the years concerned was 76.3 percent (65). Reportedly, teachers and bank tellers have fared much better in total increments in that period than health care workers (81.3% and 85.0% salary increases, respectively). Table 3 indicates that the relatively superior salary position of medical social workers has somewhat attenuated in the last three years. They remain, however, on the top of the list presented here and are the only group of those chosen earning an income higher than the RNs. Wage growth for RNs, during a period of wage constraint (1980-1982), has been more stable than for those in other nursing categories. The table indicates that annual wage increases for PNs and Nurses' Aides eroded in the latest period. As a proportion of RN wages, other nursing personnel were earning much less in 1981 and 1982 than they had in 1976. Traditional economic analyses have shown that labour force participation rates usually are positively correlated with increased wage rates. Unfortunately, data on nurses' labour force participation rates (measured as number of personnel employed in a field divided by number trained in that field) are unavailable and could not easily be obtained. A proxy measure, labour force activity rate (number of personnel employed in the field as a percentage of the number of personnel licensed to practice in that field), was therefore estimated to provide a general idea of the traditional relationship between wages and participation rates for RNs. Table 4 shows that the labour force activity rate was very stable for the period 1975-1980 inclusive. - 22 -Table 1: Salaries as a Percent of Total Operating Expenses for B.C. Hospitals, by Regional Hospital District, 1978-1981 1978 - 1979 1979 - 1980 1980 - 1981 Regional Hospital District s s s Alberni-Clayoquot 71.84 71.95 72.23 Bulkley-Nechako 69.92 70.57 70.41 Capital 72.48 71.87 72.26 Cari boo 73.20 73.12 72.68 Central Coast a.no 0.00 o.oo Central Fraser Valley 73.30 72.55 74.28 Central Kootenay 71.55 71.08 71.67 Central Okanagan 73.89 73.66 73.06 Columbia-Shuswap 70.43 70.64 71.35 Comox-Strathcona 73.35 73.29 72.08 Cowichan Valley 73.46 74.09 73.85 Dewdney-Alouette 73.81 73.10 73.09 East ICootenay 71.77 71.09 71.43 Fraser-Cheam 73.15 72.07 72.31 Fraser-Fort George 70.92 70.01 67.93 Greater Vancouver 72.44 72.33 71.46 Ki ti ma t-S ti kine 66.98 66.43 67.75 Kootenay Boundary 72.11 71.03 71.85 Mount Waddington 72.91 74.50 75.33 Nanaimo 74.54 73.45 73.93 North Okanagan 70.84 71.63 72.58 Okanagan-Sim11 kameen 71.13 70.83 71.06 Peace River-Liard 70.62 69.06 69.95 Powel 1 River 73.37 72.90 72.77 Skeena-Queen Charlotte 71.47 70.63 71.49 Squamish-Lfllooet 71.81 71 .51 71.00 Stikfne o.oo o.oo 0.00 Sunshine Coast 70.19 71.16 71.02 Thompson-Nicola 70.65 70.39 70.53 Table 2: Representative Annual Starting Salaries of Selected Personnel in B.C •• 1976-1982 TEAR Registered - Practical Nurses' Health Record Medical Social All Snings I Credit Food Nurses• Nurses2 Aides'•' Achin. 5 WorkersS Teachers' Workers' Canners' 1976 13.464 11 .532 10,572 12,912 14,658 10, 141 11 ,187 10,947 1977 14,208 12,225 11,412 13.686 15,537 10,923 11.787 10,749 1978 14,772 12,678 11,868 14,232 16,158 11,579 12,622 10,954 1979 15,660 13,566 12,699 15,264 17,328 12,494 13,543 12,447 1980 18,440 16,344 14,700 16~485 19,434 13,691 14,872 12,862 1981 21,216 17,652 15,876 19,467 21,918 15,704 18,008 15,739 1982 22,800 19,064 17, 146 21,447 24,144 18,387 20,692 18,112 I increase (1976-82) 69.31 65.31 62.21 66.11 64.71 81.31 85.0I 65.51 l Data provided by 8CNU (for General Duty RN, step 1). 2 Data provided by Health Labour Relations Association of B.C. (for 11inimin starting salary rates). The 1980 rate htc:ludes January and August increases; the 1981, 1982 rates are as of August of the respective years. !I Canada Department of labour, WAGE RATES, SALARIES AND HOURS OF LABOUR, 1970-1973, unpublished data for 1974-1977 pertod provided by Canada Departllent of labour. Annual salaries estimated from average monthly salaries. It Increases estimated according to 1978-81 Master Agreenent between Hospital Eillployees Union and Health Labour Relations Association of B.C. 5 Data provided by Health Sciences Assoc:1ation (Grade 1, 1st year salaries). g Data provided by B.C. Teachers' Federation (11tnimU11 baste salary). 1 Statistfc:s Canada, Elaplolll!ent, Earnings and Hours, Cat. 72-002. Annual figures estimated fro11 average weekly wages. N w Table 3: Representative Annual Salaries of Selected Personnel as a Percentage of Registered Nurses' Salaries, 1976-1982 YEAR 1976 1977 1978 1979 1980 1981 1982 l 2 3 .. 5 6 7 Registered Practical Nurses Hea 1 th Record Medical Social All Savtngs I Credit Food Nursesl Nurses2 Aides3,lt Admtn.5 Workerss Teachersli Workers7 Canners' lOM 85.71 78.51 95.91 108.91 75.31 83.lS 81.31 lOM 86.0S 80.31 96.31 109.41 76.91 83.0S 75.71 1001 85.81 80.31 . 96.31 109.41 78.41 85.41 74.21 1001 86.61 81.11 97.51 110.71 79.BS 86.51 79.51 1001 88.61 79.7S 89.41 105.41 74.21 80.71 69.BS lOOS 83.21 74.81 91.81 103.31 74.01 84.91 74.21 1001 83.61 75.21 94.lS 105.91 80.61 90.81 79.41 Oita provfded by BCNU (for General Duty RN, Step 1). Data provided by Health Labour Relations Associatfon of B.C. (for •fnilllUlll starting salary rates). The 1980 ratefncludes January and August increases; the 1981, 1982 rates are as of August of the respective years. Canada Departllent of Labour, WAGE RATESf SALARIES AND HOURS OF LABOUR, 1970-1973, unpublfshed data for 1974-1977 perfod provided by Canada Deparbnent of Labour. Annual sa artes estimated fT'Olll average monthly salaries. Increases estfmated according to 1978-81 Master Agreement between Hospital E'lllployees Unfon and Health Labour Relatfons Assoc1at1on of B.C. Oita provfded by Healbh Sciences Association (Grade 1, 1st year salaries). Dita provided by 8.C. Teachers' Federatfon (mini•1111 basic salary). Statistics Canada, EllploJll!ent, Earnings and Hours, Cat. 72-002. Annual figures estimated from average weekly wages. N ~ - 25 -Table 4: Labour Force Activity Rates for RNs, B.C., 1975 - 19831 YEAR Registered with Employed in Activity Practising Status Nursing Rate ! 1975 14,066 12,347 87.8 1976 - - -1977 15,239 13,493 88.5 1978 - - -1979 17,390 15,389 88.5 1980 18,712 16,458 88.0 1981 19,962 18,556 93.0 1982 20,608 18, 191 88.3 1983 20,374 19, 184 94.2 1 Compiled from ROLLCALL Reports of the respective years. Figures represent RNABC Membership status as at July of each year for 1975 - 1980, June of each year for 1981 - 19~3 . Table 5: labour Force Activity Rates for LPNs, B.C. 1979 - 19831 YEAR Licensed to Practice £mployed in Activity in B.C. Nursing Rate ! 1979 6,511 4,270 65.6 1981 6,991 4,943 70.7 1982 7,024 4,754 67 .7 1983 6,891 4,777 69.3 1 Compiled from ROLLCALL Reports of the respective years. Figures represent manbership with the Council of Practical Nurses as at Septanber of each year. - 26 -However, 1981 witnessed a large activity rate increase followed by a decrease in 1982, immediately followed by another increase in 1983. Clearly, the erratic period of recent years is the outcome of a number of factors as evidenced by the registration figures. In 1981, the number of nurses registered as practising members of the Registered Nurses' Association of British Columbia (RNABC) had increased by 6.7 percent over the 1980 figures while the proportion employed in nursing during the same period had increased by 12.7 percent. As membership figures continued to rise modestly in 1982 (3.2%), employment figures declined due to budgetary restraints and various measures of cost containment that the hospital sector undertook during that year (66). The ensuing rate for 1983 is artificially high (as a proxy for participation rate) at 94.2 percent, due to an unprecedented (at least in the last eight years) decrease in membership. It is interesting to note that had membership figures increased at the 1980-81 rate of 6.7 percent, the activity rate for 1983 would have been approximately 87.0 percent, in line with previous years. Even a moderate 3.2 percent increase in membership (as in 1981-82) would have resulted in a more comparable 90.0 percent activity rate. In surrmary, data from Tables 3 and 4 indicate that when RNs had the edge over their other nursing counterparts in terms of relative wages in 1981, the RN activity rate climbed up in astounding proportions. Although a similar RN" activity rate has been estimated in Table 4 for 1983, caution should be exercised in interpreting these results. In particular, opportunity structures in the nurse labour market have undergone a number of changes in the past two years in terms of funded positions, vacancies, and the incidence of part-time employment (67). When employment opportunities are perceived to be limited, those who are currently unemployed are less likely to renew their professional membership in the "Practising" category. Thus, the "Non-Practising" membership category decreased between 1981-82 and subsequently increased in 1983 (68). A brief examination of the same information on licensed Practical Nurses indicates a very similar but somewhat less pronounced trend in the four years for which reliable data for this group were available. Although the labour force activity rates are much lower for LPNs than for RNs, a similar increase in the number licensed in 1981 with a concurrent increase in the activity rate are discernible for LPNs as well. The latter have also experienced a drop in membership for the year 1983. Relative Salaries of Nursing Personnel The next question addressed pertained to actual relative salaries. In the light of the information in Tables 2 and 3, what were the actual relative salaries paid to RNs and LPNs given the manpower composition of each group for each of the years under study? - 27 -Relative wages were estimated as: "RN E RNSAL1 1~ "RN ~ RNHRS1 RELWAGE • "PN E PNSAL1 1~ "PN E PNHRS1 1~ where: RELWAGE is the average hourly RN wages relative to average hourly PN wages; RNSAL is the biweekly salary of each RN; RNHRS is the biweekly hours worked by the RN; PNSAL is the biweekly salary of each PN; NHRS is the biweekly hours worked by the PN. Table 6 provides a regional overview. The estimated provincial mean shows that 1980 was the lowest year; the average RN salary paid was 20 percent more than the average salary paid to practical nurses. The following year, 1981, marks the highest year of the four-year period with the average RN salary climbing to 28 percent more than that of the practical nurse. Almost the same margin is maintained for 1982. The table also provides useful information on a regional basis in terms of which Regional Hospital Districts diverge appreciably from the annual provincial means or show a discernible four-year pattern different from ·the overall provincial pattern. Kitimat-Stikine, for example, appears to have much lower than average relative salaries and the year to year variances are in the opposite direction from those of the provincial average. Columbia-Shuswap, on the other hand, appears to be on a straight-line incrementation trend reporting, in 1982, an average RN salary one and one-half times that of the practical nurse. Conversely, the average RN salary in Squamish-Lillooet in 1980 was 14 percent lower than that of the practical nurse. Greater Vancouver and Capital Regional Hospital Districts, due to the large numbers of stock, are the - 28 -Table 6: Average Relative Sa1ariesl of Registered Nurses to Practical Nurses in B.C., by Regional HOsp1ta1 District, 1979 - 1982 REGIONAL HOSPfTAL DISTRfCT 1979 1980 1981 1982 Alberni-Clayoquot 1.34 1.24 1.30 1.31 Bul kl ey-Nechako 1.20 1.10 1.25 1.26 Capital 1.24 1.24 1.28 1.29 Cari boo 1.31 1.30 1.39 1.30 Central Coast 0.00 0.00 0.00 0.00 Central Fraser Valley 1.01 1.03 0.96 1.15 Central Kootenay 1.33 1.27 1.31 1.24 Central Okanagan 1.19 1.18 1.34 1.30 Columbia-Shuswap 1.14 1.38 1.46 1.51 Comox-Strathcona 1.39 1.26 1.23 1.17 Cowfchan Valley 1.09 1,24 1.20 1.11 Dewdney-Alouette 1.18 1.26 1.38 1.31 East Kootenay 1.36 1.22 1.12 1.30 Fraser-Cheam 1.10 1.05 1.24 1.49 Fraser-Fort George 1.30 1.20 1.29 1.21 Greater Vancouver 1.27 1.25 1.29 1.26 Kitimat-Stikine 1.12 1, 16 1.08 1.17 Kootenay Boundary 1.18 1.19 1.36 1.32 Mount Waddington o.oo o.oo o.oo 1.33 Nanaimo 1.31 1.22 1.30 1.28 North Okanagan 1.25 1.30 1.33 1.42 Okanagan-Similkameen 1.24 1.25 1.33 1.35 Peace River-Liard 1.26 1.10 1.45 1.28 Powel 1 River 1.21 1.27 1.38 1.40 Skeena-Queen Charlotte 1.20 1.26 1.27 1.33 Squamish-Lillooet 1.06 0.86 1.03 0.94 Stikine o.oo o.oo 0.00 0.00 Sunshine Coast 1.23 1.07 1.23 1.16 Thompson Nicola 1.19 1.19 1.37 1.25 Provincial Mean 1.22 1.20 1.28 1.27 1 (RN salaries/RN hours)/(PN salaries/PN hours) - 29 -major contributors to the provincial average and as such are quite similar to mean figures. Two interrelated variables influence these regional relative salary variances, statistically speaking. The first is size of stock, the second occupational level of stock in the region. A small RN stock, the majority of whom are in the higher salary brackets, compared to a large number of new practical nurses in a region, provides the extreme example of higher than average relative salaries. It is therefore important to interpret Table 6 in conjunction with additional information from Tables 7-14 on nurse paid hours distributions and Table 17 on rated bed capacity (total beds including bassinets). The latter table provides some perspective on the scope for potential utilization of nursing services. Tables 7-14 provide a detailed account of nurse deployment patterns on a regional basis. Interestingly, while RN relative salaries increased in 1981 and remained equally high in 1982, the ratio of registered/practical nurse paid hours steadily increased during the four years under study from 2.35 to 2.97 in favour of the RNs. The 7 percent increase in average relative salary costs during 1980-81 (Table 6) does not appear to have been a deterrent in the continuous proportional increase of RN paid hours (Table 16). More importantly, Tables 7-14 provide a clear picture of the differential service utilization patterns of the two nurse categories in terms of regular/casual full-time/part-time employment. For RNs (Tables 7-10), 11 regular 11 paid hours steadily decreased from 84.1 percent of total paid hours in 1979 to 80.3 percent in 1982. Conversely, 11 part-time11 hours increased proportionately during that same four-year period while 11 casual 11 hours picked up the remainder of the decreased regular hours in 1980 and 1981. The 1982 cost-containment measures appear to have affected 11 casual 11 hours the most. Table 15(a) provides a concise yet revealing picture of the scope of change in each of the three categories of paid hours. Clearly, the 1982 decrease in total paid hours was mainly due to the large decrease in 11 casual 11 hours and only a slight decrease in 11 regular 11 hours. Part-time hours were still on the increase, albeit in reduced proportions. Tables 11-14 describe the situation for PNs, somewhat similar in trend to that of the RNs in terms of the proportionate mix of regular, part-time, and casual paid hours. Table 15(b) indicates, however, that part-time hours were more preponderant among PNs than RNs and that the 1981-82 crash has hit the former harder than the latter in all three categories of paid hours. Table 16 provides a 11 relative 11 view of the situation. The data indicate that, while in 1979 there were slightly more than two RN paid hours (2.35) for each one hour of PN, in 1982 that ratio had climbed up to almost three (2.97) RN hours to one PN hour. - 30 -Table 7: Nurse Pa~d Hours Re ular Part-Time Casual B.C. I 9 REGIONAL HOSPITAL DISTRICT -Regular Alberni-Clayoquot 164,995 Bul kl ey-Nechako 132, 157 Capital 2, 154.329 Cari boo 244,183 Central Coast 0 Central Fraser Valley 473,320 Central Kootenay 174.562 Central Okanagan 402.953 Columbia-Shuswap 155,572 Comox-Strathcona 331,756 Cowichan Valley 248,597 Oewdney-Alouette 261,794 East Kootenay 239,064 Fraser-Cheam 251,928 Fraser-Fort George 464,864 Greater Vancouver 9,006,442 Kitfmat-Stikine 207,896 Kootenay Boundary 268,226 Mount Waddington 0 Nana imo 382,372 North Okanagan 302,007 Okanagan-Similkameen 360,208 Peace R1ver-L1ard 167 ,695 Powell River 134,395 Skeena-Queen Charlotte 139,740 Squamish-Lillooet 47,030 Stikine 0 Sunshine Coast 67,592 Thompson-Nicola 600,308 Totall 17,383,984 I (84.U) Registered Nurses Part-Time 16,218 15,537 108,838 30,173 0 49,099 37,021 43,932 20.567 54,393 25.519 31.634 29, 160 22.568 28,242 700,271 17 ,712 60,643 0 21,178 20,009 16,609 18,296 12,891 6,636 5,647 0 8,786 53,538 1,455,115 (7.01) 1 Row and column totals 111y not add due to rounding. - -TOTALl Casual 31,933 213,146 24,246 171,940 ll5,398 2,378,565 37.746 312,102 0 0 103,392 625,811 21,538 233,121 56,220 503,105 22,898 199,037 47,993 434, 142 58,817 332,933 47,832 341,260 36,484 304,708 46,080 320,576 69,282 562,388 737,910 10,444 ,623 22, 111 247,719 30,416 359,285 0 0 64,558 468, 108 49,040 371,056 57,333 434, 150 18,480 204,471 20,288 167 ,574 17,394 163 ,770 8,946 61,623 0 0 8,959 85,337 77,999 731,845 1,833,290 20,672,389 (8.91) (100.0S) - 31 -Table 8: Nurse Paid Hours Re ular Part-Time Casual .c. p 98D ercent totals b Re ional Hos ital District REGIONAL HOSPITAL DISTRICT Registered Nurses TOTAL1 Regular Part-Tt111e Casual Alberni-Clayoquot 175.564 20.335 30.983 226,882 8ul kley-Nechako 130.761 18,975 18,193 167 ,929 Capital 2.342,569 286,825 323,698 2.95;s,D~2 Cari boo 253,828 38,562 37.544 329,934 Central Coast 0 0 0 D Central Fraser Va 11 ey 535,473 67.046 107.188 7D9,7D7 Central Kootenay 239,613 48.972 39.331 327,916 Central Okanagan 430,089 47,667 65.692 543,448 Columbia-Shuswap 161,544 18.416 3D.826 21D,786 Comox-Strathcona 348.565 57,274 54.64D 46D,479 Cowithan Valley 283,506 40,826 63.614 387,946 Dewdney-Alouette 3D5.526 49.430 58.541 413,497 East Kootenay 253.071 33,493 42, 156 328,72D Fraser-Cheam 251,574 29.179 46,799 327,552 Fraser-Fort George ... 512, 101 37,742 7D,535 62D,378 Greater Vancouver 9,746.33D 824,901 872.144 11,443,375 K1t1mat-Stikine 195,756 23,139 24,0D2 242,897 Kootenay Boundary 275,912 51.918 35,632 363,462 Mount Waddington 29,09D 3,6DD 10,046 42,736 Nana1mo 395,431 34,543 69,799 499,773 North Dkanagan 323,682 22,874 54,133 4DD,689 Okanagan-S1m11kameen 372,883 19,78D 63,993 456,656 Peace R1ver-L1ard 165,4D3 13,883 19,503 198,789 Powell River 134,141 12,455 19,12D 165,716 Skeena-Queen Charlotte 135,99D 10,035 17 ,717 163,742 Squamish-Lillooet 51,158 6,988 14,35D 72 ,496 St1k1ne 0 0 0 D Sunshine Coast 73,671 6,2D1 11,351 91,223 Thompson-Ni co la 676,633 57,874 68,649 8D3,156 Totall 18,799,856 1,882,933 2,270,178 22,952,967 I (81.H) (8.21) (9.91) (1 DD. DI) 1 Row and col1111n totals •Y not add due to rounding . - 32 -Table 9: Nurse Paid Hours, Regular, Part-Time, Casual, (percent totals) by Regional Hospital District, B.C., 1981 Registered Nurses REGIONAL HOSPITAL DISTRICT TOTAL1 Regular Part-Time Casual Albernt-Clayoquot 183,889 26,452 26,634 236,975 Bul kl ey-Nechako 135,085 15,689 23,195 173,969 Capital 2,552,945 344,943 376,413 3,274,301 Cari boo 259,125 51,160 32,830 343, ll 5 Central Coast 0 0 0 0 Central Fraser Valley 583,474 87,201 119,803 790,478 Central Kootenay 316,703 801709 51,520 448,932 Central Okanagan 475,986 59,784 81,076 616,846 Columoia-Shuswap 174,980 20,597 35,370 230,947 Comox-Strathcona 388,568 64.507 75,553 528,628 Cowichan Valley 310, 139 50,590 62,707 423,436 Oewdney-Alouette 317.802 56,727 67,380 441,909 East kootenay 291 ,711 44, 181 47,676 383,568 Fraser-Cheam 282,440 38,650 53, 109 374,199 Fraser-Fort George 568,261 50,780 73,180 692,221 Greater Vancouver · 10,130,473 978,866 1,006,332 12, 115,671 Kitimat-Stikine 220,134 28.378 33,051 281,563 Kootenay 8oundiry 312.357 52,644 41,924 406,925 Hount Waddington 43,862 3,704 10,716 58,282 Nanaimo 418,828 47,902 60,672 527,402 North Okanagan 344,144 28,388 58, 131 430,663 Okanagan-Similkameen 393,792 23,333 70, 178 487,303 Peace River-Liard 213,676 23,535 16, 741 253,952 Powell River 138,736 12,504 19,944 171, 184 Skeena-Queen Charlotte 147,456 17,443 13,738 178 ,637 Squamish-Lillooet 52,915 9,528 14,851 77 ,294 Stikine 0 0 0 0 Sunshine Coast 82,599 16,850 11, 143 ll0,592 Thompson-Nicola 828,737 68,613 66,743 964,093 Total l 20,168,816 2,303,656 2,550,607 25,023,079 I (80.61) (9.21) (10,21) (1 DO.OS) 1 Row and column totals 111y not add due to rounding. I i : ' ' i I I ' i I I ' i I I I ' I - 33 -Table 10: Nurse Paid Hours Re ular Part-Time Casual .c., 1982 Registered Nurses REGIONAL HOSPITAL DISTRICT Regular Part-Tf111 Alberni-Cla,yoquot 163,332 26,470 Bul kl ey-Nechako 126,936 8,553 Capital 2,508,027 411,415 Cari boo 254,692 60,634 Central Coast 0 0 Central Fraser Valley 590,134 107,563 Central Kootenay 299, 141 70,607 Centra 1 Okanagan 479,246 81,047 Columbia-Shuswap 165,298 27,548 Coruox-Strathcona 379,754 70,665 Cowichan Valley 297,727 58,087 Dewdney-Alouette 311,542 59, 100 East Kootenay 280,097 35,461 Fraser-Cheam 280,342 43,743 Fraser-Fort George 558, 184 63, 191 Greater Vancouver 10,170,008 1,094,163 Kitimat-Sti kine 207,034 28,643 Kootenay Boundary 313,267 57,789 Mount Waddington 51,228 8,670 Nanainio 392,374 45,614 North Okanagan 349,494 41,728 Okanagan-Similkameen 392,404 20,430 Peace River-Liard 226,949 29,143 Powell River 126,668 20, 164 Skeena-Queen Charlotte 133,183 12,544 Squamish-Lillooet 50,676 13,596 Stikine 0 0 Sunshine Coast 82,441 20,396 Thompson-Nicola 816,552 78,599 Total 1 20,006,720 2,595,567 I (80.31) (10.41) 1 Row and column totals .. Y not add due to rounding. TOTAL1 Casual 21,886 211,688 23,287 158,776 336,466 3,255,908 35,582 350,908 0 0 92,583 790,280 50,574 420,322 70,075 630,368 31,670 224,516 71, 185 521,604 53,838 409,652 56,705 427,347 47,639 363,197 40,205 364,290 87,133 708,508 899,080 12,163,251 25,722 261,399 41,250 412,306 11,323 71,221 54,469 492 ,457 50,960 442, 182 77,683 490,517 18,663 274,755 15,774 162,606 13,460 159,187 14,069 78,341 0 0 9,900 112,737 67,221 962,372 2,318,402 24,920,689 (9.31) (100. 0%) - 34 -Table 11: Nurse Paid Hours, Regular, Part-Time, Casual, (percent totals) by Regional Hospital District, B.C., 1979 --Practical Nurses REGIONAL HOSPITAL DISTRICT TOTAL1 Regular Part-Time Casual Alberni-Clayoquot 74,739 10,021 16,722 101,482 Bul kl ey-Nechako 76,921 5,816 13,825 g6,562 Capital 870,390 41,831 19,042 931,263 Cari boo 144,982 14,718 26,647 186,347 Central Coast 0 0 0 0 Central Fraser Ya 11 ey 175,056 18,611 20,406 214 ,073 Central ICootenay 108,367 22,013 14,316 144,696 Central Okanagan 262,794 19,556 33,068 315,418 Columbia-Shuswap 68,611 6,448 6,927 81,986 Comox-Strathcona 185,457 22, 110 30,353 237,920 Cowfchan Valley _ 159,718 4,792 36,407 200,917 Oewdney-Alouette 133,825 21,518 15,236 170,579 East ICootenay 96,400 18,331 12,356 127,087 Fraser-Cheam 104 ,916 3,394 16,361 124,671 Fraser-Fort George 150,888 24,531 16,511 191,930 Greater Vancouver 3,776,905 237,823 189,537 4,204,265 Kiti111at-Stikine 86,888 B,789 5,983 101 ,660 ICootenay Boundary 134,454 21,124 12 ,281 167 ,859 Mount Waddington 0 0 0 0 Nanaimo 180,624 7,071 15,621 203,316 North Okanagan 165,486 13, 195 15,331 194,012 Okanagan-Similkameen 207 ,594 19,629 23,756 250,979 Peace River-Liard 67,405 2, 181 11, 158 80,744 Powell River 51,539 715 4,005 56,259 Skeena-Queen Charlotte 67,555 1,605 5,863 75,023 Squamish-Lillooet 22,926 3,317 2,286 28,529 Stikine 0 0 0 0 Sunshine Coast 30,294 1,099 5,511 36,904 Thompson-Nicola 227,939 35,689 15,668 279,296 Total 1 7,632,675 585,928 585,177 8,803,780 ' (86.71) (6.71) (6.61) (100.0S) 1 Row and col1111n totals 1111 not 1dd due to rounding. - 35 -Ta~le 12: Nurse Paid Hours, Re ular, Part-Time, Casual, ( ercent totals) b Re ional Hos ital District, B.C., 80 Pra~tical Nurses REGIONAL HOSPITAL DISTRICT TOTAL1 Regular Part-Time casual Alberni-Clayoquot 70,408 12,218 17,309 99,935 Bul kl ey-Nechako 78,183 8,812 10,009 97,004 capital 744,638 135,690 93, 147 973,475 Cari boo 139,649 12,356 30,550 182,555 Central Coast 0 0 0 0 Central Fraser Valley 189,206 20,176 24,667 234,049 Central Kootenay 134,244 28,035 17,659 179,938 Central Okanagan 281,091 22,647 36,287 340,025 Columbia-Shuswap 72,280 6,687 9,788 88,755 Comox-Strathcona 186,938 22,774 30,413 240,125 Cowichan Valley 177 ,800 12,031 40,069 229,900 Dewdney-Alouette 161,240 20,435 21,058 202,733 East Kootenay 109,000 24,214 18,099 151,313 Fraser-Cheam 110,116 3,289 17 ,033 130,438 Fraser-Fort George 148,353 23,080 14,668 186, 101 Greater Vancouver 3,853,906 294,956 204,294 4,353,156 K1 ti1111t-St i k1ne 92,933 10,301 9,912 113, 146 Kootenay Boundary 130,875 14,863 16,410 162, 148 Mount Waddington 0 0 0 0 Nanaimo 183,991 31,867 18,620 234,478 North Okanagan 158,624 13,774 13,022 185,420 Okanagan-Similkameen 218,758 24,014 23,068 265,840 Peace River-Liard 82,341 6,476 13,322 102,112 Powell River 50,324 0 3,767 54,091 Skeena-Queen Charlotte 74,237 1,658 5, 189 81,084 Squam1sh-Lillooet 21,656 2,049 3,292 26,997 Stfkine 0 0 0 0 Sunshine Coast 30,721 1,566 4,528 36,815 Thompson-Nicola 255,853 25,834 20,016 301 ,703 Total l 7,757,365 779,800 716, 193 9,253,358 I (83.81) (8.51) (7.71) (100.01) 1 Row and column totals may not add due to rounding. - 36 -Table 13: Nurse Patd Hours, Regular, Part-Time, Casual, (percent totals) by Regional Hospital District, B.c •• 1981 Practical Nurses REGIONAL HOSPITAL DISTRICT TOTAL1 Regular Part-Time Casual Albernf-Clayoquot 69,194 20,990 5,904 96,088 Bu1 kl ey-Nechako 80,563 12,816 14,498 107,877 Capital 762,257 151,279 96,821 1,010,357 Cari boo 150,508 14,977 33,328 198,813 Central Coast 0 0 0 0 Central Fraser Valley 198, 127 29,317 21,550 248,994 Central Kootenay 160,002 32,724 19,080 211,806 Central Okanagan 301, 150 32,348 38,034 371,532 Columbfa-Shuswap 75,622 6,525 8,933 91,080 Comox-Strathcona 94,603 17,457 14,811 126,871 Cowfchan Valley 188,309 15,546 42,021 245,876 Dewdney-Alouette 166, 185 17 ,303 16,244 199,732 East Kootenay 129,330 28,807 22,799 180,936 Fraser-Cheam 128,854 3,850 21,051 153,755 Fraser-Fort George 148,091 22,957 12,450 183,498 Greater Vancouver 3,731,726 379,668 181,286 4,292,680 Kftf111at-Stf kfne 106,204 12,004 12,360 130,568 Kootenay Boundary 155, 112 16,315 22,404 193,831 Mount Waddington 3,616 1,168 1,385 6, 169 Nanafrao 191 ,639 40,359 17,217 249,215 North Okanagan 157,920 16,443 13, 177 187,540 Okanagan-Sfmilkameen 233,189 32,404 25,913 291,506 Peace Rfver-Lf ard 95,680 11,540 12,860 120,080 Powell Rfver 51,536 85 6,148 57,769 Skeena-Queen Charlotte 74,084 540 8,027 82,651 Squamfsh-Lfllooet 20,388 1,842 2,062 24,292 Stf kf ne 0 0 0 0 Sunshine Coast 31,017 4,313 5, 180 40,510 Thompson-Nicola 286,415 19,716 15,551 321,682 Total 1 7 ,791,322 943,294 691,092 9,425,708 s (82.JS) (1 O.OS) (7.3S) (1 DO.OS) I Row and col1111n totals uy not add due to rounding. ' - 37 -Table 14: Re ular Part-Ttme Casual Practical Nurses REGIONAL HOSPITAL DISTRICT TOTALl Regular Part-Time Casual Alberni-Clayoquot 53,742 17,692 4,736 76,170 Bul kl ey-Nechako 64,685 17 ,615 8,922 91,222 Capital 833,359 173,264 79,967 1,086,590 Cari boo 128,565 12,687 25,452 166,704 Central Coast 0 0 0 0 Central Fraser Valley 166,560 27,298 7,123 200,981 Central Kootenay 149,383 34,095 18,483 201 ,961 Central Okanagan 281,063 38,733 31,669 351,465 Columbia-Shuswap 73,976 9,231 6,896 90, 103 Comox-Strathcona 93,883 22,629 14,506 131,018 Cowichan Valley 163,251 12,285 38,726 214,262 Dewdney-Alouette -.... 136,6)7 23,193 15,296 175, 106 East Kootenay 98,239 14,919 16,524 129,682 Fraser-Che am 123,339 7,840 17,374 148,553 Fraser-Fort George 115,876 3,661 17,986 137,523 Greater Vancouver 3,184,487 320,385 105,820 3,610,692 Kitimat-Stikine 100,256 12,228 ll ,023 123,507 Kootenay Boundary 152,842 16,300 19,416 188,558 Mount Waddington 4,353 998 518 5,869 Nanaimo 152. 784 36,530 13,624 202,938 North Okanagan 129,425 9,208 12,807 151,440 Okanagan-Similkameen 227,864 31,428 21,357 280,649 Peace River-Liard 92,236 12,040 ll ,315 115,591 Powell River 46,856 l, 187 5,525 53,568 Skeena-Queen Charlotte 68,682 211 7,396 76,289 Squamish-Lillooet 19,377 2,630 1,521 23,528 Stikine 0 0 0 0 Sunshine Coast 25,235 4,345 3,403 32,983 Thompson-Nicola 280,390 30,853 13,975 325,218 Total 1 6,967,325 893,482 531,356 8,392, 163 s (83.0S) (10. 7S) (6.3S) (100 .OS) 1 Row and col1111n totals may not add due to rounding. - 38 -Table 15: Annual Percent Changes for Nurse Paid Hours, in B.C., Regular, Part-Time, casual, 1979-1982 Registered Nurses YEAR Regular Part-Time casual 1979-80 8.1 29.4 23.8 1980-81 7.3 22.3 12.4 1981-82 -o.8 12.7 -9.1 Practical Nurses YEAR Regular Part-T1me Casual 1979-80 1.6 33.1 22.4 1980-81 0.4 21.0 -3.5 1981-82 -10.6 -5.3 -23.1 Table 16: 1979 1980 1981 1982 Registered Nurses 20,672,389 22,952,967 25,023,079 24,920.689 (11.0S) (9.0S) (-0.4S) Practical Nurses 8,803,780 9,253,358 9,425,708 8.392,163 ( 5.lS) (1.9S) (-11.0S) RN/PN 2.35 2.48 2.65 2.97 Ffgure 2: Annual Percent Changes for Nurse Pafd Hours fn B.C. Regular, Part-Tfme, casual, 1979-1982 B. PRACTICAL NURSES M ~ ~ ~ n ~ ~ ~ ~ ~ z z ~ ~ ~ ~ u u ~~ ~,.,1 '- " I ~ u 'J u 5 ffi ffi 11. ,, 11. 0 ~ ~ ~ _: ~ -~j "'~ Legend -l'J -15 ma. PAllT-TK ~ ~ ' Sff!'&.. -:~ -21J I I 1979-80 1980-81 1981-82 1979-80 1980-81 1981-82 TEARS \'EARS - 40 -Table 17: Rated Hospital Bed Capacity!, by Regional Hospital District, B.C., 1979-1982 --REGIONAL HOSPITAL DISTRICT 1979 19BO 1981 1982 Alberni-Clayoquot 199 199 199 199 Bul kl ey-Nechako 236 241 241 241 Capital 2.756 2.714 2.76B 2,715 Car1boo 305 289 289 289 Central Coast 58 51 41 41 Central Fraser Valley 572 570 570 645 Central Kootenay 399 407 407 395 Central Okanagan 477 477 527 527 Columbia-Shuswap 204 203 203 203 Comox-Strathcona 362 362 362 379 Cowichan Valley 357 350 344 343 Dewdney-Alouette 315 339 339 339 East Kootenay 366 37.0 370 370 Fraser-Cheam 352 352 352 353 Fraser-Fort George 455 508 508 516 Greater Vancouver 8,591 8,417 8,968 9,060 Kitimat-Stikine 339 341 284 282 Kootenay Boundary 356 356 356 344 Mount Waddington 102 92 92 91 Nanaimo 402 398 398 398 North Okanagan 329 319 319 328 Okanagan-Similkameen 426 428 428 433 Peace R1ver-Liard 371 362 362 362 Powell River 150 142 142 142 Skeena-Queen Charlotte 205 205 205 203 Squamish-Lillooet 71 71 71 71 Stikine 9 9 9 9 Sunshine Coast 78 78 78 78 Thompson-Nicola 661 658 658 675 Total 19,503 19,308 19,890 20,031 1 Source: Hospital Programs, B.C. M1n1stry of Health. - 41 -Estimation of FTEs The data in Tables 7-16 can also be used to estimate full-time equivalent (FTE) figures for each of the categories and years. However, this deceptively simple-sounding exercise has more than one caveat. The estimated FTEs should not be taken as a firm measure of funded positions because they are mare than "worked" hours and contain compensation components such as vacation, leave, and holidays. While we do have data on actual "worked" hours, it is impossible to ascertain from the aggregate data what proportion of those worked hours are the relief or replacement components. There is no systematic documentation of this "relief" phenomenon and our research indicates that it varies according to year, size of hospital and level of care, and management style. Thus, it is not a simple task to calculate the difference between FTEs of paid hours (70) and worked hours to arrive at an estimate for absenteeism rates and current funded positions. The data have to be disaggregated to the level of each individual nurse and a longitudinal design developed (far beyond the scope of this report) before attempting the perilous task of estimating total funded positions. Generally, however, there is some indication that as much as 13 percent of budgeted monies may be paid to "relief" nursing personnel. It is reasonable to expect, as the data in Table 15 confirm, that during times of cost containment "casual" hours are the ones most likely to be curtailed and the difference in decreased total hours may be picked up by increased productivity. Yet it is not reasonable to expect this type of scenario to occur in small facilities where the absolute number of nursing staff makes it physically impossible to increase workload and to expect one nurse to be on duty in more than one ward, or, in service areas such as ICU/CCU, surgical wards, etc., where the weight of the workload makes it impossible to decrease numbers of personnel and expect two nurses to do the work of three for an indeterminate length of time. Our preliminary analysis of the number of warm bodies (as opposed to FTEs) provided a broader perspective on available nurse stock. While the number of employed RNs (not including graduate nurses) reconciled reasonably well with data from the RNABC membership files (68), the number of employed LPNs did not. Further examination of the data and the consideration of a possible explanation suggest that more than a third of those employed as practical nurses or orderlies are not members of the B.C. Council of Practical Nurses. Since there is no mandatory registration legislation for nurses in B.C. and the Master Agreement between the Hospital Employees' Union (HEU) and the Health Labour Relations Association contains no remunerary advantage for those practical nurses licensed by the Council, the discrepancy in numbers is, albeit misleading, understandable. Caution should, therefore, be exercised when using only the B.C. Council of Practical Nurses' Membership records to count current PN stock. As the employment data in Table ~ l8 demonstrate, available figures from that source should be inflated by at least one-third. These data also indicate that while the rate of change of total paid hours for PNs declined during the four-year period (Table 16), the actual number of persons gradually increased, - 42 -presumably due to the rates of increase in part-time work. It should be noted that this report always refers to the larger number of employed practical nurses as PNs, and the LPN designation is used only when reference is made to the smaller group who are licensed with the Council. This distinction was made only after the preliminary data analysis was completed and, combined with other "new" information from this stage, led to a reconsideration of the initial operationalized model (discussed in the next chapter). Budget and Average Length of Stay We also examined regional variations of hospital net income from patient services per patient day (the per patient capita BUDGET variable). Table 19 provides the intertemporal and cross sectional perspective on per patient-day costs for the years 1978-81 (since this is a one-year lagged variable). Generally, cross sectional variations are similar in magnitude to intertemporal ones. It is beyond the scope of this report to examine or attempt to explain such variations (71). The discussion on variable intercorrelations in the next chapter may shed some light on these results. Finally, a detailed examination of average length of stay, measured as: n ~ {ACUDAYS)i i=l AVLNST = n }.: (ACUSEPS)1 i=l where: AVLNST is the average number of days spent in short-term units; ACUDAYS is the number of days spent in short-term (Acute, Psychiatric and Newborn) units for those separated during that year; ACUSEPS is the total number of separations (discharges and deaths) from short-term units, indicates that although the variation between years is quite small, there is appreciable cross-sectional variation (Table 20). - 43 -Table 18: Number of Employed Practical Nurses in B.C. by Regional Hospital District, 1979 - 1982 REGIONAL HOSPITAL DISTRICT 197t 1980 1981 1982 Alberni-Clayoquot 82 81 96 94 Bul kl e,y-Nechako 71 69 86 81 Capital 7071 624 869 898 tar1boo 161 156 179 181 Central Coast 0 0 0 0 Central Fraser Valley 150 157 195 196 Centra 1 ICootenay 116 153 181 190 Central Okanagan 245 240 306 324 Columbia-Shuswap 57 66 87 90 Comox-Strathcona 149 160 99 99 Cowichan Valley 191 174 210 211 Dewdney-Alouette 134 163 191 202 East kootenay 112 125 140 137 Fraser-Cheam 89 111 131 141 Fraser-Fort George 123 123 115 102 Greater Vancouver 2,610 2,642 2,951 2,9742 Kitimat-Stikine 80 73 95 105 kootenay Boundary 116 112 153 165 Mount Waddington 0 0 0 7 Nanaimo 166 168 194 188 North Okanagan 140 132 162 157 Otanagan-Similkameen 203 190 229 250 Peace River-Liard9 60 79 93 114 Powel 1 River 48 45 55 59 Skeena-Queen Charlotte 71 17 76 65 Squamish-Lillooet 20 23 29 31 Stikine 0 0 0 0 Sunsh.ine Coast 24 31 38 41 Thompson-Ni co 1 a 187 221 286 305 Total 6,112 6,195 7,246 7,407 l Includes estimates frOID Hospital Annual Returns of 1979-80 for one Tertiary and two Extended Care Facilities which were not available from the BCHA data set for that year. 2 Includes estimates from one large facility based on the previous year's record. • Does not include a small Acute tare Facility. - 44 -. Table 19: Per Patient-Day Costs for B.C. Hospitals, by Regional Hospital District, 1978-1981 REGIONAL HOSPITAL DISTRICT Budget 78/ Budget H/ Budget 80/ Budget 81/ Pat-Day78 Pat-Day79 Pat-DaySO Pat-Day81 Alberni-Clayoquot 102.96 118.45 167.11 183.60 Bul kl ey-Nechako 123.56 144.50 174.62 202.23 Capital . ___ 1 ___ 1 156.05 155.55 Cari boo 115.20 126.89 154.04 176.81 Central Coast o.oo o.oo 0.00 o.oo Central Fraser Valley 96.49 108.09 129.49 156.77 Central Kootenay 137.58 162.70 153.48 159.77 Central Okanagan 98.88 109.68 134.16 150.1 g Columbia-Shuswap 107. 71 123.06 150.74 164.63 Comox-Strathcona 99.08 113.93 146.70 161.88 Cowichan Valley 91.60 99.18 123.43 140.06 Dewdney-Alouette 91.90 103.79 127.27 143.97 East Kootenay 115. 76 129.10 154.26 168.31 Fraser-Cheam 104.27 116.55 127.83 150.91 . Fraser-Fort George 131 .52 150.07 191.67 219.07 Greater Vancouver 130.76 146.95 183.51 212.93 Kitimat-Stikine 133.27 156.79 188.12 210.14 Kootenay Boundary 132.07 141.30 180.11 178.36 Mount Waddington o.oo 175.562 243.512 319.432 Nanaimo 106.39 122.52 160.32 171.40 North Okanagan 100.37 113.27 137 .61 155.52 Okanagan-Similkameen 104.46 111.23 141.07 157. 75 Peace River-Liard 159.04 181 .10 151.64 188.67 Powel 1 Rfver 108.11 119.81 139.96 169. 91 Skeena-Queen Charlotte 108.87 121. 95 155.60 183 .11 Squamish-Lillooet 109.59 131.45 189.17 218.71 Stik1ne o.oo o.oo 0.00 o.oo Sunshine Coast 96.49 112.83 141.64 160.98 Thompson-Ni co 1 a 147.40 162.26 166.58 192.45 Provincial Mean 114.13 129.09 154.85 174.37 1 Missing data from several facilities not included in the BCHA payroll system for that year would render the calculation of a regional average somewhat misleading. 2 Omitted from the Provincial Mean calculations. The artificially high per patient-day cost· for this region is due to the low denominator for (a) a newly in-service small facility and (b) decreased bed capacity 1n another small facility. - 45 -Ttble 20: AYeraae Length of Stay by Reaional Hospital Distri~t. B.C., 1979 • 1982 REGIONAL HOSPITAL DISTRICT 1979 1980 1981 1982 Alberni-Clayoquot --' __ , _1 7.33 Bul kl ey-Nechako 5.34 5.16 4.76 4.93 Capital _ _, 9.11 8.95 8.88 Car1boo 5.38 5.64 5.17 5.02 Central Coast o.oo 0.00 o.oo o.oo Central Fraser Va 11 ey 6.92 7.24 6.91 7.18 Central ICootenay 6.85 6.46 6.57 6.64 Central Okanagan 8.30 8.32 7.85 7 .94 Columbia-Shuswap 6.01 5.92 5.94 5.90 Comox-Strathcona 7.02 6.96 6.50 6.27 Cowichan Valley 7.09 7.23 6.97 6.89 Dewdney-A1ouette 7.20 7.25 6.54 7.37 East ICootenay 6.94 6.75 6.28 6.15 Fraser-Chea111 7.64 7.17 6.95 7.47 Fraser-Fort George 6.05 5.88 6.43 6.41 Greater Vancouver 9.19 9.05 8.60 9.03 Kit1mat-Stikine 5.63 5.62 5.56 5.69 ICootenay Boundary 7.93 8.26 7.29 7.53 Mount Waddington 3.70 3.49 3.62 4.20 Nanaimo 8.08 7.83 6.90 7.41 North Okanagan 7.42 7.65 7.30 7 .16 Okanagan-S1m11kameen 7.89 7.66 7.59 8.04 Peace River-Liard 5.83 5.84 6.16 5.71 Powell River 7 .81 8.07 a.so 7.52 Skeena-Queen Charlotte 6.27 6.66 5.96 5_4g Squamish-Lillooet 5.79 6.02 4.74 4.94 Stikine o.oo o.oo 0.00 o.oo Sunshine Coast 7.59 6.85 6.82 6.61 Thompson-Nicola 5.70 8.55 8.47 7_g7 All of B.C. 8.24 7.95 8.02 7.75 1· Missing infonnation from both facilities in this Regional Hospital District. 2 fffss1ng data from several f1cilities not included in the BCHA payroll syst1111 for that year would render the calculation of a regional average somewhat •fsleading. - 46 -7. Requirements for RNs in Institutional Care The initial testing of our operationalized model for nurse requirements (discussed in Chapter 5) yielded some unexpected, as well as expected, results. The most important among these was the somewhat linear relationship of RN and PN regular paid hours (Figures 3-6). Since our endogenous (dependent) variable RELPDHRS (total relative paid hours) per capita consisted of: "RN L (RNPDHRS)i RELPDHRS = i =l "RN "PN I ( RNPDHRS) i + l; ( PNPDRHS) i i=l i=l scarcity of cross-sectional variation in relative regular paid hours for all four years necessitated a revised model. The data indicate that there was not much regional variation in nurse staffing patterns during the study period (1979-1982). The small amount of variation was due basically to the differential deployment patterns of casual and part time RNs. Although the 3:1 (approximately) provincial ratio presented in Table 16 was generally applicable across the twenty-six Regional Hospital Districts for which complete data were available, the regions differed in their deployment patterns of part-time and casual RNs. Since regular hours comprised the major portion of total paid hours, the regional variations in the other two categories did not affect the latter. It was, therefore, futile to attempt a regression analysis of relative paid hours if there was no story to tell. Clearly, nurse manpower substitution did not appear to be taking place in the institutional sector, at least not on a scale large enough to be statistically discernible from the data. Interestingly, our findings indicate that while actual relative wages increased substantially in 1981, so did relative paid hours. The latter increased again in 1982 when relative wages remained comparable to those of 1981. The dependent variable was then modified to an absolute rather than relative measure of RN paid hours and a similar separate model developed for estimating PN paid hours. We then proceeded to verify a number of methodological limitations imposed upon the analysis by the current state of the art. Since our initial interest was the explanatory power of the postulated explanatory variables taken together, rather than the relationship between the dependent and each of the independent variables taken separately, the Figure 3: Ill "' :3 ::@ 'O -~ "' .. ... :3 en QI a: z: a: .. ~ -a.. lJ "' GI a.. Per capita Registered fllrses' Regular Paid Hours by Per Capita Practical Nurses' Regular Paid Hours, for Regional Hospital D1str1cts, 1979 .•----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----•. 1000 i i 120 + + I · I 140 + + I · . I 780 + + I . I 810 + • I . · I eoo • • • 120 I . . . I • • + I . : . . . I • • • 440 + • • I · I 360 + • + I · . I 280 • + I . I 200 + + .•----·----·----·----·----·----·----·----·----·----·----·----·----·--·----·----·----·----·----·----•. 100 140 tao 220 2&0 300 3•o 310 •20 •&0 !500 Per Capita PN Regular Paid Hours ~ '-I Figure 4: Per Capita Registered Nurses' Regular Paid Hours by Per Capfta Practical Nurses' Regular Pafd Hours, for Regional Hospital Dfstrtcts, 1980 Ill I. ::I ~ "a -~ I. .. .... ::I i a: ;!!: a: .. ~ -D. ~ ~ a. ,000 920 140 7'10 810 llOO !120 440 380 210 200 .•----·----·----·----·----·----·----·----+----·----·----·----·----·----·----·----·----·----·----·----+. • • I · I • • I I • • •• I I • • I . I • • • I . . I • • I . . . · · I • • • • I · . ·· · · . I i . . i • • I · . I • • • I I • • .+----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----•. too t30 tao tso 220 2!IO 210 3to 340 370 400 Per Capita PN Regular Pafd Hours .i:=. CX> Ftgure 5: ttoo 10;0 820 Ill ... ::I ~ 130 ,, -• a. ... 740 • ...::I en ., 1111: z HO 1111: 3 -a. 1110 l1 t a. 470 3IO 280 200 Per toP1ta Req1stere!f ft!rses' Regular Paid Hours by Per Capita Practical Nurses' Regular Paid Uoyrs for Regional Hospital Districts, 1981 .+----·----·----·----·----·----+----+----·----+----·----+----·----·----+----·----·----·----·----·----+. + + I . I + + I • I + + I : I : • + + I . I + + l . I . + + I • . I .. • + • • • + I .. I • . • . • • + I • . I • • + I . I + • + I . I • + .+----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----+. o Bo too 190 200 2so 300 3!IO •oo 490 soo Per Capita Pl Regular Paid Hours ~ ID Ftgure &: "' ... ::3 ~ ,, -~ ... .. -::3 en .. ai: z ai: .. .. -Q. lJ ... .. a. tooo 920 840 180 880 IOO •20 440 360 280 200 Per Captt1 RerJstered Nurses' Regular Pafd Hours by Per Capita Practfcal Nurses' Regular Pafd Hours for Regtonal spttal Dtstrtcts, l98Z .•----·----·----·----·----·----·----+----·----·----·----·----·----·----·----·----·----·----·----·----+. + + I . . I + • I . I + • I · I • • I I + • • I . · . I + + I . = · · I + • + I · . . · 2 • I + + I · · . I + • I . · I + • I . I + • . •----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----•. o !IO too t!IO 200 2so 300 350 400 4!IO soo Per Capf ta PN Regular Paid Hours U'1 0 Figure 7: General Practitioners per 10,000 Population by Specialists per 10,000 Population for Regional Hospital Districts, 1979 c 0 -... .. -ii e. 8 0 .. 0 ~ a. "' ~ c 0 -... -... u .. .t -~ c ., C!ll 19 14 12 tt • • 7 9 4 2 .•----+----·----·----·----+----·----·----+----+----·----·----·----·----·----·----·----·----·----·----+. + + I I + • I . · I + • + I . I + + I . : I + • • I . 2 • • • • I 2 •• i . . .. . i i i + + I I + + I I • + I I • + .•----·----·----·----·----·----·----·----·----·----·--·----·----·----·----·----·----·----·----·----+. I 3 9 7 9 I I 14 16 18 :ZO 22 Specialists per 10,000 Population U1 __. Figure 8: General Practitioners per 10,000 Population by Specialists per 10,000 Population for Regional Hospital Districts, 1981 .•----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----•. ts + + t4 I I + + t2 I . I + • + I: 0 -... .. ... " & 0 A. 0 8 • . 0 .... ... • °" • I . · I + • + I· . . · I +• • • • • • 1 · . · ·· · I • • • + tll t I: 0 - 7 ... 1 · = . . I + + -... u .. it: 15 .... • t I: ., C!J 4 I I i i + + 2 I I i i + + .•----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----·----•. t 3 9 7 9 t t 14 16 ti 20 22 Specialists per 10,000 Population U'1 N - 53 -multiple correlation coefficient (R) was the desired statistic. The stepwise multiple regression procedure involves the introduction of a first variable (the one with the highest F statistic) from the model to do all the explaining it can. A second independent variable is then permitted to go to work on that portion of the variation left unexplained by the first (controlling for the first). This is repeated as many times as there are variables in the model, controlling eaeh time for the other variables already introduced into the equation. It is, therefore, preferable to include independent variables which are relatively unrelated to each other (moderate zero-order correlations) in order for each to explain a different proportion of the total variation, but which have, also, moderately high correlations with the dependent variable. To the degree that independent variables are highly intercorrelated, both partial correlations and slope estimates will be increasingly sensitive to measurement errors (multicollinearity). The second important limitation in regression analysis is its assumption regarding the magnitude of variability about the regression equation (homoscedasticity). The amount of spread about the least squares line is assumed to be equal. The nature of the subject under study and the inherent data availability problems, thus, required careful transformation of data and several redefinitions of constructs to overcome the major technical limitations. After detailed examination of the zero-order correlation matrix for each of the years under study, several variable measures were modified. In addition to the modified dependent variable discussed earlier, the independent variable ACUBEDS -- a relative measure of the proportion of acute beds to total beds in each region -- was substituted by TBEDS. As a ratio measure, the former was too sensitive to the large regional variations in Extended bed capacity and did not accurately reflect the availability of complementary capital input. Total bed capacity per 10,000 population yielded better results both in terms of assumptions of homoscedasticity and problems of multicollinearity. Similarly, the variable DOCSUPPLY measured as the General Practitioner/Specialist ratio was taken initially as an estimate of complementary manpower availability. After closer examination of the ensuing results, however, this measure was modified to the total number of physicians per 10,000 population in the region (DOCS). Figures 7 and 8 provide a scattergram picture of GP/Specialist distributions for 1979 and 1981. The data clearly depict the stability of GPs per 10,000 population distributions and the contribution made by Specialist (per 10,000 population) distributions to the variation in total number of physicians per 10,000 population. For both bed capacity and physician supply variables the possibility of using two measures, that is, Acute bed capacity and - 54 -Extended bed capacity, General Practitioner stock and Specialist stock, were duly examined. The addition of these and other possible multi-measure variables would put considerable strain on one of the major methodological constraints imposed by the mathematical properties of regression analysis -- the maximum acceptable number of independent variables given the relatively small number of "observations" (RHDs). This same limitation was extended to our population variable, POPAGE. Ideally, this factor should enter the regression as twenty-two variables (in the usually presented age-sex groupings) which can possibly have differing explanatory values. Clearly, our data could not tolerate taking such liberties. The health services utilization literature suggests that females over the age of 70 are the highest users of health services and recent data from B.C. substantiate that premise (72). In order to further validate our initial subjective choice of the age-sex category that is most likely to influence the demand for nurses, we performed an age-sex factor analysis of the 1981 Population Census data. The purpose of this factor analysis was to explore the possibility of data reduction as well as the search for simple and interpretable factors. Three basic steps are customarily used in this type of analysis. The first involves the calculation of appropriate measures of association for a set of predefined relevant variables, in this case, the twenty-two age-sex sub-groups. The second step is to explore the data-reduction possibilities by constructing a set of new variables based on the interrelations indicated by the data. Briefly, this step entails determining the best linear combination of variables that would account for more of the variance in the data than any other linear combination of variables. The first principal component is, therefore, considered the single best summary of the linear relationships exhibited by the data. Initial factors are either extracted in orthogonal fashion, that is, one factor is independent of the other, or by the oblique rotational method when the factors may be correlated. The final step in factor analysis involves the rotation of factors to terminal factors, since the exact configuration of the factor structure is not unique and the implication is that there are many statistically equivalent ways to define the underlying dimensions of a data set. Since our primary interest in using factor analysis was mainly to provide us with some insight into the possible summary measures of the POPAGE variable as well as to further support our choice of a single population measure (females aged more than 70), we shall not delve deeply into the intricacies of factor-analytic techniques. As anticipated, the twenty-two age groupings were highly intercorrelated as indicated by the correlation matrix (Appendix A). The factor-pattern matrix tells us the composition of a variable in terms of hypothetical factors. Our data showed the first factor with generally high loadings for most age-groups (Appendix B). However, the three oldest (55-64, 65-69, 70+) and the two youngest groups (0-4, 5-9) for both sexes loaded considerably higher than the rest. The second factor had high loadings for the 15-19 age-group, male and female. On the basis of this matrix, - 55 -the researcher would usually decide how many factors to retain (ordered in descending importance) and assess how thorough the analysis is. The importance of a given factor is indicated in terms of the amount of total variance in the data it accounts for (eigenvalue) and the proportion of total variance accounted for by that variable. The analysis indicated that the two factors would explain 88.2 percent of the variance (see Appendix B). Therefore, ideally, we should include 1n our regression analysis the two factor scores su1T1Tiarizing the eight older subgroups as the first factor and the two younger subgroups as the second factor. We have already mentioned the methodological constraints brought upon by the utilization of a relatively small number of observations (RHDs). For this report we chose to use the result of the factor analysis as a confirmatory tool rather than as a measuring device. The single age-sex group included in the regression analysis is among those with high loadings in the first factor. Our final explanatory note pertains to the BUDGET variable, hospital net income from patient services. To minimize the variance between RHDs (heteroscedasticity), data transformation was undertaken. The usual per capita budget measure, however, correlated so highly with the dependent variable, RN paid hours per capita, that it was predestined for exclusion. Conversely, changing the denominator of the dependent variable to one that was based on service utilization, RN hours per patient day, did not prove any more profitable. In that case, there was no correlation, and there should not be, between that dependent variable and most of the independent variables as currently defined. The BUDGET variable used, then, was net income from patient services per patient day (73). Alternatively, this could be expressed as: budget patient day = budget x population population patient day We hypothesized that since nurse paid hours per capita is positively correlated with patient days per capita, its inverse, population per patient day, would have a negative correlation counteracting the strong positive one of the first part - budget/capita. The conclusive regression equation was: where: RNHRS = a0 + a1TBEDS + a2DOCS + a3PNSTK + a4BUGyr~l + a5RELSAL + a6POP + a7~VLNST RNHRS is the number of RN paid hours per capita in a year; - 56 -TBEDS is the total number of beds per 10,000 population; DOCS is the total number of physicians per 10,000 population; PNSTK is the total number of employed Practical Nurses per 10,000 population; BUGyr-1 is the net income from patient services (budget) per patient day in the previous year; RELSAL is the average hourly RN wages relative to average hourly PN wages; POP is the proportion of total population who are female 70 and older; ~VLNST is the average length of patients' stay (for those separated that year) in short-term units. Tables 21, 22, 23 and 24 provide a detailed analysis of the variable interrelationships for each of the years under study. The correlation matrix does not indicate any clear or pronounced intertemporal trends. The dependent variable in each model, RN hours and PN hours, have moderately high correlations with all but relative wages and budget per patient days. A modest positive correlation between relative salary and RN hours in 1980 is no surprise since relative wages that year registered an appreciable decrease. The corresponding positive correlation between relative salaries and PN hours is, however, rather unexpected. The almost nil correlation (inclined towards the negative) between paid hours and budget per patient day reaffirms the underlying dimension of that variable discussed earlier. The correlation between the dependent variable and total beds per 10,000 population appears to have increased between 1979-1981 and somewhat decreased in 1982. The dependent variable also shows increased intertemporal correlation with the population variable, females 70 and older. The correlation with physician supply is highest for 1980 (.84) and lowest for 1979 (.78). The remaining two years (1981 and 1982) are almost identical (at .80). Among the independent variables included in the requirements equation, correlations are generally moderate and in the expected direction. It is interesting to note that the correlation between total beds per 10,000 population and total physicians per 10,000 population is no higher than .6 for all four years. For the 1979 regression, the relative salary variable proved to be of no explanatory value and was automatically omitted from the computerized calculations. Physician stock per 10,000 population as expected, accounted for most of the variation (60.4%) for that year (as for 1980 and 1982). Total beds per 10,000 population ranked second in - 57 -Table 21: Matrix of Correlation Coefficients of Variables in Requirements Model, B.C., 1979 Data RN HRS PNHRS T8EDS DOCS PNSTK RNSTK BUG78 ·POP AVLNST REL SAL RN HRS 1 .00000 PNHRS D.75078 1.00000 T8EDS 0.57348 0.61431 1.00000 DOCS 0.77734 0.48241 0.58410 1.00000 PNSTK 0.51439 0.87903 0.67139 0.19642 1.00000 RNSTK 0.93937 0.75264 0.66334 0.63194 0.64111 1.00000 8UG78 -0.07523 -0.23308 0.05474 -0.08616 -0.27753 -0.14824 1.00000 POP 0.52761 0.52961 0.05867 0.72097 0.31400 0.46535 -0.41990 1 .00000 AVLNST 0.65191 0.56016 0.22072 0.77620 0.36995 0.58652 -0.39670 0.85773 1.00000 REL SAL 0.11833 0.22283 0.00629 0.10241 0.16277 0.05579 -0.18214 -0.02856 0.13458 1.00000 Table 22: Matrix of Correlation Coefficients of Variables in Regu1rements Model, B.C., 1980 Data RN HRS PNHRS TB EDS DOCS RNSTK PNSTK BUG79 POP AVLNST REL SAL RN HRS 1.00000 PNHRS 0.69709 1.00000 T8EDS 0.74246 0.71049 1.00000 DOCS 0.83917 0.49624 0.60430 1.00000 RNSTK 0.93969 0.65779 0.79585 0.75633 1.00000 PNSTK 0.51836 0.90865 0.64772 0.28224 0.58701 1.00000 8UG79 -0.24646 -0.31266 0.05621 -0.19911 -0.17491 -0.34485 1.00000 POP 0.64228 0.56788 0.48132 0.72781 0.59956 0.43290 -0.49514 1.00000 AVLNST 0.52398 0.40780 0.36896 0.66131 0.48776 0.29137 -0.22356 0-.59474 1.00000 RELSAL 0.40835 0.53072 0.40156 0.16826 0.37313 0.53719 -0.13803 0.18187 0.06370 1.00000 - 58 -Table 23: Matrix of Correlation Coefficients of Variables fn Regufrements Model, B.C., 1981 Data RM HRS PNHRS TB EDS DOCS · RNSTK PNSTK . BUG80 Pee AVLNST RELSAL RN HRS 1.00000 PNHRS 0.64397 1.00000 TB EDS 0.84415 0.73739 1.00000 DOCS 0.79712 0.41665 0.63858 1.00000 RNSTK 0.95530 0.63039 0.83551 0.71540 1.00000 PNSTK 0.58549 0.94769 0.67409 0.32472 0.61584 1.00000 BUG80 0.04524 -l>.14405 0.04437 0.03874 0.02616 -0.25552 1.00000 POP 0.64048 0.55485 0.56691 0.69985 0.58409 0.54193 -0.44571 1.00000 AVLNST 0.77453 ·0.43081 0.51166 0.76141 0.73625 0.41524 -0.22549 0.70546 1.00000 RELSAL 0.18584 0.27660 0.16862 0.08270 0.09911 0.32750 -0.11712 0.12738 0.26895 1.00000 Table 24: Matrix of Correlation Coefficients of Variables in Requirements Model, 8.C., 1982 Date RN HRS PNHRS TB EDS DOCS RNSTK PNSTK BUG81 POP AVLNST REL SAL RN HRS 1.00000 PNHRS 0.69302 1.00000 TBEDS 0.79334 0.75729 1.00000 DOCS o. 79987 0.48247 0.61559 1.00000 RNSTK 0.97244 0.68178 0.82885 0.79442 1.00000 PNSTK 0.58007 0.93878 0.60995 0.34390 0.58133 1.00000 BUG81 -0.17785 -0.35047 -0.08028 -0.06528 -0.13796 -0.46101 1.00000 POP 0.65079 0.59885 0.46388 0.70914 0.61268 0.57742 -0.61226 1.00000 AVLNST o. 76191 0.50950 0.48031 0.77907 0.71870 0.46403 -0.34184 0.81033 1.00000 REL SAL 0.15703 0.26149 0.26931 0.12709 0.10982 0.27243 -~.36161 0.27802 0.27105 1.00000 - 59 -explanatory value for 1979, accounting for 18.4 percent of the variance in the dependent variable. The next variable, PN stock per 10,000 population came a distant third explaining only 1.3 percent of the variation. The contributions of POP and ~VLNST were negligible. Table 25 provides a summary of selected statistics for the four-year period. The multiple R coefficient indicatas the strength of the dependence between the endogenous variable and the exogenous ones. The R2 statistic reflects the overall accuracy of the prediction equation by indicating the proportion of variation explained by the variables included in the regression equation. The 1980 data show that DOCS and TBEDS jointly accounted for an appreciable proportion of the variation in RN paid hours. In comparison to their relative contributions in 1979, however, the TBED variable proved less enlightening in 1980 and DOCS slightly more so. RELSAL added a modest contribution (almost 3.0% of the variation). The year 1981 appears to have been atypical for DOCS. TBEDS topped the list, accounting for more than 70 percent of the variation, followed by ~VLNST explaining 16 percent of the total variation. BUDGET ranked third with less than two percent. Other contributions w~e negligible. It should be noted that 1981 data elicited the highest R (90.0%) in the four years studied. The regression equation for 1982 yielded the usual DOCS, TBEDS rank-order, with respective contributions comparable to those of 1979. AVLNST accounted for approximately five percent of variation in paid hours for RNs and RELSAL accounted for one percent. Contributions from other variables were negligible. As is customary with this type of analysis, we simplified our regression equation for each year to include only those independent variables which contribute significantly to the explanation of the variation in the dependent variable. Therefore, for 1979 the equation was: RNHRS = a0 + a1TBEDS + a2DOCS + a3PNSTK; for 1980: RNHRS = a0 + a1TBEDS + a2DOCS + a3RELSAL; for 1981: RNHRS = a0 + a1TBEDS + a~VLNST + a3BUG80; and for 1982: RNHRS = a0 + a1TBEDS + a2DOCS + a~VLNST Table 25: Variable - - -DOCS TBEDS PNSTK POP AVLNST DOCS TB EDS REL SAL 8UG79 POP PNSTK AVLNST T8EDS AVLNST BUGBO DOCS PNSTK RELSAL POP DOCS TBEOS AVLNST REL SAL PNSTK POP 8UG81 - 60 -Selected Statistics fr0111 the Multi~le Regression oo Registered Nurse Regu1rements, B.C., 1979-198 Multiple R R Square RSQ Change U79 0.77734 0.60425 0.60425 0.88810 0.78872 0.18447 0.89545 0.80182 0.01310 0.89694 0.80450 0.00267 0.89730 0.80515 O.DOD65 1980 0.83917 0.70421 0.70421 0.88964 0.79146 0.08725 0.90596 0.82077 0.02931 0.91355 0.83457 0.01380 0.91572 0.83855 0.00398 0.91603 0.83911 0.00056 0.91645 0.83988 0.00077 l 981 0.84415 0.71259 0.71259 0.93360 0.87160 0.15902 D.94241 0.88814 D.01654 0.94454 0.89216 0.00402 0.94570 0.89435 0.00219 0.94623 0.89534 0.00100 D.94649 D.89584 D.00049 ~ 1982 0.79987 0.63980 0.63980 0.88636 0.78563 0.14583 0.91343 0.83436 0.04873 0.91956 0.84559 0.01123 D.92169 0.84951 0.00392 0.92345 D.85275 D.00324 0.92556 0.85666 D.00391 - 61 -Multiple regression analysis provides statistics that go beyond the mere description of direction and strength of the independent variables/dependent variable relationship. The unstandardized regression coefficient B, and its standardized version Beta, provide the coefficient of predicted scores indicating the expected change in the dependent variable, with a change of one unit in each independent variable when all others in the equation are held constant. For example, a computed B statistic for TBEDS would indicate the expected change in RNHRS with a change of one unit of TBEDS when the effects of DOCS, PNSTK, BUGyr-1, RELSAL, POP, and AJLNST are taken out. In short, it is the pure and direct impact of DOCS on RNHRS, with the combined "effects" of each of the variables being additive. Such effects are described in Table 26. Using the 1982 results, for example, RNHRS = -296.488 + ll.865{DOCS) + 6.420(TBEDS) + 65. 728{_j\VLNST) With this prediction equation, the planner could compute a forecasted RNHRS figure for any given combination of independent variable values. The statistics in Table 26 indicate that the direct effect of each independent variable varies with the combination of the other independent variables included in the equation. While the zero correlations {bivariate correlations presented in Tables 21-24) include the confounding effects of all other correlations between all factors (including those not being considered), and the partial regression coefficient, B, includes the effect of one independent variable when the impact of other variables in the equation are controlled for, the standardized regression coefficient - Beta - quantifies the relative effect of each of the variables. Thus, the Beta coefficient for DOCS, 0.68 in 1979, is appreciably reduced to 0.23 in 1982 indicating that the other two variables in the equation have increased in relative importance. However, another equation model introducing other variables not considered here may provide a different Beta score for DOCS. The standard error of estimate provides an evaluation of the accuracy of the predictions by examining the amount of absolute error in the prediction. These figures, presented in Table 26, indicate the "average" error in predicting RN paid hours per capita for each of the years respectively. For example, in 1982 the RN paid hours per capita of approximately 95 percent of the RHDs will fall within the range Y + 91.96. This standard error of estimate is calculated from several statistics that reflect the average size of the residuals (Table 27). - 62 -Table 26: More Selected Statistics from the Multiple Re~ress1on on Registered Nurse Requirements, 8.C., 1979-198 1979 B Beta Standard Error DOCS 28.000 0.681 TBEDS 4.670 0.336 PNSTK 3.608 0.155 (CONSTANT 1 0 ) -84.748 81 .203 1980 DOCS 29.512 0.637 TBEDS 4.265 0.282 REL SAL 335.043 0.188 (CONSTANT a0) -381 .501 85.928 1981 T8EDS 7.718 0.578 AVLNST 93.833 0.509 8UG80 1.377 0.134 (CONSTANT a0 ) -583.241 74.249 1982 DOCS 11.865 0.228 TB EDS 6.420 0.484 AVLNST 65.728 0.352 (CONSTANT 1 0 ) -296.488 91.964 -• .. --• c: ~ .. a: ... ID 'i .. u i ~ c: • 'i E .. ... 0 ... ... .. :! l .. .. u .,_ ..... _., ... .,_ a: Cl ... N :! ... t! • ill r :::I l!I a . ~i .. f • - 63 -~~~---~-~D•W••··~-~D•N•l~W~-~ I '-~-'-~~~~~~~~~~ .. ~ .. ~~....~--~~~~~-.. ~N~N~N~N~N~-N~N~N~N~N~ 11 t"' .. ii • f • ill . r :::I l!I ... c: -.. "' ii • - 64 -8. Requirements for PNs .!!l. Institutional Care The descriptive analysis of data {discussed in Chapter 6) revealed the virtual absence of variation in RN/PN staffing patterns for all four years studied. A sep~rate regression equation for PN requirement was then estimated based largely on the RN model. As such, it is more of an exploration and afterthought and not specifically designed for this category of nursing personnel. The regression equation for PNs was: where: PNHRS = a0 + a1TBEDS + a2DOCS + a3RNSTK + a4BUGyr-l + asRELSAL + a6POP + a7~VLNST PNHRS is the number of PN paid hours per capita in one year; TBEDS is the total number of beds per 10,000 population; DOCS is the total number of physicians per 10,000 population; RNSTK is the total number of employed RNs, per 10,000 population; BUGyr-l is the net income from patient services per patient day in t~e previous year; RELSAL is the average hourly RN wages relative to average hourly PN wages; POP is the proportion of females 70 and older in the total population; ~VLNST is the average length of patients' stay in short-term units. In general, the proportion of explained variation in PN paid hours per capita was lower than that for the RNs. The two possible explanations in descending order of likelihood are {i) that the model has not captured the important factors affecting PN requirements, or, {ii) that the PN situation is less amenable to prediction. Table 28 provides a su111Rary view of selected statistics for each of the years 1979-1982, using the minimum number of independent variables which account for the optimum amount of variation in the dependent variable. Total beds per 10,000 population is on top of the list accounting for approximately 50 percent of the variation (R2) for every year except 1979. For those same years {1980-1982) budget per patient day is second in rank, although its relative .contribution {Beta) is - 65 -Table 28: Selected Statfst1cs from the Multfple Regressfon on Practfcal Nurse Requirements. B.C. 0 1979-lYH~ Varfable Multfple R R Square RSQ Change Simple R 8 Beta 1979 RNSTK 0.75264 0.56647 0.56647 0.75264 2.492527 0.31868 POP o. 77944 0.60753 0.04107 0.52961 20.89807 0.36509 TBEDS 0.81814 0.66935 0.06182 0.61431 2.689075 0.38016 RELSAL 0.84512 0.71423 0.04487 0.22283 195.1755 0.21309 (CONSTANT) -282.6887 Standard 50.787 Error 1980 TB EDS 0.71049 0.50480 0.50480 0.71049 4.398447 0.63510 BUG79 0.79342 0.62952 0.12472 -D.31266 -1.293811 -0.31633 RE LS AL 0.82054 0.67328 0.04376 0.53072 189.3938 0.23202 (CONSTANT) - 26.35018 Standard 53.117 Error 1981 TBEDS 0.73739 0.54375 0.54375 0.73739 4.773232 0.72153 BUGBO o. 75832 0.57505 0.03131 -0.14405 -0.8116582 -0.16011 RELSAL 0.76992 0.59278 0.01773 D.27660 117.0862 0.13618 {CONSTANT) - 3.553530 Standard 70.130 Error 1982 TBEDS 0.75729 0.57348 0.57348 0.75729 4.685956 0.71932 BUG81 0.81114 0.65794 0.08446 -0.35047 -0.2587316 -0.05766 POP 0.81980 0.67207 0.01413 0.59885 26.27685 0.41120 DOCS 0.82935 0.68782 D.01575 0.48247 -6.551299 -0.25570 (CONSTANT) 41.09237 Standard 63.482 Error ' - 66 -appreciably diminished over time. Relative wages account for approximately four percent and two percent of the variation in PN paid hours in 1980 and 1981, respectively. POP and DOCS explain a modest amount of the variation in 1982. Total variation explained for PNs is highest in 1979, yet this appears to be an atypical year in terms of the combination of independent variables and their equal appearing Beta coefficients. As well, this is the year which had PNSTK enter the regression equation for estimating RNHRS. The converse of that relationship is indicated for estimating PNHRS. The strength of the Beta coefficient for TBEDS in the last three years of regression analysis for PNs is perhaps indicative of the nature of the latter's roles and functions and their task-oriented work style. - 67 -IV. DISCUSSION 9. Surmnary of Findings In a report of this kind which contains large quantities of information presented for the first time in a systematic fashion, it is difficult to establish superimposed boundaries of what are considered minor details and what are considered major findings. Nevertheless, we shall attempt to summarize some of the important information including that which may have been only intuitively suspected heretofore. Our extensive literature review unearthed a large variety of studies directly and indirectly addressing nurse manpower requirement issues. Among the various modelling perspectives, the most relevant to our concerns are those dealing with the nurse education market, the nurse manpower market and the nurse services market. Manpower resource models, health care delivery models, health consumer behaviour models, and incidence/prevalence of illness models contribute to understanding the processes affecting the three aforementioned markets. Our model was developed based on what had already been studied and utilized, taking into consideration more current concerns and data limitations. For the six-year period examined (1976-1982) starting salaries for new RNs have increased by 69.3 percent, compared to a 66.l percent increase for the equivalent Health Record Administrator category, 65.3 percent for Practical Nurses and 64.7 percent for Medical Social Workers. Although teachers and bank tellers have obtained much higher salary increases for the six-year period, salaries for these groups have been appreciably lower; these salary rates are proportionally better in 1982 than they were in 1976 or, conversely, RN relative wages are less favourable in 1982 than they were in 1976. The labour force activity rate (number employed in nursing expressed as a percentage of the number registered as practising members) was very stable at approximately 88 percent for 1975-1980 inclusive. However, in 1981 this activity rate increased to 93 percent, decreased to 92 percent in 1982 and escalated again to 94 percent in 1983. Actual relative wages indicated the average RN wage was 22 percent higher than the average PN wage in 1979; 20 percent higher in 1980; 28 percent higher in 1981; and 27 percent higher in 1982. While RN relative wages increased in 1981 and remained high in 1982, the ratio of registered/practical nurse paid hours steadily increased during the four-year study period from 2.35 to 2.97. - 68 -Regular paid hours for RNs comprised 84 percent of total paid hours in 1979, decreased to 82 percent of the total in 1980, and further to 80 percent in 1981 and 1982, respectively. Part-time and casual hours compensated for the drop in regular hours. The impact of the 1982 crash, however, was absorbed mainly by the casual RN hours. The PN data indicate that part-time work is preponderant for this group. Moreover, the 1mpact of the 1982 decrease in total PN paid hours is eleven times the decrease in total RN paid hours. As well, the data indicated that there were at least one-third more employed PNs than the number licensed with the 8.C. Council of Practical Nurses. The initial regression model, Relative paid hours as a function of (ratio of acute to total beds; ratio of GPs to specialists; number of LPNs per 10,000 population; per capita hospital net income from patient services; average hourly wage of RNs relative to PNs; proportion of population who are female and 70 years and older; and average length of stay in short-term units), was tested and modified. The testing of our operationalized model indicated that there was no variation in the dependent variable as currently defined. In addition, proportion of acute beds and proportion of General Practitioners were not accurate estimates of availability of complementary capital input and complementary manpower, respectively. Budget per capita was modified to budget per patient day. The conclusive RN regression equation was: RN paid hours as a function of (number of beds per 10,000 population; total number of physicians per 10,000 population; number of PNs per 10,000 population; budget per patient day; average hourly wage of RNs relative to PNs; proportion of population female 70+ years; average length of stay in short-term units). In general physician stock and bed stock accounted for, jointly, an appreciable proportion of the variation in RN paid hours. The variation explained with seven variables in the equation ranged between 85-90 percent for each of the four years examined. We further deleted the number of independent variables to include only those with substantial contributions to explain optimum variation in the dependent variable. For 1979, 1980 and 1982 a third independent variable (PN stock, relative salary, average length of stay, respectively) provided a moderate increase in predictive value. For 1981, an altogether different set of variables was evident: total bed stock, average length of stay and budget per patient day. - 69 -The regression equation for 1979 was: RNHRS = -84.748 + 4.670{TBEDS) + 28.000{DOCS) + 3.608{PNSTK); for 1980: RNHRS = -381.50 + 4.265{TBEDS) + 29.512{DOCS) + 335.043(RELSAL); for 1981: RNHRS = -583.240 + 7.718{TBEDS) + 93.833{~VLNST) + l.377{BUG80); and for 1982: RNHRS = -296.488 + 6.420{TBEDS) + ll.865{DOCS) + 65.728(~VLNST). The predictive value of the PN regression equation was lower than that of the RN, approximately 70 percent at best. - 70 -10. Concluding Remarks As previously stated, the primary purpose of this study was to place requirements issues in a broader context than the usual personnel:population ratio analyses, in order to improve our understanding of some of the complex underlying factors. The four specific questions raised in our chapter on the scope and objectives of the study will form the framework for this discussion: (i) What is the level of utilization (effective demand) of nursing .services? Does it vary appreciably over time? The data from 1979-82 provide clear indications of nurse personnel mix preferences of employers and deployment of regular/casual, full-time/part-time personnel within each of the two categories. For both RNs and PNs the proportion of "regular" hours steadily decreased over time, although absolute hours increased in the 1979-81 period and dropped in 1982 (with a louder thud for PNs than RNs). The crash of 1982 was felt mostly in the reduction of "casual" hours for RNs but 11 part-time11 hours were still on the increase. For PNs, all three "types" of hours were decreased. Thus, RN utilization patterns appeared fairly stable for regular and part-time nurses during the first three years under study, with appreciable variation reported for casual nurses. The largely unanticipated fluctuation of 1982 was absorbed by the decreasing requirements for casual RNs. The PN picture was much less stable pre-1982 as well as during that year and yet the PN services market appears to be parallel to that for RNs. ii) Is there any manpower substitution among nurse categories by hospitals? Does this vary over time or across geographic areas? There does not appear to be substitution of RNs by PNs in the four years since 1979. The RN/PN ratio steadily increased from 2.35 to 2.97. However, our regression analysis for 1979 indicates that the availability of substitute manpower in that year (our PNSTK variable) contributed to the explanation of the variation in RN paid hours, and the converse was true for PN paid hours, but not in the a priori intuitive direction. The direction of nurse manpower substitution appears to be, perhaps surprisingly, counterintuitive, that is, RNs substituting for PNs. The degree of substitution also varies counterintuitively. When relative wages (RN/PN salaries) declined in the 1979-80 period the RN/PN ratio increased by approximately five percent. When relative wages increased appreciably in 1980-81, the RN/PN ratio increased by seven percent followed by an increase of twelve percent in 1981-82 when relative salaries remained stable. An attempt to explain this type of trend can only be made after a close examination of the changing roles - 71 -and functions of each of the nursing categories, and the small 'p' political environment within which decisions are made, within a broader context of the changing nature and complexities of health delivery systems. iii) What are the major factors that affect nurse requirements? Starting from the premise that there is, and always will be, some positive correlation between available supply and effective demand and, that the highly regulated and largely non-profit nature of the sector determines the parameters of the capital input availability factor, we delved into an examination of other planning and policy variables that are more amenable to change and control. Seven such factors were closely examined for each of the four years under study and for both nurse categories, using multiple regression analysis techniques. Expectedly, physician stock per 10,000 population was the most important factor in terms of explanatory value for RN requirements and total bed capacity per 10,000 population was the most important factor for PNs. As well, this factor was second in rank for estimating RN requirements. Budget per patient day was ranked second for PN requirements. Other variables made minor contributions and were different one from the other for the different years. The predictive value of the model was generally high at 85-90 percent for RNs. These findings for both nurse categories, however, were not applicable to all four years. In each case one year differed from the three others, albeit not the same one. For RNs, 1981 appears to be the atypical year, and for PNs, 1979 is the atypical one. The findings from these years provide some insight into the complex network of interrelationships among the independent variables. Even among the three years which yielded common factors, the results indicate the variation in relative contributions made by the respective factors. iv) What are the interrelationships among the major quantifiable factors that affect nurse requirements? The study findings indicate the complexity of the interrelationships among even the minimum three independent variables which account for the optimum variation in the dependent variable. The "unusual" year for RN requirements is 1981 for which the important factors were total bed capacity per 10,000 population and average length of stay, and to a lesser extent budget per patient day. This is also the year in which total bed capacity increased by 582 beds. In general, these were Extended Care or Rehabilitation beds and as such had some impact on the average length of stay. The capital input availability variable, budget per patient day, is naturally of some importance in a year when salary costs for RNs relative to PNs escalated appreciably. - 72 -Finally, the direct influence of physician stock (mostly due to the variability in specialist stock) on RN requirements (for the years it does enter the regression equation) varies more than threefold between the two early years and the last year under study. Thus, simply knowing that in year Y, physician stock will be X, will not yield as accurate a prediction of RN requirements as knowing that physician stock will be X and total bed capacity will be Z. The interactive effect of bed capacity and physician supply on RN requirements is clear from the model. The estimation of provincal future requirements (1983, for the convenience of using actual data) using 1982 equations, entails the following steps: a) calculation (or verification) of physician supply per 10,000 population (69); b) calculation (or verification) of total beds per 10,000 population (69); c) calculation of provincial average length of stay; this latter has not changed much in the period under study. (from 6.99 to 6.83, see Table 20) 1983 estimated per capita RN hours = · = - 296.488 + 6.420 (67.82) + 11.865 (19.67) + 65.728 (7.75) = - 296.488 + 435.404 + 233.385 + 509.392 = 821.223 = 881.693 This estimate is a province-wide average requirement figure for 1983; the standard error of+ 91.964 from the 1982 data (Table 26) indicates that approximately 95 per cent of the Regional Hospital Districts would fall somewhere between 789.729 and 973.657 (881.693 ! 91.964). The impact of proposed policy changes manpower requirements, thus, can be accurately estimated once the interrelationships between these variables are understood and quantified. Application to Policy The more interesting feature of regression analysis, at least for the planner and policy maker, is its contribution to the development of alternative scenarios. Hypothetical values of the variables in the equation can be introduced to estimate hypothetical RN requirements. As well, the rate of change incurred in the dependent variable as a result of change in an independent variable can be measured and specified. For example, if the physician per 10,000 population ratio were to increase 2 - 73 -percent from the 19.67 figure for the base year 1983 (69}, the ensuing per capita RN paid hours would be: -296.488 + 6.420 (67.82} + 11.865 (20.06} + 65.728 (6.83} =825.850, this indicates a 4.63 hours per capita increase (0.56%}. A five per cent increase in the physician per 10,000 population ratio results in a one per cent increase in RN requirements (7.63 additional hours per capita}. The salary costs for such an increase in RN hours can easily be estimated using an average hourly wage rate. Similarly, the impact on requirements of a change in the bed capacity per 10,000 population variable can be estimated. For example, if a target of 60 total beds per 10,000 population were to be set (a decrease of 11.5 per cent from the 1983 rate}, the ensuing decrease in RN requirements would be approximately six percent (a reduction of 50.20 hours per capita}. The underlying implication of our model regarding RN manpower markets is that the most important factors are neither patient nor budget related. The former is, perhaps, not so surprising since RNs very rarely provide primary care and are not the gatekeepers to the delivery system. Their strong dependence on physicians for the provision of services is clearly quantified here with some qualification as to how strong that dependence has been, based on other intervening factors in the four years studied. The budget variable is conspicuous by its lack of measurable influence. One logical explanation may be that the model is not designed to capture the impact of a generally uniform cut across all regions. A pooled time-series analysis which measures budget fluctuations relative to the base year may yield different results. Another explanation for the absence of budget may be that the influence of the budget per patient day variable is manifested through the bed capacity per 10,000 population variable. The results from a regression analysis which included all four years and seven variables were similar to the single year analyses, (Appendix C} yielding the following equation per capita RN paid hours = - 74.761 + 4.784 (TBEDS} + 27.488 (DOCS} + 4.240 (PNSTK} = - 74.761 + 4.784 (67.82} + 27.488 (19.67} + 4.240 (26.45} = 902.527 The standard error for this province-wide average requirements figure is ± 88.926, indicating that approximately 95 per cent of the RHDs would fall somewhere between 813.601 and 991.453. - 74 -The implication here is that the underlying factors (which affect the data) from four years instead of one, have been taken into consideration and therefore, given that the same underlying factors are operant in the future, these results should provide a higher degree of predictive accuracy (74). The purpose of this report has been to present a cross sectional and across time analysis of RN requirements in the Institutional Sector. The descriptive analysis has provided a detailed examination of aggregate deployment patterns and other employer practices. The model, developed to forecast future RN requirements in the Institutional Sector, has provided the quantification of the important factors (as postulated by the model) and has delineated their relative contributions. This report, however, describes only the second phase of the three-phase study which will culminate in a report discussing net requirements for RNs in this province. That report will provide an--analysis of current and future supply vis a vis requirements, and forecast manpower imbalances or equilibrium as indicated by the data. - 75 -FOOTNOTES - REFERENCES 1. For a clear sumnary of fiscal responsibilities, see: Statistics Canada (1978), Canada Year Book 1978-79, Hull, Quebec: Supply and Services Canada. -- -2. For an historical perspective, see: Marsh, L. (1975), Report on Social Security for Canada, 1943, Toronto: University of Toronto Press. 3. Drunmond, M.F. (1980), Principles of Economic Appraisal in Health Care, Oxford: Oxford University Press. 4. Feldstein, P.J. (1979), Health Care Economics, New York: John Wiley and Sons, Inc. --5. Ward, R.A. (1975), The Economics of Health Resources, Massachusetts: Addison-Wesley Publication Co. ~ 6. Benham, L. (1971), "The Labor Market for Registered Nurses: A Three-Equation Model", The Review of Economics and Statistics 111(3), 246-252. 7. Hurd, R.W. (1973), "Equilibrium Vacancies in a Labour Market Dominated by Non-Profit Firms: The 'Shortage' of Nurses", The Review of Economics and Statistics, May, 234-240. -8. Keaveny, T.J. and R.L. Hayden (1978), "Manpower Planning for Nurse Personnel", American Journal of Public Health 68(7), 656-662. 9. Kazanjian, A. and G. Wong (1982), Registered Nurses in British Columbia: ~Report Q!l the Current Supply Situation, Report S:ll, Division of Health Services Research and Development, The University of British Columbia. 10. See, among others: Marmor, T.R. (1970), The Politics of Medicare, Chicago: Aldine. 11. Hall, T.L. (1978), "Demand", in T.L. Hall and A. Mejia (eds.), Health Manpower Planning, Geneva: World Health Organization. 12. Mejia, A. and T. Fulop (1978), "Health Manpower Planning: An Overview", in T.L. Hall and A. Mejia (eds.), Health Manpower Planning, Geneva: World Health Organization. 13. For an excellent discussion see, Thurow L.C. (1983), Dangerous Currents, New York: Random House. 14. Adapted and greatly simplified from the' elaborate health manpower - 76 -classification scheme in Vector Research, Inc., (1973), A Health Manpower Model Evaluation Study.:. Volume II: Health Manp'Ower Model Inventory, Department of Health, Education and Welfare, NTIS. 15. Fuchs, _V.R. and M.J. Kramer (1972), Determinants of Expenditures for Physicians' Services in the United States 1948=68, DHEW P'iiDlication No. (RSM) 73-101~ 16. Human Resources Research Centre (1972), Implementation of Micro-Simulation Model of Health Manpower Demand and Supply, Uiliversity of Southern California, Research Institute for Business and Economics, Quarterly Progress Report to Division of Health Manpower Intelligence on U.S. Public Health Service Contract No. NIH 71-4065. 17. Feldstein, P.J. and S. Kelman (1972), An Economic Model of the Medical Care Sector: ~ Description of the Model and I l luSfrillve Aaplication, University of Michigan: Bureau of Hospital A ministration, School of Public Health. 18. Feldstein, P.J. (1972), "Financing Dental Care: An Economic Analysis", unpublished paper, University of Michigan. 19. Feldstein, M.S. (1967), "An Aggregate Planning Model of the Health Care Sector", Medical Care _V(6) November-December, 369-381. 20. Feldstein, M.S. (1970), "The Rising Price of Physicians' Services", The Review of Economics and Statistics Lil (2). 21. Andersen, R. (1968), "A Behavioral Model of Health Service Use", in A Behavioral Model of Families' Use of Health Services, University of Chicago Research-Series No. 2-S:-Center of Health Administration Studies. 22. Navarro, V., R. Parker and K.L. White (1969), "A Stochastic and Determinfstic Model of Medical Care Utilization", Health Services Research 4(2) Summer. 23. Andersen, R.M. et al. (1983), "Exploring Dimensions of Access to Medical Care", Health Services Research 18(1) Spring, 49-74. 24. Berki, S.E. and B. Kobashigawa (1976), "Socioeconomic and Need Determinants of Ambulatory Care Use: Path Analysis of the 1970 Hea 1th Interview Survey Data", Medical Care 14(May), 405-421. 25. Chiu, G.Y. et al. (1981), "An Examination of the Association of 'Shortage' and 'Medical Access' Indicators", Health Policy Quarterly l(Sunmer), 142-58. 26. Rosenthal, G.D. (1964), The Demand for General Hospital Facilities, Hospital Monograph Series No. 14. ~ - 77 -27. Baer, R.R. (1971), Patient Characteristics, Hospital Services and Length of Stay0 A Pilot ~udy, Ph.D. Dissertation, The University of Minnesota, MI 07830 • 28. Research Triangle Institute (1972), Simulation of Hospital Ut111zat1on and Health Manpower Regufremenn,. _VOTumes I, It, Technical Report No. l, National Institutes of Health: Bureau of Health Manpower Education. 29. Fitzmaurice, J.M. (1972), The Demand for Hospital Services: An Econometric ¥tMdy of Mar01ancr Counties;l>h.D. Dissertation, The University o aryTand, MI 72-18, 947. 30. Hopkins, C.E. et al. (1967), Methods of Estimating Hos~ital Bed Needs, Los Angeles, California: The Uilfversity of Cali ornia-;IJSPHS Grant No. AP-04, 03c, UMI HE 1094. 31. Jeffers, J.R. et al. (1971), "On the Demand Versus Need for Medical Services and the Concept of Shortage", Amerfcan Journal of Public Health 6l(Jan.), 46-63. -32. Auster, R., I. Leveson and D. Sarachek (1969), "The Production of Health, An Exploratory Study", The Journal of Human Resources .I.V(4). - -33. Anderson, J.G. (1972), "Causal Model of a Health Services System", Health Services Research 7(1) Spring, 23-42. 34. 35. Larmore, M.L. (1967), An In0uiry into an Econometric Production Function for Health in~e nited--stites, Ph.D. Dissertation, Northwestern University, UMI 68-3200. Freeburg, L.G. et al. (1979), Health Status, Medical Care Utilization and Outcome: An Annotated Bibliography of Empirical Studies, Washington, D.C.: U.S. Government Printing Office, DHEW Publication No. (PHS) 80-3263. 36. Maurana, C. et al. (1981), The Use of Health Services: Indices and Correlates .::..B. Research BibTiOgraphy 1981, laYfayette, IN: Healtfl" Services Research and Training Program, Purdue University. 37. Connor, R.J. (1960), B_ Hospital Inpatient Classification Slstem, Ph.D. Dissertation, The John Hopkins University, OMI 60-33 9. 38. Singer, S. (1961), B_ Stochastic Model of Variation of Categories of Patients within.! Hospital, Ph.D. Engineering Dissertation, The Johns Hopkins University, UMI 61-3839. 39. Wolfe, H. (1964), A Multiple Assignment Model for Staffing Nursing Units, Baltimore: The Johns Hopkins Hospital, Operations Research Division. - 78 -40. Jelinek, R.C. (1964), Nursing: The Develo~ment of an Activitl Model, Ph.D. Dissertation, University ofichigan,IJMI 65-59 2. 41. Thomas, W.H. (1964), Model for Short-Term Prediction of Demand for Nurs1ng Resources, Ph.D. D1ssertatT'on, Puraue UniversTfy, OMI - 6~ 8720. 42a. Laberge-Nadeau, C. and M. Feuvrier (1972), "Nursing Staff Utilization", Canadian Hospital, May. 42b. Laberge-Nadeau, C. and M. Feuvrier (1972), "Utilization of a MEDSIM Program in a Task Study of Nursing Staff", Canadian Hospital, May. 43. Chagnon, M., L. Audette and C. Tilquin (1977), "Patient Classification by Care Required", Dlr.ENSIONS in Health Service, September, 32-36. ~ 44. Chagnon, M. et al. (1978), 11A Patient Classification System by Level of Nursing Care Requirements", Nursing Research 27(2), 107-113. 45. Chagnon, M. et al. (1978), "Validation of a Patient Classification Through Evaluation of the Nursing Staff Degree of Occupation", Medical Care ~VI(6), 465-475. 46. Robinson, G.C., C.P. Shah and C. Kinnis (1973), "Patient Care Classification of Children in Hospital: Predicted Changes in Patient Care Group Classification in a Future Hospital System", British Columbia Medical Journal 15(2) February, 31-33. 47. Shah, C.P., C.G. Robinson and C. Kinnis (1973), "Patient Care Classification of Children in Hospital - Study Points to Effective Utilization Methods in all Hospitals", Canadian Hospital 50(7) July, 38-44. 48. Baligh, H.H. and D.J. Laughhunn (1969), 11An Economic and Linear Model of the Hospital", Health Services Research 4(4) Winter, 293-303. 49. Shuman, L.J. and H. Wolfe (1972), "Mathematical Programming", in L.J. Shuman et al. (eds.), Operations Research in Health Care.:..~ Critical Analysis, Baltimore, Maryland: The Johns Hopkins University Press. 50. Hershey, J.C., W.J. Abernathy and N. Baloff (1972), Comparison of Nurse Allocation Policies -- A Monte Carlo Model, Graduate SchoOT of Business, Stanford University, August. - 79 -51. Smallwood, R.D., E.J. Sondik and F.L. Offensend (1971), "Toward an Integrated Methodology for the Analysis of Health Care Systems", Operations Research 19, 1300. 52. Maki, D. (1967), ~Forecasting Model of Manpower Requirements .in. the Health Qc_cupations, Ph.D. Dissertation., Iowa State University, URI 68-5964. 53. Deane, R.T. (1971), Simulating an Econometric Model of the Market for Nurses, Ph.D. Dissertation, -UCLA, UMI 72-13, 601:- -54. Hixson, J.S. (1969), The Demand and Supely of Professional Hospital Nurses: Intra-Hospita'flfesource 10Tocat1on,l5h.D. Dissertation, Michigan State University, UMI 69-16, 146. 55. Johnson, W.L. (1980), "Supply and Demand for Registered Nurses --Some Observations on the Current Picture and Prospects to 1985. Part 2", Nursing and Health Care, September. 56. Levine, E. (1978), "Nursing Supply and Requirements: The Current Situation and Future Prospects", in Political, Social and Educational Forces on Nursing: Impact of Political Forces, New York: National League for Nursing, Pub. No. 15-1754. 57. Engler, D. (1981), The Costs and Benefits of Baccalaureate Education for RNs, Thesis, Ohio State University (Ann Arbor, Michigan: University Microfilm International). 58. Yett, D.E. (1975), An Economic Analysis of the Nurse Shortage, Toronto: Lexington Books, D.C. Heath and Company. 59. For a discussion of the relative contributions of the different methodologies, see: Kriesgberg, H.M. et al. (1976), Methodological Approaches for Determining Health Manpower Supply and Requirements. Volume l.:. Analytical Perspective, Health Planning Methods and Technology Series, DHEW Publication No. (HRA) 76-14511. 60. Aday, L.A. and R. Andersen (1975), Development of Indices of Access to Medical Care, Ann Arbor: Health Administration Press. 61. Blalock, H.M. Jr. (1960), Social Statistics, New York: McGraw-Hill Book Company. 62. The University of British Columbia (1984), ROLLCALL 83. A Status Report of Health Personnel in the Province of Britisfi""""C"oTumbia, Division-of Health Serviceslfesearch and Development, Office of the Co-ordinator of Health Sciences, Report R:23. 63. U.S. Department of Health, Education and Welfare (1976), Health Manpower Planning Process, Health Planning Methods and Technology Series, DHEW Publication No. (HRA) 76-14013. - 80 -64. ROLLCALL 83, op. cit. 65. Statistics Canada, Canadian Statistical Review, Cat. ll-003E, Dec.'73 - Sep.' 83. 66. Barer, M.L., A.J. Stark and C. Kinnis, "Manpower Planning, Fiscal Restraint and the 'Demand' for Health Care Personnel", Inquiry, forthcoming. 67. Ibid. 68. ROLLCALL UPDATE 82 - A Status Re5ort of Selected Health Personnel in the Province or BrTtish Colum ia, DTvision of Health Services Research and Development, Office of the Coordinator of Health Sciences, Report R:20, and ROLLCALL 83, op. cit. 69. ROLLCALL 83, op. cit. 70. 1957.5 paid hours per year constitutes one FTE. 71. See, for example, Tatchell, M. (1983) "Measuring Hospital Output: A Review of the Service Mix and Case Mix Approaches", Social Science and Medicine, 17(13), 871-883. 72. Densen, P.M., S. Shapiro and M. Einhorn (1979), "Concerning High and Low Utilizers of Service in a Medical Care Plan, and the Persistence of Utilization Levels Over a Three Year Period", The Milbank Memorial Fund Quarterly 37(3), July. -For B.C. see Division of Health Services Research and Development reports F:3 (1979) and F:4 (1980), Inpatient Hospital Services per 1,000 Population and Average Length of Stay, 1976 and 1978, For Selected Service Categories by School District of Residence and Anticipated Numbers for 1981. 73. HSl data provide patient days for those separated in that year. 74. PN stock per 10,000 population was estimated using PN data from Table 18 of this report and population data from ROLLCALL UPDATE 82, op. cit. APPEUDICES Appendfa A: Correlation Matrfx for Input Variables fn POPAGE Factors MOI M02 M03 ll04 ..,. MO& MOT llOll llOll MOIO 9!01 1.00000 0.110211 0.151093 0.90321 0.71•"4 0.1•7H o.T02114 -0.113399 -0.9183T -0.1&113 MO:Z 0.110211 1.00000 o.•Hll3 0.111•01 0.1111111 o. 79T1'7 O.T20I• -o . 303117 -0.1738!1 -O.Tl84T lll03 0.11&013 0.411111:1 1.00000 0.1120ll o ...... o.23211 0.3'740I -0 .08111 -0.110 ... ·O .• Ol3• MCM 0 .1103:11 0.111401 0.1120ll 1.00000 o.ICM11 0.21313 0.21422 0.021121 -0.117118 -0 .• 1670 MOii 0.71•114 0.1118611 0.'86118 0.10<!11 1, DOOQO O.'l\D_I 0.-llOUll -o .a1ae• •O.T•l4a •o .••••• MO& o ... 7n 0.'78Tff 0 .232H 0.21313 0.112111 1 .00000 0 .13146 -0.11148 -0.14137 -0.12T62 MOT o. 702114 0.720 .. 0.37406 0.21422 0.907111 0 . 13841 1.00000 -o ' 111282 -0.80113 -o. 78110 MOI -0.113:1111 -0.30317 -0 .09871 0.021121 -o.a1211 -0 . 31148 -0.111212 1 .00000 0.43070 0.31260 MOI -0.81137 -0.173111 -0.110<!4 -0.111111 -o . 1•1141 -0.141:117 -0.10773 0.43070 1.00000 0.862811 MOIO -0.11173 -0.71847 -O.IOB34 -0 .11&10 -o.1•111 -0.12112 -0.11110 0 .31210 0.86215 1.00000 M011 -0.11843 -o.1•210 -0.60<!02 -0.111170 -0.10'9117 -0 .173&1 -0.13411 0.23114 0.141!17 0.1611!1 FOi 0.17211 0.11411 0.11418' 0.110:1114 0.811711 0.1110<!3 0.12471 -0 .463!14 -0.11179 -0.111211 F02 0 .811120 0.81309 0.717113 0.&6120 O.MTll 0.81731 O.&TOllll -0 .408TI -0.811030 -0.81909 F03 o. 7148• 0.81&•8 0.19811 0.&110151 0.1118'8 0.11111121 0 .1111211 -0 . 23062 -0.110 ... -0 .THBll FO<I 0 . 4Qa.O 0 . 311281 0 .85271 0.11832 0.112100. 0.0811311 0 . 111240 0.00614 -0.48883 -0.127311 FOii 0.11121 0.110714 0.3880ll 0.111837 0 . 82181 o.1ooe1 0 . 111111 -0.40313 -0.78141 -0.1 ... H FO& 0 . 7713T 0.113114 o. 14114 0 .08411 0.1111114 0 .131188 0 .171116 -0 .317110 -0. 122118 -0.71600 F07 -O.Oll103 0.08483 0.23174 0 . 0'98111 -0 .0241• 0 .01111 o ... n2 0.31317 -0.21637 ·0.26108 FOB -0.84•40 •0.84TH -o . •631111 -0. 37318 -0.81423 ·O.HHI -0 .77181 0.61490 0 921130 0 .841182 Foe -0 .9207• -0.18803 -0 .18183 -0.111111 -0.71483 -0.17400 ·0.820<IO 0.311438 0.81633 0.86660 F'010 -0.820B3 -0.89120 -0.62038 -0.62233 •O . 788111 -0.811193 -0 .11330 0 .21320 0 . 86198 0 .96089 F'011 -0 .11142 -0 .878118 ·O.H846 -0.62111 -0.102•0 -O.lllOT -0 .7827• 0 21(1!!18 0.1116211 Cl.RHlll!I MOU FOi F02 F03 Fa. FOii F06 F07 FOii F09 MOI -o .1111143 0.97281 0.111520 O.Tl .. 4 o.•~o 0.111211 0.77137 -0 .011103 -0.8 ... 40 -0.92076 M02 -0.14270 0.81•31 0.11309 0.16U9 0.311218 0.11071• 0.77311• 0 011<113 -0 .84T96 -o . 8881)3 M03 -0 .60'902 0.!1<1117 o. 787113 0 .11818 0 . 111271 0.388011 o. 1•11• 0. 2387' ·0.•6355 -0.59993 MO<I -0.!11670 0.110311• 0.661120 0.180!11 0 . 81632 0.111837 0.01469 0 .04659 -0 . 37311 ·0.!195P1 MOS -0.BO<IST 0.611761 0.1068 0 . 111148 0.1121100 0 .12911 0 .115114 -0 .02494 -0 .611423 -0.78493 11106 -0.87361 0.1110<!3 0.69738 0.1111128 0.0811311 O. TOOllT 0 .831188 0 .01971 -0.8836£1 ·0.17400 MOT -0.83<1811 o. 72471 0.&1055 0.1188211 o. 1112<10 0.111189 0 .171116 0 .41172 ·O. 77 IH ·0.12040 MOB 0.23114 -0.463114 -0.40871 -0.23062 0 .00614 -0.40313 -0 . 38750 0 . 31317 0.61480 0.35"38 11()9 0.8<11117 -0.81678 -0.88030 -0.810<!4 -0 . •8883 -o. 71141 -0 .12258 -0 .21637 0 .821130 0 .91633 M010 0 . 86115 -0.8!1211 -0 .11808 -o. 7611111 -0.112T35 -0 ... 41111 ·O.THOO -0.26109 0.841182 0.86660 lllO" 1.00000 -0.18053 -0.16173 -0 .71677 -0.48838 -0.11482 -0 .1121211 -0 . 224111 0 .8!1969 0.96645 F01 -0 . 1180!l3 1 .00000 0 .18621 0 .80806 0.31831 0 .63005 0 .7 ..... -0.0'9091 ·0.9UOO -0.8249(1 F02 -0.16173 0.111628 I .00000 0.11220 0.83061 o.&o3•3 0.611131 0 .05514 -0.89!!6' •O. IH92 F03 -0 . 71&11 0.80806 0 .11220 1 .00000 0 . 117182 0 .4422T o.14363 0.20!l16 -o. 71621 ·0.12!;48 M011 F01 F02 F03 Fa. FOi FO& F07 FOB FOi FO<I •0 ... 838 0.39831 0 .63068 0 .1171182 1 .00000 o . •61118 -0.011111 0 ' 16010 -0.30115 ·0 ... 700 F05 -o .77 .. 2 0.63005 0 .103"3 0 . 44227 o . •&1111 1.00000 0.621162 O.O!l466 -0.69058 -o . 11013 F06 -0.12125 0.78 .. 4 0.61531 0 .11"3&3 -0 .01111 0 .412862 I . 00000 0 .22970 -0.1"718 -o 82719 F07 -0.22481 -0.0'9091 0.05514 0.205141 o . '8010 O.O!l•&6 0 . 22170 , .00000 0.00129 -o . !7733 FOB 0.811868 -0.1 .. 00 -0 .15562 -o . 71621 -0 . 301711 -0 .18058 -0 .84718 0.00129 1.00000 0 .91246 FOB 0.864145 -0.12 .. 0 -0 .81882 -0 .82548 -0 .41700 -0 .77013 -0 . 112719 -o. 11133 1).81241! , .00000 F010 0.8U23 -0.83207 -0.89173 -0.82818 -0.112713 -0.76286 -0.80108 -o. 17685 0 . 89531 o . 91!817 F011 0.84UI -0.81628 -0.11088 -0 .13221 -0.132111 -0 .63614 -o . T3520 -o . 13772 0 . 15161 l) . 93nf'>' F010 FOii M01 -0 . 82083 -0.111•2 lllDD M02 -0.11120 -0.871!18 M03 -0 .62031 -0 . 1111•& ... ., .. .. ... , .. 11111 .,. -0 .62233 -0 .12111 - ~· 11111 11()5 -0.71855 -0.702•0 M06 -0.111183 -0.11807 .. . "~" .. MOT -0.11330 -o. 7127• .. . '~" .. MOii 0.28320 0.21881 MOB 0.861188 0.18625 - . ..... .. M010 O.HOB8 0.118111 - . ..... .. 11()11 0.88•23 0.84848 F01 -0.8320T -0.81628 - . .... .., F02 ·0.118173 -0.81088 . F03 -0.12111 -0.13226 - .... .. FO<I -0 .112713 -0.1132118 .. . .... .. FOii -O.T6216 -0.8368• F'O& -0.1080B -0 .731120 .. D . .... no F07 -o. 176115 -o. 13772 .. , .... "' FOi 0 .181131 0.811161 F09 0.81817 0.83082 FOIO I .00000 0.15872 FOii 0 .811172 1 .00000 Appendix B: Factor Matrf x for Population Variables FACTOR 1 FACTOR 2 FACTOR 3 FACTOR 4 MOt 0.84611 -0.11080 -0.20530 -0.16414 M02 O.BB337 -0.08682 0.03818 -0.28861 M03 0.0693 o.nno 0.01109 -0.12393 M04 0.61683 0.73105 -0.081561 0.02862 MO!S 0.781158 0.08609 -0.24148 0.47774 M06 0.86077 -0.45245 -0.023151 0.088157 M07 O.B0087 -0.30611 0.4B114 0.02321 MOB -0.37274 0.28038 0.4B242 0.08643 MOB -0.8B271 0.036150 -0.02811 -0.02!587 M010 -0.861!5!5 -0.02388 -0.07787 -0.20877 M011 -0.97381 0.01382 -0.12412 -0.08B87 F01 0.84414 -0.11600 -0.14240 -0.23012 F02 0.91576 0.17706 -0.08577 -0.20B19 F03 O.B28!54 0.23119 0.1078!5 -0.295B1 F04 0.!52646 O.B06!51 -0.04002 0.07148 FO!S 0.77199 -0.00518 -0.23576 0.56B58 F06 O.B0403 -0.1547B4 0.17357 0.05727 F07 0.14547 0.09743 0.76221 0.1620B FOB -0.82688 0.25800 0.16980 o. 1115B7 F09 -0.99237 0.034B3 -0.04390 -0.02126 F010 -0.99294 -0.00909 -0.06688 -0.02252 F011 -0 .85013 -0.06121 -0.08313 0.11BB2 FACTOR EIGENVALUE PCT DF VAR CUM PCT 1 15.13124 11.a 7!5.1 2 2.47555 12.4 IB.2 3 1.35413 ••• 95.0 4 1.00171 1.0 100.0 UIEllll 1111 11111 M Fel1u m - • 1-1 • '112 - • , .. ,. • .. - • 11-11 • "' - • II-IC • .. - • 11-M • .. 1117 • ..... • P07 1111 . ...... • POI 1119 • M-14 • POI 1110 • ... .. • '10 1111 • 71• • '11 Appendi• C: . Selected Statistics froaa the Pooled Tfme-Series Multiple Regression on Registered Nurse Requirements, B.C. ·- -- •• - - ---- cxwx ·---· Variable Multiple R R Square RSQ Change Simple R B Beta DOCS o. 79281 0.62854 0.62854 0.79281 27.486 o.s&& TB EDS 0.88492 0.78307 0.15453 0.75042 4.784 0.341 PNSTK 0.89561 0.80212 0.01905 0.54609 4.239 0.118 (constant) -74.761 Standard 88.926 Error Nlnt BSRD SERIBS ROLLCALL REPORTS (continued) R:12 ROLLCALL 79. A Status Report of Health Personnel in the Province of British Columbia. March, 1980 R:13 Registered Psych1atric Nurses in British Columbia, 1977. A Descriptive Report. March, 1980 R:14 Physiotherapy Manpower in British Columbia, 1979. A Survey of Registrants of the Associa-tion of Physiotherapists and Massage Practitioners of British Columbia 1979 with 1977 Comparisons. August, 1980. (W.G. Hanning) R:15 Occupational Therapy Hanpower in British Columbia, 1979. A Survey of Members of the B.C. Society of Occupational Therapists and of all Other Known Occupational Therapists in British Columbia. August, 1980. (W.G. Hanning) R:16 ROLLCALL UPDATB BO. A Status Report of Selected Health Personnel in the Province of British Columbia. December, 1980 R:17 Place of Graduation for Selected Health Occupations, l9BO. July, 1981. (B. Mccashin, A. Kazanjian, M.L. Barer) R:l8 ROLL.CALL Bl. A Status Report of Selected Health Personnel in the Province of British Columbia. March, 1982 R:19 Place of Graduation for Selected Health Occupations - l9Bl. July, 1982. (S. Chan, A. Kazanjian, M.L. Barer) R:20 ROLLCALL UPDATE B2. A Status Report of Selected Health Personnel in the Province of British Columbia. March, 1983 R:21 Licensed Practical Nurses in British Columbia 1975 - l9B2 - A Descriptive Rep~rt. March, 1983 (S. Jansen, S. Chan) R:22 Place of Graduation for Selected Health Occupations - 1982. August, 1983. (S. Chan, C. Jackson) R:23 ROLLCALL B3. A Status Report of Health Personnel in the Province of British Columbia. March, 1984 R:24 Plac6 of Graduation for Selected Health Occupations - 1983. June, 1984. (S. Chan, c. Jackson) FACILI'l'Y REPORTS: F:3 Actual Number, l97B and Projected Number, l98l, G.V.R.D. and Non G.V.R.D. Hospital Cases and Days Stay by Selected Service categories. September, 1979 F:4 Inpatient Hospital Services Per l,000 Population and Average Length of Stay, 1976 and 1978, SPECIAL REPORTS: for Selected Service categories by School District of Residence and Anticipated Numbers for l98l. April, 1980 · S:l Health Care or Health - The Development of a Plan to Address the Health Care of the Elderly in British Columbia. May, 1977. (M. McPhee) 5:2 Accreditation, Certification and Licensure of Health llanpower in British Columbia. August 8, 1977. (J.S. Britt) 5:3 Requirements for Dental Auxiliary Hanpawer in British Columbia - Present and Projected to 1980. February, 1979 NEW HSRD SERIES SPECIAL REPORTS (continued) 5:4 Respondents in the Canadian Dietetic Association survey of Dietetic Personnel in British Columbia - A Preliminary Report. January 26, 1979 5:5 Likely Demand for Medical Laboratory Technologists in the Period 1979-1984. March, 1979. (A.J. Stark, C.W. Kinnie) 5:6 General Manpower Stock Simul.ator (GMSS) Simulation of Possible Numbers of Physicians in British Columbia to 1993. July, 1979. (G. Muir, A.J. Stark) 5:7 Study of Dietetic Personnel in British Columbia - Research Results of the Dietetic Man-Power Pilot Project for the canadian Dietetic Association. 1978-1979. (C.W. Kinnie) 5:8 Diagnostic Ultrasound in B.C., 1979-1980, Provision and Utilization. September, 1980. (M.L. Barer, C.W. Kinnie, s. Ross) 5:9 A Survey of Difficult-to-Fill Positions for Registered Nurses and Other Health Care Disci-plines in British Columbia, 1980 - First Year-End Report. (A.J. Stark, C.W. Kinnie) 5:10 case Mix Adjustment in Hospital Cost Analysis: Information Theory Revisited. June, 1981. (M.L. Barer) 5:11 Registered Nurses in British Columbia - A Report on the Current Supply Situation. April, 1982. (A. Kazanjian, G. Wong) 5:12 Occupational Therapy and Physiotherapy Manpower in British Columbia, 1981: Current Supply and Future Requirements as Reported by Employers and Selected Senior Therapists. September, 1982. (J, MacKinnon, C. Matthews, P. Wong Fung, A.J. Stark) 5:13 Health Record Personnel in British Columbia. September, 1982. (M.L. Barer, S.E. Ross. K. Brothers, S. Jansen, B. McCashin, A.J. Stark) 5:14 Referral Patterns, Full-Time-Equivalents and the 'Effective' Supply of Physician Services in British Columbia. November, 1983. (M.L. Barer, P. Wong Fung, D. Hsu) 5:15 Refining the Need for Post-Basic Instructional Programs for Registered Nurses and Suggestions for the Provision of Learning Opportunities - A Discussion Paper. June, 1984. (A.J. Stark, C.W. Kinnis) The Division of Health Services Research and Development Office of the Coordinator, Health Sciences 400 - 2194 Health Sciences Hall University of British Columbia Vancouver, Canada, V6T 1Z6 


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