Canadian Nurses Association 50 Driveway Ottawa, ON K2P 1E2. Tel.: or Fax: Website:

Transcription

1

2 This report has been prepared by CNA to provide information on a particular topic or topics. The views and opinions expressed in this report do not necessarily reflect the views of the CNA Board of Directors. All rights reserved. No part of this document may be reproduced, stored in a retrieval system, or transcribed, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without written permission of the publisher. Canadian Nurses Association 50 Driveway Ottawa, ON K2P 1E2 Tel.: or Fax: Website: May 2009 ISBN

3 Authors Gail Tomblin Murphy Professor School of Nursing /Faculty of Health Professions and Director WHO/PAHO Collaborating Centre in Health Workforce Planning and Research Dalhousie University, Halifax Nova Scotia Stephen Birch Professor Centre for Health Economics and Policy Analysis McMaster University, Hamilton, Ontario, Canada Rob Alder Associate Professor of Epidemiology Faculty of Medicine and Dentistry University of Western Ontario Adrian MacKenzie Analyst School of Nursing/Faculty of Health Professions WHO Collaborating Centre on Health Workforce Planning and Research Dalhousie University, Halifax Nova Scotia Lynn Lethbridge Analyst School of Nursing/Faculty of Health Professions WHO/PAHO Collaborating Centre in Health Workforce Planning and Research Dalhousie University, Halifax Nova Scotia Lisa Little Director Public Policy Canadian Nurses Association Amanda Cook Analyst TMCI Toronto Ontario

4

5 FOREWORD Dear Colleagues, The Canadian Nurses Association (CNA) is dedicated to advancing the registered nursing profession, to ensuring that Canadians have timely access to quality registered nursing care, and to promoting health for all. CNA recognizes that health human resources planning is complex. Finding solutions to the problem of nursing shortages is a responsibility shared between federal, provincial and territorial governments, educators, professional associations and/or colleges, unions, employers and other stakeholders. As a policy-focused organization, CNA has a long history of working with governments and other partners to track and analyze issues surrounding the health workforce. CNA contributes research findings, policy documents, statistics, quantitative analysis and other resources to support policy and planning decisions that affect the nursing workforce. Tested Solutions for Eliminating Canada s Registered Nurses Shortage the third in a series of reports on the RN shortage presents a picture of the supply of and requirement for RNs in direct/clinical care (excluding nurse practitioners) in Canada over 15 years. Most importantly it shows how several viable policy options were tested, and highlights their ability to address the projected shortfall. It is these results that are the basis for the report recommendations. CNA is committed to sharing information and collaborating with nursing and other health-care stakeholders to ensure Canada has a sustainable, self-sufficient supply of health human resources. We hope that this report will bring us closer to achieving this. Kaaren Neufeld, RN, MN President Tested Solutions for Eliminating Canada s Registered Nurse Shortage i

6

7 EXECUTIVE SUMMARY There are a quarter-million registered nurses (RNs) in Canada. Representing the largest profession in the health-care workforce, the RN labour market must be carefully managed to meet the diverse and changing health-care needs of Canadians now and in the future. To this end, the Canadian Nurses Association (CNA) has created a national planning model, specific to RNs who provide direct clinical care, which estimates the supply of and requirement for RNs (excluding nurse practitioners) in Canada for each year over 15 years. Unfortunately, due to data limitations, the health needs of the Aboriginal population are not specifically accounted for in this model instead the model is based on the needs of the entire Canadian population, including Aboriginals. This model is aligned with federal, provincial and territorial policy, which calls for population health needs-based health human resource planning. Based on the model, this report estimates: 1. The future requirements for RNs services based on the size, distribution and levels of health-care needs of the population; and 2. The future availability of RNs services based on the size and characteristics of the current workforce as well as trends in entries to and exits from the workforce. The model goes one step further by enabling planners to gauge the effects of a policy or combination of policies on the supply of RNs and requirements for them in the short, medium and long term. As a result, this report is able to estimate the extent to which the implementation of certain policies would alleviate the projected RN shortage in Canada. Based on the best available data and a number of planning assumptions, the simulation model suggests that: There was a shortage of nearly 11,000 full-time equivalent (FTE) RNs in Canada in If the health needs of Canadians continue to change according to past trends, and if no policy interventions are implemented, the shortage of RNs in Canada will increase to almost 60,000 FTEs by Six policy scenarios rooted in these and other findings, and validated by nursing stakeholders were evaluated to determine their impact on the projected shortage. The results show that, on their own, each of the policy scenarios would yield the following results: Short-term: Increasing RN productivity 1% per year would have a dramatic effect on the gap, cutting the shortage by about 47% over 15 years. The effects of this policy would be seen within the first year of implementation. Reducing RN absenteeism (which currently averages 14 days/year) by half over three years would be equivalent to 7,000 new RN FTEs entering the workforce between 2007 and Tested Solutions for Eliminating Canada s Registered Nurse Shortage iii

8 Long-term: Reducing RN exit rates to 2% for RNs under the age of 60 and to 10% for those 60 and over would reduce the RN shortage by about half, or 30,000 FTEs, by Reducing attrition rates in RN entry-to-practice education programs (from 28% attrition to 15%) could cut the gap by about 24% to roughly 45,000. This scenario has a slightly more pronounced effect earlier than an enrolment increase. Increasing enrolment in RN entry-to-practice education programs by 1,000 per year from 2009 to 2011 would result in a substantial reduction in the 15-year RN gap (a shortage of 45,000 vs. 60,000 RNs). However, this scenario would have no noticeable effect on the gap until about Reducing international in-migration by 50% would result in a larger shortage of RNs; however, the effect of this change is not at all substantial (less than 10%), even in the long term. This is due to the fact that internationally educated RNs still represent a relatively small fraction of the national RN supply. Most interestingly, the simulation model showed that the combined effects of these six policy scenarios would be sufficient to eliminate the RN shortage within 15 years. In fact, the results suggest that should all the necessary policies be implemented, Canada could eliminate its shortage of RNs while halving existing levels of in-migration and recruitment from other countries. Canada s policy-makers, decision-makers, educational organizations, professional associations and/or colleges, employers and others are in a position to start addressing the RN shortage immediately. With these results, and future investments in data, stakeholders and governments will be able to make the right policy decisions for Canada s health-care workforce and for the health of Canadians. Recommendations: 1. Governments, employers, unions, professional associations and/or colleges, RNs and other health providers should work together to consider how they can enhance the productivity of the RN workforce. For example, removing non-nursing tasks and providing support staff, appropriate technology and equipment, interprofessional practice and/or effective organization of services would allow RNs to remain as focused as possible on the provision of quality RN patient care. 2. Governments, employers, unions, professional associations and/or colleges, RNs and other health providers should collaborate to focus workplace improvement efforts on strategies to improve the health and well-being of RNs. For example, addressing high role overload, acquiring technologies and equipment that help reduce injuries, and addressing workplace morale would all contribute to reducing the injury and absenteeism of RNs. 3. Governments, employers, unions, professional associations and/or colleges, and RNs should collaborate to improve the retention of RNs in the workforce. Although retention issues may be generation-specific, they generally include having control over one s work (autonomy), reducing high role overload, feeling valued and respected by one s employer, being included in decision-making, and having opportunities for continuing education and professional development. iv Tested Solutions for Eliminating Canada s Registered Nurse Shortage

9 4. Educational organizations, professional associations and/or colleges, student associations and governments should partner to examine opportunities to improve the retention of nursing students. Factors to consider include pre-admission requirements, guidance and campus counselors, remediation, availability of faculty, student financial support and teaching methods. 5. Governments, educational organizations and professional associations and/or colleges should collaborate at a pan-canadian level to increase enrolment in RN education programs by considering a variety of delivery models, availability of faculty, location of programs and opportunities for interprofessional education. 6. HHR planning should employ a continuous, comprehensive, multifaceted approach considering a variety of policy options (such as those tested in this report) to achieve greater self-sufficiency. Investments in one policy area, such as improving work environments, may simultanelously affect many issues, such as retention, health of nurses and student attrition. 7. Governments, employers and professional associations and/or colleges should invest in data including coordinating and linking data currently collected with a particular focus on: the amount and type of services RNs provide according to the health needs of the patient/population; any aspect of the work done by RNs working outside acute care (e.g., long-term care, home care, in the community), the level of service they provide, activity rates, participation rates and productivity; level of retention of nurses, both practicing and newly educated; and rates of attrition among RN education programs. 8. Governments, employers, unions and professional associations and/or colleges should invest in a national health provider unique identifier to provide more accurate and reliable HHR data. Tested Solutions for Eliminating Canada s Registered Nurse Shortage v

10

11 CONTENTS FOREWORD...i EXECUTIVE SUMMARY...iii INTRODUCTION...1 METHODS...3 Conceptual Framework...3 Analytical Framework...4 Simulation Model...5 Data Elements...7 The Training Module...7 The Supply Module...8 The Work and Productivity Module...8 The Needs Module...10 RESULTS...14 Initial Data Analysis...14 Training Module...14 Supply Module...14 Work and Productivity Module...14 Needs Module...15 Simulation Modelling Results...25 Development and Testing of Policy Scenarios...29 CONCLUSIONS...37 RECOMMENDATIONS...38 GLOSSARY...40 REFERENCES...41 Tested Solutions for Eliminating Canada s Registered Nurse Shortage vii

12 Appendix I: Developing an Aboriginal-Specific HHR Planning Model...45 Health Needs...45 Age Groups...46 Service Levels...46 Appendix II: List of Data Sources...47 Appendix III: Data Limitations and Associated Assumptions...49 Appendix IV: Measurement of Low Income...51 Appendix V: Forecasting Future Needs Based on Past Trends...53 Population...53 Health Needs...53 Trending Methodology...55 Health Status and Income...56 viii Tested Solutions for Eliminating Canada s Registered Nurse Shortage

13 INTRODUCTION The quarter-million registered nurses (RNs) in Canada constitute the largest profession in the health-care workforce. They play a key role in illness prevention and health promotion, as well as treating illness and helping individuals, families and communities throughout the life cycle. Given the importance of this profession to the health and well-being of those it serves, it is crucial that Canada carefully manage its RN workforce so it can meet the diverse and changing health-care needs of the Canadian population. Health human resources (HHR) planning is a priority for many stakeholders, including governments, professional associations and/or colleges, unions, employers, and research funding agencies. HHR accounts for about 70% of health-care operating budgets (Advisory Committee on Health Delivery and Human Resources, 2004). The strategic value of meaningful HHR planning is clear provincial, territorial and federal (jurisdictional) health-care systems need enough health-care providers to meet the requirements for service. While this objective is quite straightforward, identifying the best strategies for ensuring an adequate supply of HHR can be challenging. To project future service requirements, traditional approaches to HHR planning have relied primarily on age- and gender-standardized provider-to-population ratios. However, it is widely acknowledged that provider-to-population ratios are of limited value (Birch et al., 1994; Lavis and Birch, 1997; Birch, 2002; Birch et al., 2003). First, these approaches tend to assume that the future demand for health care is a function of the current supply of health-care providers (this is typically referred to as supply-based planning). Where the workplace deployment and use of RNs, licensed practical nurses (LPNs) and registered psychiatric nurses (RPNs) is changing consider, for example, the rapidly evolving role of the nurse practitioner (NP) in Canada existing measures of service provision (i.e., current provider-topopulation ratios) are of no relevance to future planning. Second, provider-to-population ratio approaches to HHR planning do not take into account the fact that health-care needs of populations will vary between communities and over time, despite the fact these variations are well documented (Evans, 1994; Roos et al., 2001; O Brien-Pallas et al., 2007). Third, supply-based approaches to HHR planning do not account for variations in the productivity of providers between health-care settings and over time. New technologies and new models of service delivery can have a marked effect on provider productivity (Gray, 1982). However, recent reviews have found little evidence of these considerations being factored in by planners or researchers (Tomblin Murphy et al., 2007; Tomblin Murphy et al., 2004). CNA and members of the current project team have been involved in efforts to develop and implement more comprehensive and effective HHR planning processes. Recent work in this field includes the Atlantic Health Human Resources Planning Study (Birch et al., 2005) and work on the Canadian Nurse Practitioner Initiative (Tomblin Murphy et al., 2006). In both these projects, the authors developed models for needs-based HHR planning. These models estimate the future requirements for the services of a given group of providers based on the size, distribution, and levels of health-care needs of the population. In addition, they estimate the future supply of services, based on the size and characteristics of the current workforce as well as trends in entries to and exits from the workforce. With these models, planners can gauge the effects of various policy scenarios, as well as combinations of policies, on the balance between supply of and requirements for health-care providers over short-, medium- and long-term planning periods. Tested Solutions for Eliminating Canada s Registered Nurse Shortage 1

14 The objectives of this study were to estimate, for each year over 15 years, the requirements for and supply of direct/clinical care RNs in Canada (excluding those who are also licensed as NPs), and to develop a simulation model that permits the testing of various HHR policy scenarios. To this end, CNA and the project team have collaborated in building a planning model specific to RNs at the national level (though the model can be adapted to the provincial/territorial level as well). The present study is directly informed by the Atlantic and Canadian Nurse Practitioner Initiative projects, and builds on that work. The conceptual and analytical frameworks used in this study and the earlier ones are discussed in the following chapter. One of the initial objectives of this study was to develop a planning model specific to the health-care needs of Canada s Aboriginal population. Unfortunately, after substantial exploration and consultation with the National Aboriginal Health Organization and the First Nations and Inuit Health Branch, it was determined that the data presently available on the health-care needs and health-care utilization of Aboriginals in Canada are not sufficient to support needs-based-planning for this population, and as such this model has not been developed. However, substantial progress was made in this study in HHR planning for the Aboriginal population, particularly in the identification and assessment of data sources. For a detailed discussion of the effort to develop an Aboriginal-centred HHR planning model, see Appendix I. 2 Tested Solutions for Eliminating Canada s Registered Nurse Shortage

15 METHODS Conceptual Framework The methodological approach in this study is informed by a conceptual framework developed by O Brien- Pallas et al. (2005). This framework has been adopted as a guiding framework for use in HHR planning by Canada s Advisory Committee on Health Delivery and Human Resources (ACHDR, 2005). As seen in Figure 1, the outer oval represents the context of social, political, geographical, technological and economic factors in which HHR planning takes place. Fundamentally, however, HHR planning starts with the health-care needs of the country, province/territory or region in question. Figure 1: Conceptual Framework. Tomblin Murphy & O Brien-Pallas, 2006 Adapted from O Brien-Pallas, Tomblin Murphy & Birch (2005), O Brien-Pallas, Tomblin Murphy, Birch & Baumann (2001) and O Brien-Pallas & Baumann (1997) Tested Solutions for Eliminating Canada s Registered Nurse Shortage 3

16 Analytical Framework The conceptual framework is the foundation for the analytical model, which can then generate simulations and recommendations. The analytical model (Birch et al., 2007) consists of two broad elements: provider supply and provider requirements. Provider supply is, in essence, the answer to the question How many providers are available to deliver health-care services to the population? Supply can be seen as the outcome of two broad determinants: 1. The stock of individuals, namely the number of providers in each age and sex group who are potentially available to provide health-care services 2. The flow of provider time from the stock, influencing the quantity of service output in short, time spent in the production of services. This time depends on: a. The proportion of the current stock participating in providing health care, or the participation rate b. The quantity of time devoted to service provision by those who do participate in the provision of health care, or the activity rate Participation and activity rates represent policy levers for HHR policy-makers and, hence, alternative or complementary approaches for changing provider supply. In addition to changes in the flow of provider time, the size of the stock changes with new entrants (inflows of health-care providers from other countries together with new graduates from within Canada) and departures from the stock (outflows of providers to other countries, and retirements and deaths among providers). In terms of policy responsibilities, education and training (i.e., the production of new providers) are generally separate from the management and regulation of providers (the use of existing providers). Thus provider supply can be seen as the combination of two components: training of new providers and management of existing providers. The second element of the analytical model, provider requirements, has four distinct components: 1. Demography: the number of people by age and gender group in the population. 2. Epidemiology: the rate of health and illness as well as risk factors for future illness across the population subgroups. 3. Level of service: the amount of health-care services to be provided for individuals at different levels of illness or risk of illness. 4. Productivity: the amount of health-care services a full-time equivalent (FTE) provider performs per unit of time. Because each of these components varies across age and gender groups in a population, the analytical model is applied to each age-gender group to come up with the provider requirements for each group. These results are then added together to provide an estimate of total provider requirements. Combining the first three components of the framework demography, epidemiology and level of service yields an estimate of the number of health-care services required by a population, given its size, demographic mix, levels and distribution of health and illness, and levels of service. The fourth component productivity 4 Tested Solutions for Eliminating Canada s Registered Nurse Shortage

17 translates the number of services required into the number of health-care providers required to perform them. The framework for this needs-based approach can be written mathematically as follows: (1), where R t is the number of providers required to meet the service requirements of a population at time t; N i,j,t is the number of providers required to perform each service to patients of age group i and sex j at time t (i.e., the inverse of productivity); Q i,j,t is the number of services required by level of need per person of age group i and sex j at time t; H i,j,t is the proportion of the population by level of need (in the simplest case this would be the proportion of the population who are sick ) for age group i and sex j at time t; P i,j,t is the size of the population of age group i and sex j at time t. Traditional HHR planning methods have limited attention to demographic change, by applying current levels of service use by age and gender to the estimated changes in the size and mix of the future population. This assumes that health-care needs of the population, the types and quantity of health-care services delivered to meet those needs, and the ways in which such services are delivered are constant over time. In contrast, the needs-based model allows policy-makers to explicitly consider the separate effects of each of these factors as they plan for their health-care workforces. Simulation Model Building on the analytical framework, a simulation model (Kephart et al., 2005) has been adapted that simultaneously estimates present and future HHR requirements and present and future HHR supply. The model was designed using a system dynamics approach (Forrester, 1968; Richardson, 1991; Sternman, 2000) and implemented using Vensim (2002) simulation software. Tested Solutions for Eliminating Canada s Registered Nurse Shortage 5

18 Figure 2: Simulation model Adapted from Kephart et al., 2005 The components of the model are grouped into four distinct modules. The training module estimates the flow of new graduates out of education/training programs. The supply module estimates the size of the stock of providers available to deliver health-care services. The needs module estimates the number of health-care services required by a population. The work and productivity module translates service requirements and counts of available providers into standard FTE providers required and available, respectively. The model is not necessarily designed to predict the future, but rather to integrate knowledge of different components of the health-care system to better understand how various factors affect the supply of and/or requirements for health-care providers. The model enables policy-makers to rehearse potential policy changes; by altering variables in the model, they can see the effects of each change on the supply of and requirements for a given type of health-care provider. Moreover, because many of the variables in the model lie beyond the scope of HHR policy (e.g., needs for care, population size), it supports consideration of the effects of HHR policies in different future contexts. In this way, it gives policy-makers a means of testing and evaluating policy options to determine the most efficient and effective ways to manage HHR under different future scenarios. Often, a central concern for policy-makers is the gap, or difference, between the number of providers available and the number required hence the prominence of this component in Figure 2; the model has been designed to calculate this gap automatically. 6 Tested Solutions for Eliminating Canada s Registered Nurse Shortage

19 The accuracy of any model s simulations depends greatly on the quality of the data used to run it. Unfortunately, timely access to reliable data is a challenge for HHR planners and researchers alike. This model, therefore, has been designed so that users can easily update it as newer and/or better data become available. The model was originally designed to be able to simulate HHR gaps over a 40-year period. In this study, however, it was used to consider the supply of and requirements for direct/clinical care RNs (with the exception of NPs) in Canada over a 15-year period. This time frame lines up with policy-making in Canada, which uses short- (1- to 2-year), medium- (2- to 5-year) and long-term (5- to 15-year) time frames. The individual elements of the simulation model follow, presented according to the module to which they belong. A list of the sources for each data element is provided in Appendix II, and a tabulated list of any associated assumptions appears in Appendix III. Data Elements The Training Module Domestic nursing education programs account for the bulk of new entrants into the stock of RNs in Canada, and as such are essential to maintaining Canada s supply of RNs. The training module estimates the future number of graduates from Canadian RN entry-to-practice (ETP) education programs using data on enrolment, length of programs, attrition rates and graduate retention rates. All figures for the training module are from the CNA/Canadian Association of Schools of Nursing (CASN) National Student and Faculty Survey of Canadian Schools of Nursing. Enrolments: The number of first-time entrants in all RN ETP education programs in Canada. Program length: The duration of RN ETP education programs in Canada. The value used in the simulations is the average (weighted by program size) of all programs. Program attrition: The proportion of entrants to all Canadian RN ETP education programs who leave the programs before completion. This can also be seen as the opposite of the graduation rate. Seventeen programs at 22 schools/school sites spanning seven provinces provided data as part of the National Student and Faculty Survey of Canadian Schools of Nursing, which supported calculation of attrition rates. Graduate out-migration: The proportion of new graduates from Canadian RN ETP education programs who do not enter practice as RNs in Canada. This includes graduates leaving Canada to practice in other countries. Estimates of this proportion were available from British Columbia, Ontario and Saskatchewan, and the average of these three provincial rates was used to approximate the national rate. The Ontario report, however, does not specify whether out-migration from Ontario is to other parts of Canada or to other countries; thus this report may overestimate the rate of national outmigration of Ontario graduates. As previously noted, the model can be easily updated should a better estimate become available. New graduates: The number of new graduates from Canadian RN ETP education programs each year is determined by enrolment, program length, attrition and out-migration rates. The age distribution of these new graduates is based on the average age distribution of Canadian-educated first-time writers of the Canadian Registered Nurse Examination over the past seven years (provided by CNA). Tested Solutions for Eliminating Canada s Registered Nurse Shortage 7

20 The Supply Module The supply module estimates the future size of the RN stock in Canada. This is a function of its current size and the flow of RNs into and out of the stock. This estimate is based on the following factors: Existing RN stock: The number of RNs in Canada who are potentially available to provide nursing services. This includes all individuals with the appropriate RN licensing, whether or not they are currently practising direct nursing care as an RN. This does not include RNs who are also licensed as nurse practitioners. Data on the current size of the stock of RNs in Canada by age were obtained from the Canadian Institute for Health Information (CIHI) s national Regulated Nursing Database. In-migration: The number of RNs entering practice in Canada from other countries. This is an estimate based on the number of internationally educated writers of the Canadian Registered Nurse Examination in In-migration is combined with new entrants from the training module (minus out-migrating graduates) to arrive at the total flow of new RNs into the stock. Exit rates: Every RN, at some point, will cease to practice, as is true of any professional. They do so for a variety of reasons, such as relocation to another country, burnout, retirement or death. Exit rates are the age-specific rates at which RNs in Canada cease renewing their professional registration and hence are no longer licensed to provide RN services. Ideally, national exit rates would be measured using a national RN database with a unique identifier for each nurse. In the absence of such a system, CNA obtained data on age-specific provincial/territorial-level non-renewal rates from provincial/ territorial RN regulatory bodies. These were then adjusted for interprovincial/territorial migration, since these relocations do not represent a loss of RNs to Canada. Based on these factors, the supply module provides estimates of the number of RNs available to provide care over time. As we have seen, however, the quantity of services provided by the RN stock depends on factors associated with working practices, such as the productivity and activity rates of the RNs in the stock. These are accounted for in the work and productivity module. The Work and Productivity Module The degree to which the supply of RNs (i.e., the RN stock) is sufficient to meet requirements depends on the amount of time RNs contribute to service delivery and their productivity during that time. Hence, the work and productivity module translates the size of the RN stock into the supply of RN hours and the rate at which RNs can provide services over time. Thus the work and productivity module allows for the comparison of the supply of and requirements for RN services. The module consists of the following components: Participation rate: The proportion of all licensed RNs in Canada who are employed in direct patient/clinical care. This does not include licensed RNs who are working in other sectors (such as research, education, policy or administration), who are not currently in the labour market (e.g., on parental or educational leave) or who have become licensed as nurse practitioners. Activity rate: The proportion of an FTE s hours that the average RN employed in direct/clinical care provides. Ideally, this would be estimated using Canada-wide administrative data (such as CIHI Management Information System data) on the worked hours of RNs. However, figures were available only for acute-care RNs and only for three provinces New Brunswick, Nova Scotia and Ontario. 8 Tested Solutions for Eliminating Canada s Registered Nurse Shortage

21 Therefore, the ratio of worked hours to actual RNs in these three provinces was used as a proxy for the ratio for all of Canada. As noted, these values are easily replaced should actual data become available. The National Survey on the Health and Work of Nurses (CIHI, 2006) included nurses own estimates of the hours they worked in an average week. However, when extrapolated to yearly estimates, these self-reported data differed substantially from all available administrative data, perhaps indicating that self-reported weekly estimates are not an accurate basis for estimating yearly worked hours. Productivity: The number of services performed per FTE RN per year. Because of the significant differences in the type of care provided across various sectors, this value was estimated separately for RNs in acute care, long-term care, home care and community care (which includes primary health care and public health nursing). Estimating productivity was complicated, however, by a general absence of administrative data on the actual amounts of care provided by RNs, except in the acute-care sector. In each sector, productivity was estimated by taking the best available measure of the services performed per unit of time and dividing that by the number of FTE RNs working in that sector during that time. In acute care, the best available measure of RN services was acuity-adjusted episodes of care. These data were provided by CIHI and broken down by in-patient, day surgery and emergency episodes of care. No figures were available from Quebec, day surgery data were not available from Alberta, and emergency data were available only from Ontario. In an attempt to estimate the total number of acuity-adjusted episodes of care in Canada, the model relied on provinces for which data were available. The ratio of the population to the number of acuity-adjusted episodes of care by age and sex in those provinces for which data were available was used as an adjustment factor to amplify the provincial data to a national estimate. As noted, CIHI data on productivity of RNs working in acute care were limited to RNs providing in-patient, day surgery and emergency care. Without information on the contribution of RNs to out-patient care in acute-care organizations, the model underestimates the total productivity of acute-care RNs. However, the impact of this on the model is more or less negated by corresponding underestimates in the levels of service they provide (see the level of service discussion that follows). As noted, the model can easily be updated should more relevant or complete data become available. In long-term care, the only available measure of services provided by RNs was the total patient days of care provided in long-term care facilities. This information was obtained from Statistics Canada s Residential Care Facilities Survey (RCFS Statistics Canada, 2008a). In home care, the only measure of RN services was information from patients about the number of home-care visits they received. These self-reported data came from Statistics Canada s Canadian Community Health Survey (CCHS Statistics Canada, 2005a). No information was available on the services performed during visits or their duration. Similarly, in the community sector, the only available measure of the services provided by RNs was reports from patients on their use of community nursing, again from the CCHS. Information was limited to the number of consultations with a nurse, with nothing on the services performed or the duration of the consultations. Tested Solutions for Eliminating Canada s Registered Nurse Shortage 9

22 In this way, the work and productivity module translates the number of available RNs (estimated by the training and supply modules) into the number of FTEs available based on the proportion of RNs participating in patient care and the hours they work. The rate at which an RN FTE can provide patient care services translates the needs-based service requirements (estimated by the needs module) into the number of RN FTEs required. The Needs Module Estimating the service requirements for RNs in Canada based on the health-care needs of the population uses the following components: Population: The size of the Canadian population by age and sex, at present and projected into the future. Statistics Canada s (2005) medium-growth population projections were used to estimate future population size. Need: An estimate of the level of need for RN services. Because of the significant differences in the services provided by RNs across sectors, different measures for needs were used for each sector; see Table 1. The aim was to identify indicators of population health needs that were appropriate to the particular sector, and that were independent of service use. However, choice of need measures was limited by availability of data over several years, which is required to consider trends over time. For acute-care RN services, the measure of need was a combination of rates of injury and chronic conditions, and self-assessed unmet requirements for health-care services. Utilizing injuries and chronic conditions as need indicators was based on the need for acute care RN services arising from either acute episodes of need brought on by chronic conditions (e.g., asthma attacks, hypertensionrelated cardiac incidents) or injury (e.g., fractures, lacerations). Unmet need was included since it suggests a level of service is required beyond that reflected in current utilization patterns. A potential problem with using chronic conditions and injury as needs indicators is their dependence on utilization. Because individuals are identified as having a chronic condition or an injury based on a health professional s diagnosis, as access to health care and diagnostic criteria expand, rates of diagnoses will increase even if the prevalence in the population does not change; the population may appear to be becoming less healthy when in fact it is not. Combined rates of injury and chronic conditions were estimated based on self-reported data from the CCHS. The proportion of the population within each level of need reporting unmet needs for hospital services (emergency, overnight and outpatient) was calculated from the CCHS. Categories are not mutually exclusive; that is, individuals reporting unmet need in more than one service are counted more than once. To quantify a level of service for this group, the total number of services for each need level was inflated by the percentage reporting unmet need. The total inflation amount can be considered the level of service for those with unmet need 1. 1 This is used as a possible scenario using the assumption that the required proportionate increase in level of service to accommodate current self-reported unmet need for care is equal to the proportion of the population within each need level reporting unmet need for care. However, there is no objective basis for this assumption and alternative scenarios for the levels of unmet need for care can easily be accommodated in the model should information become available. 10 Tested Solutions for Eliminating Canada s Registered Nurse Shortage

23 For long-term care, the proportion of the Canadian population requiring services was estimated by combining: (1) individuals already in long-term care facilities (on the assumption that they have met explicit needs criteria for admission); (2) individuals occupying acute-care beds while waiting for beds in long-term care facilities (i.e., alternate level of care, or ALC, patients); and (3) individuals living alone and being unable to perform personal care and/or mobilize in their homes without help. The population in long-term care was estimated using the Residential Care Facilities Survey. The number of individuals living alone and requiring assistance with personal care and/or to move about their homes was estimated using data from the CCHS. The number of ALC patients was not available at a national level. Corresponding data for Ontario were provided by the Ontario Hospital Association. These were used to generate a rate of ALC patient population. This rate was then used as a proxy for the average rate of ALC patients in Canada. For home-care RN services, a combination of met and unmet need was used. The proportion of the Canadian population requiring home-care RN services includes both those who already receive such services (having met explicit criteria of need used in the assessment procedures of home-care programs) and those who do not but identify themselves as requiring such services. The assumptions of this approach are that Canadians do not receive home-care nursing services unless they meet objective criteria of need and that self-reports of unmet need for home-care services are valid. Data on both measures were obtained from the CCHS. For community RNs, the distribution of self-assessed general health status by income level was used. This is because community RNs provide primary health-care and/or public health services, with much of the public health work aimed at low-income families or individuals. Self-assessment prompts people to visit a primary health-care provider, and such providers respond to general health issues. Measures of self-assessed general health status have also been found to correlate with a wide range of health and socioeconomic variables, both at the population and individual level (Birch et al., 1996). Data on self-assessed health by income level were obtained from the CCHS. For a detailed discussion of how income levels were measured, see Appendix IV. As with the acute-care sector, unmet needs were estimated using self-assessed reports of health-care services perceived to be required, but not received. The percentage of respondents in the CCHS who report trying unsuccessfully to receive services from a doctor s office, walk-in clinic, appointment clinic or a community health centre was calculated within each level of need for community nursing care, with multiple selections counted more than once. Note, however, that only a fraction of the respondents reporting unmet care from a doctor s office was included as not all offices employ an RN. A lack of data on the proportion of services performed at a doctor s office that could be administered by an RN meant expert consultation was required to estimate this value. Twenty-five percent of those reporting unmet care from a doctor s office was chosen, but this should only be viewed as a possible scenario until more detailed data become available. Finally, as was done in the acute-care sector, the level of service for each need category was inflated by the percentage with unmet needs 2. 2 As in Note 1, this is used as a possible scenario using the described assumptions, and alternative scenarios for the levels of unmet need for care can easily be accommodated in the model should information become available. Tested Solutions for Eliminating Canada s Registered Nurse Shortage 11

24 Table 1: Needs Indicators and Data Sources by Sector Sector Needs Indicators Data Source Acute care Long-term care Home care Community care Injury Number of chronic conditions Living alone and requiring assistance with activities of daily living (ADLs) Use of alternative level of care (ALC) beds in hospitals Residence in long-term care facilities Receipt of home-care nursing services (publicly or privately funded) Self-reported unmet need for home-care nursing services Self-assessed health status by income level Self-assessed unmet need CCHS CCHS CCHS OHA RCFS CCHS CCHS CCHS CCHS When planning to allocate resources based on health-care needs, it is necessary to estimate what those future needs will be. In this study, population demographics population and health-care needs were projected over a 15-year time horizon and fed into the planning model. No one can predict with certainty how the health-care needs of Canadians will change in the future. Still, if these needs continuing to change as they have in the recent past, it is possible to simulate the effects of these changes by fitting statistical models to the historical data and projecting forward. For a detailed discussion of the methodology used to project past trends in health needs into the future, see Appendix V. Level of service: This measures the amount of care or services a person requires by level of need. In the absence of gold standards, existing levels of service (where available) were used as a baseline, keeping in mind that the model can be changed or updated as desired. Moreover, unlike the other elements of the model, the level of service will be influenced by non-clinical factors such as the resources available for the delivery of care and the other demands on those resources. The same unit of service volume was used in the calculation of both level of service and productivity, before these measures were multiplied to estimate FTE RN requirements. Because of significant differences in the type of care provided across sectors, level of service was estimated separately for acute care, long-term care, home care and community care. In acute care, data on acuity-adjusted episodes of care were not available by level of need. Instead, estimates were made by combining data from various sources. Specifically, the distribution of actual hospitalization by chronic conditions and injuries was assumed to be the same as the self-reported distribution in the CCHS data. Institutional-level data (provided by CIHI, as described previously) on actual hospitalizations by age and sex were combined with self-reported population-based data on hospitalizations by age and sex for the same year to calculate the ratio of actual utilization to self-reported use by age and gender. This ratio was then applied to self-reported use by level of need (i.e., chronic conditions and injuries). 12 Tested Solutions for Eliminating Canada s Registered Nurse Shortage

25 As no administrative data were available on the amount of outpatient services provided by RNs, this value could not be included in simulations; thus the level of service provided by RNs is underestimated in these models. However, the impact of this on the accuracy of the model is more or less negated by the fact that productivity is similarly underestimated due to this same data gap. In long-term care, the amount of care measured by days of care in long-term care per patient per year used was from work by Tomblin Murphy et al. (2008) for Ontario. The level of service for long-term care patients in Ontario was used as a proxy for the average level of care for long-term care for Canada. Because the level-of-service measure was not estimated by patient age and gender, the same rate was applied to all age and gender groups admitted to long-term care. The report by Tomblin Murphy et al. (2008) also provided an estimate of level of care for the homecare sector, but the same limitations applied to using this source of data as those listed previously for long-term care. For the community sector, the number of community nursing consultations per person by income level and self-assessed health status was estimated from self-reported data in the CCHS. For each year of the projection period, the needs module estimates the future requirements for the healthcare services of RNs by combining the projected size of the population and its health status rates with estimates of the amounts of RN health-care services it receives by health status. Using RN productivity data (number of services delivered per FTE RN per year), the requirement for services is converted to the requirement for FTE RNs in each year. This value is then compared with the estimate of the future availability of FTE RNs in each of those years to determine whether there will be a surplus or shortage of RNs in Canada. The results produced by the simulation model (see the following section) are dependent on the range and quality of data available and should be interpreted with caution; still, they are based on the best data available. The qualitative findings, in particular, provide valuable insight to planners. Further, this approach will become more valuable as new and better-quality data become available to support efficient, effective HHR planning. Tested Solutions for Eliminating Canada s Registered Nurse Shortage 13

26 RESULTS Initial Data Analysis Before the model was populated and simulations run, substantial analysis of the data from various sources was required to extract the necessary information. The results of this analysis are organized by the module to which the data pertain. Training Module According to data from the Nursing Education in Canada Statistics, report, in 2007 there were an estimated 13,000 first-time entrants in all ETP programs. Recently announced funding increases in several provinces will increase this number by about 900 by the end of the academic year. Based on figures from the 17 programs that provided sufficient data to calculate graduation rates, approximately 28% of students admitted to RN ETP education programs do not complete them. In addition, recent estimates from three provinces put the proportion of new graduates staying in the province to practise at 95% (CRNBC, 2005), 98% (Insightrix, 2007) and 96% (OMHLTC, 2008) in British Columbia, Saskatchewan and Ontario, respectively. Data from the Nursing Education in Canada Statistics, report indicate that the number of RN graduates in Canada nearly doubled between 1999 and 2007, increasing from 4,833 to 9,447. The average age of these graduates was approximately 26. Supply Module Data from the Regulated Nursing Database show that as of 2007 there were approximately 270,000 licensed RNs in Canada. Of these, about 217,000 or 81% are employed in the delivery of direct/clinical care. The average age of these RNs is 44. There is an annual influx of approximately 1,000 RNs from other countries entering practice in Canada, with an average age of 35. Provincial/territorial registration data provided to CNA, adjusted for inter-jurisdictional migration, indicate that about 3% of RNs in Canada cease to renew their registration each year (varying from less than 2% for 35-to-49-yearolds to more than 10% for those over 60). Stated another way, these data indicate that Canada retains about 97% of its potential RN workforce from year to year. Work and Productivity Module CNA (2008) has defined an FTE RN as one who accumulates 1,950 earned hours in a year. This is based on RN collective agreements in Canada. However, since RNs do not deliver patient care during vacations or sick time, the model uses actual worked hours to measure the activity of RNs. According to CIHI Management Information System data on nursing hours in hospitals in Nova Scotia, New Brunswick and Ontario, about 83% of earned hours are actually worked; thus an FTE RN is defined here as one who works 83% of 1,950 hours, or 1,618.5 hours, per year. Other administrative data sources indicate that the 54,000 hospital-based unit-producing RNs in these provinces worked a total of about 75 million hours in 2005, or over 1,392 hours per RN. This quantity is about 86% of the 1,618.5 worked hours per year an FTE accumulates, meaning the average acute-care RN in these provinces represented approximately 86% of an FTE. As this was the best available estimate of the activity level of any type of RN, 86% was used as the activity level of all RNs. 14 Tested Solutions for Eliminating Canada s Registered Nurse Shortage