Health human resources forecasting: Understanding the current and future requirements of PSW s and nurses in Ontario s LTC sector

Transcription

1 Health human resources forecasting: Understanding the current and future requirements of PSW s and nurses in Ontario s LTC sector Presented by: Adrian Rohit Dass, MA IHPME, University of Toronto Canadian Centre for Health Economics PI s: Audrey Laporte, Ivy Bourgeault Team: Raisa Deber, Andrea Baumann, Janet Lum, Amy Hsu, Christine Kelly Preliminary - Please do not cite or quote without permission from the authors

2 Presenter Disclosure 2 Speaker: Adrian Rohit Dass Relationships with commercial interests: Grants/Research Support: N/A Speakers Bureau/Honoraria: N/A Consulting Fees: N/A Other: N/A This project is supported with funding from the Government of Ontario through the Bruyère Centre for Learning, Research and Innovation in Long Term Care. The views expressed in this publication are the views of the author(s)/presenter(s) and do not necessarily reflect those of the funder.

3 Disclosure of Commercial Support 3 This program has received financial support CIHR Operating Grant This program has received in-kind support N/A Potential for Conflict of Interest: N/A

4 Acknowledgements 4 We thank Peter Walker, Melissa Donskov, Tracy Luciani, and the team at the Bruyère Centre for Learning, Research and Innovation in Long Term Care for their input and support on the project. We also thank the Ontario Long Term Care Association (OLTCA), Ontario Association of Non Profit Homes and Services for Seniors (OANHSS), Ontario Home Care Association (OHCA), Ontario Community Support Association (OCSA), Nursing Health Services Research Unit (NHSRU), Ontario Personal Support Worker Association (OPSWA), Canadian Home Care Association (CHCA), Toronto Central Community Care Access Centre (TC-CCAC), Ontario Health Human Resources Research Network (OHHRN) and Canadian Centre for Health Economics (CCHE) for providing their expertise and facilitating connections with those in the policy realms. All errors/omissions are the responsibility of the research team.

5 Objective for today s presentation 5 Outline the main components of the model Show how we marshaled various datasets to produce the data inputs for the forecasting model Present preliminary forecasts from the model Present policy scenarios that can be explored with the model

6 Motivation 6 PSWs are essential in providing care in Long Term Care. Approximately three-quarters of direct care staff in Ontario s LTC sector are PSWs (MOHLTC, 2012). Despite their importance in administering care, relatively little is known about them Need for PSW services expected to grow over the years as proportion of population aged 65+ continues to grow From a policy point of view knowing how many PSWs will be required is essential

7 Hours of Direct Care Per Resident Day 7 H o u r s p e r R e d i s e n t D a y Registered Nurses Registered Practical Nurses Other Direct Care Staff Year Source: Residential Care Facilities Survey for Ontario ( )

8 LTC PSW Forecasting Model 8 Goal: To predict the need and supply of Personal Support Workers (PSWs) accounting for the supply of Nurses (RPNs and RNs) in Ontario s LTC sector by LHIN Approach: Multi-professional (PSW, RPN, RN) Model RN labour force as well as RPNs which has largely been neglected to date in Canada. Currently focusing on one sector-ltc but working to expand to the community Forecast at the level of the LHIN, rather than provincially Improve upon existing forecasting models, which have focused on one profession and have not accounted for the effects of different skill mixes on output.

9 Diagram for Forecasting LTC PSW labour force 9

10 Provider Supply (Continued) 10 To estimate the future stock of the Nurse and PSW workforces, we take the current stock, subtract the expected outflows, and add to it the expected inflows (all by age), which is outlined by Birch et al. (2007): For Nurses, the College of Nurses of Ontario (CNO) has a rich, longitudinal database of all nurses in Ontario For PSWs, we have the PSW registry, which is crosssectional and only (currently) mandatory for PSWs working in Home and Community Care

11 Employment Categories of PSWs 11 Source: Lum (2013) Ontario PSW Registry Data Analysis

12 Age Distribution of PSWs 12 Source: Lum (2013) Ontario PSW Registry Data Analysis

13 How to measure RN/RPN outflows from the LTC sector by LHIN? 13 RNs and RPNs: Since we have longitudinal data for nurses, we can calculate transitional probabilities to determine the proportion remaining in LTC from year to year This measure is also known as stickiness, a term coined by Alameddine, Laporte, Baumann, O Brien-Pallas, Mildon, & Deber (2006) Defined as the transition probability of a nurse working in a given setting in year t remaining in the same setting in year t+1 (sector) We have calculated transitional probabilities for RNs and RPNs over the years by age group and LHIN

14 RN Stickiness by LHIN / / / / /2010 Average Total Outflows Stickiness 82% 82% 83% 84% 87% 84% % Outflows by LHIN Central 8% 10% 9% 11% 8% 9% Central East 11% 11% 12% 9% 9% 10% Central West 4% 4% 4% 3% 4% 4% Champlain 9% 12% 11% 11% 11% 11% Erie St. Clair 7% 6% 5% 4% 6% 6% Hamilton Niagara Haldimand Brant 9% 10% 10% 11% 12% 10% Mississauga Halton 6% 5% 6% 5% 6% 6% North East 5% 5% 5% 6% 7% 6% North Simcoe Muskoka 4% 3% 5% 6% 5% 5% North West 3% 2% 3% 2% 3% 2% South East 5% 7% 5% 7% 7% 6% South West 10% 8% 9% 9% 8% 9% Toronto Central 14% 11% 13% 13% 11% 12% Waterloo Wellington 5% 5% 4% 5% 4% 5% Source: CNO Database

15 RPN Stickiness by LHIN / / / / /2010 Average Total Outflows Stickiness 82% 83% 83% 82% 84% 83% % Outflows by LHIN Central 5% 6% 6% 7% 9% 6% Central East 10% 8% 9% 12% 8% 10% Central West 2% 3% 4% 3% 4% 3% Champlain 12% 10% 11% 10% 13% 11% Erie St. Clair 7% 5% 6% 5% 5% 6% Hamilton Niagara Haldimand Brant 11% 13% 14% 12% 10% 12% Mississauga Halton 3% 4% 3% 5% 4% 4% North East 6% 10% 9% 7% 10% 8% North Simcoe Muskoka 5% 3% 4% 5% 4% 4% North West 5% 5% 5% 5% 5% 5% South East 6% 6% 6% 5% 4% 5% South West 11% 10% 10% 8% 9% 10% Toronto Central 11% 11% 10% 11% 11% 11% Waterloo Wellington 6% 6% 5% 5% 4% 5% Source: CNO Database

16 How to measure outflows of PSWs from the LTC sector by LHIN? 16 Wodchis et al (2015) found the turnover rate of full-time PSW s in LTC to be 6.5%. We therefore assume that the LTC sector retains 93.5% of it s FTE PSW workforce from year to year. In terms of outflows per LHIN, we use the same proportions as the RPN labour force In terms of outflows by age, we calculate this based on the PSW Registry Question: In the next 1-3 years, what do you plan on doing? We obtained an age distribution of PSW s intending to: retire, seek employment outside of the PSW field, return to school for a PSW Certificate, return to school for PSW professional development or continuing education, or return to school for a different career. We found intention to leave is not limited to the older ages, as a high proportion of younger PSWs intend to leave as well.

17 How to measure Nurse (RN, RPN) inflows into the LTC sector by LHIN? 17 RNs and RPNs: Alameddine et al. (2006) also have an inflow measure, defined as as the percentage of nurses that are working in a particular setting in year t, who were not working in that setting in year t-1, i.e., were fresh additions to that setting. This measure captures nurses who are new additions to the workforce (new graduates, immigration), returning to the workforce, and those switching sub-sectors. This was estimated using the CNO database for years

18 Inflows of RN s into LTC by LHIN 18 Source: CNO Database

19 Inflows of RPN s into LTC by LHIN 19 Source: CNO Database

20 How to measure PSW inflows into the LTC sector by LHIN? 20 PSWs Every year, over 7000 PSWs graduate from Ontario programs ( PSW Educational Program Accreditation According to the Ontario Ministry of Training, Colleges, and Universities website, 54% of new PSW graduates go into the Nursing and residential care facilities sector We therefore assume that 54% of 7000, or 3780 new PSW graduates enter the LTC sector in each year. In the absence of data on inflows across LHINs for PSWs, we apply the same proportions as the RPN labour force.

21 Demand for Providers 21 Before we can estimate the number of required providers, we need to estimate the number of services required, which can be done using the following indicators: (1) (2) (3) This is the needs based approach for estimating the number of services required by the population, originally created by Birch et al (2007), and used by CNA (2009) and Murphy et al (2012) to estimate the number of required services in each sector of Ontario (including Long Term Care).

22 Predicting Services Required P a t i e n t D a y s Total Patient Days (Predicted) Total Patient Days (Actual) Year Source: Residential Care Facilities Survey for Ontario ( )

23 How does this translate to the number of staff required? And what about productivity? 23 Under the traditional uni-professional approach, once the services required have been estimated, the next step is usually to divide this number by the rate at which FTE workers can perform them (productivity) to yield the required number of health workers. The conceptual framework is usually modeled as follows (adapted from Murphy et al 2012):

24 Why might this approach be incomplete? 24 From Auditor General of Ontario (2013 report), as a recommendation to the MOHLTC: further refine its forecasting models and their capabilities to assess the impact of various factors on service-provider productivity From Health Workforce Australia s National Health Workforce Productivity Modelling Final Report (2012): Current workforce planning models examine labour productivity of each health worker (e.g. number of visits per doctor), though in reality the number of visits each doctor produces is also influenced by nurses, equipment, and other resources... Models that estimate requirements for a particular group of health professionals do so independently of other groups of health professionals, leading to inaccurate results

25 Predicting the Number of FTE RN s in Ontario, F T E ' s R e q u i r e d FTE RN"s Required (Estimated) FTE RN's Required (Actual) 0 Year Source: Residential Care Facilities Survey for Ontario ( )

26 Predicting the Number of FTE RPN s in Ontario, F T E ' s R e q u i r e d FTE RPN"s Required (Estimated) FTE RPN's Required (Actual) 0 Year Source: Residential Care Facilities Survey for Ontario ( )

27 Predicting the Number of FTE PSW s in Ontario, F T E ' s R e q u i r e d FTE PSW's Required (Estimated) (Silo) FTE PSW's Required (Actual) 0 Year Source: Residential Care Facilities Survey for Ontario ( )

28 How to model workers together? 28 We need a way to relate services required (output) with the factors that influence it (PSWs, RNs, RPNs, drug expenditures, etc.). Conceptually, it would look like this for the LTC sector: This can be done by estimating a production function

29 Multi-professional Approach 29 We estimate a translog production function, which allows for interactions of the inputs, thereby allowing for the productivity of one input to depend on the level of other inputs. It also has a history in the economics literature for use of determining the complementarity/substitutability of inputs (Hsu et al, 2015). We find a statistically significant interaction effect of between PSWs and RPNs, This implies that a 1% increase in the number of RPN hours leads to a 0.016% decrease in the productivity of a PSW, all else being equal. We also found a positive (though not significant) interaction between RN s and PSW s of 0.008, implying an increase in PSW productivity of 0.008% for every 1% increase in RN hours. These results are supported by qualitative work performed by members of the team (see Kelly 2015). PSWs feel their role is very similar, overlapping, or even identical to that of RPNs The PSWs and key informants do not make comparisons of this nature with RNs

30 Predicting the Number of FTE PSW s in Ontario, F T E ' s R e q u i r e d FTE PSW's Required (Estimated) (Silo) FTE PSW's Required (Actual) FTE PSW's Required (Estimated) (Concert) (RN's + RPN) 0 Year Source: Residential Care Facilities Survey for Ontario ( )

31 31 Results of Forecasts Gap (Surplus) of Workers by LHIN

32 Results of Forecasts Gap of PSW Labour Force Central Central East Central West Champlain Erie St. Clair Hamilton Niagara Haldimand Brant Mississauga Halton North East North Simcoe Muskoka North West South East South West Toronto Central Waterloo Wellington

33 Results of Forecasts Gap of RN Labour Force Central Central East 0 Central West Champlain Erie St. Clair Hamilton Niagara Haldimand Brant Mississauga Halton North East North Simcoe Muskoka North West South East South West Toronto Central Waterloo Wellington

34 Results of Forecasts Gap of RPN Labour Force Central Central East 200 Central West Champlain 0 Erie St. Clair Hamilton Niagara Haldimand Brant Mississauga Halton North East North Simcoe Muskoka North West South East -600 South West Toronto Central -800 Waterloo Wellington

35 Policy Scenarios 35

36 Policy Scenario 1: Initial Gap 36 In our base-case scenario, we assume supply is equal to demand in the base period and do not allow for unmet need. This assumption is likely to be false in Ontario, as it does not take into account clients in the community or hospital waiting for an LTC bed According to a report by the MOHLTC, there was an average of 25,322 clients in the Community and in the Hospital on the waitlist for LTC in the 2009/10 fiscal year (MOHLTC, 2012). Unfortunately, this was not broken down by LHIN, so we approximated how many individuals per LHIN were on the waitlist this year by the number of facilities in each of the LHINs.

37 Policy Scenario 2: Diversion of home-care clients on LTC Wait-list to Community 37 Williams et al conducted 11 Balance of Care (BOC) projects across 10 of the 14 LHINs in Ontario They found that individuals on the LTC wait-list could be diverted safely and cost-effectively cared for in the family home or in supported housing. This generally ranged from 14-50% We simulate from the lower and upper ends to explore the effects on the estimated gap in LTC.

38 Policy Scenario 3: Level of staffing and Quality of Care 38 One issue related to LTC Homes in Ontario is the level of staffing used by homes. Data from RCFS suggests that residents were receiving 1.94 PSW hours, 0.91 nurse hours (RN and RPN), and 0.22 Therapist Hours per resident day in 2009, totaling 3 hours of direct care per resident day. These figures are below the recommended 2.5, 1, and 0.5 hours for PSWs, Nurses, and Other Staff recommended by Shirlee Sharkey s 2008 report on LTC Homes in Ontario, which equates to 4 hours of care per resident day. To date, these staffing levels have not been implemented (OANHSS, 2015). OANHSS s 2015 report suggests implementing these ratios to improve the quality of care in LTC homes in Ontario. We simulate the effects of these ratios on the projected gap of workers.

39 Summary of Forecasts: Gap of FTE Workers in LTC by 2025 in Ontario 39 PSWs RNs RPNs Scenario Gap % Change Gap % Change Gap % Change Base Case - Silo Base Case - Multiprofessional % Policy Scenario Unmet Need (50% cared for in HCC) % % % Policy Scenario Unmet Need (14% cared for in HCC) % % % Policy Scenario 3 - Staffing Ratios % % %

40 Limitations 40 Demand Side data Residential Care Facilities Survey (RCFS) Quality of care measures not available. This is not a limitation of the methodology, but rather, the data used in the study. Although we used a volume based measure of care (patient days of care), we could have explored quality outcomes as well. Such data was not available to use at the time of the study. Future work may look into exploring this should such data become available Only reports total direct care hours (including casuals). Future work may explore variation in hours worked across class of works (Full-time, Part-time Casual, etc.) Supply Side Data PSW Registry PSW labour force is unregulated, so we don t have a robust, representative dataset. Is only currently mandatory for PSWs working for publicly funded employers in the home and community sector, so may be unrepresentative (selection bias). In contrast to nursing registry, PSW registry is (at least for now) cross-sectional so that we cannot get a sense of dynamics that are driven by changes in the age distribution. Had to source data from other sources (accreditation websites, survey research, Ontario Ministry of Training, Colleges, and Universities, etc.) to obtain relevant data for forecasting model. Inflows from the Home and Community Care and Hospital sectors was assumed to be zero, since data on these transitions do not currently exist Future work will allow for workers to inflow from other sectors (see Berta et al 2015).

41 Value of Forecasting 41 Having the appropriate level of staffing available in LTC will become essential in the upcoming years as proportion of population aged 65+ continues to grow Knowing how may RNs, RPNs, and PSWs will be required, as well as the projected supply, will assist us as we plan for this

42 Conclusions 42 Health Human Resource planning is a complex science requiring detailed data Requires longitudinal data to get a sense of dynamics of workers over time and across regions Need to model workforces simultaneously to take into account interactive effects between workforces Vital for LTC with the expected increase in demand for these services in the upcoming years.

43 Next Steps 43 Funding leveraged through CIHR to expand model to Home and Community care sector Survey data is currently being collected from PSWs working in LTC and HCC With the data collected, we will be able to examine the factors that affect hours of work and influence intentions to leave the labour force, as well as likelihood of working in a particular sector The results of the analysis will be used as inputs to extend the current model to other sectors

44 References 44 Alameddine M, Laporte A, Baumann A, O'Brien-Pallas L, Mildon B, Deber R (2006). 'Stickiness' and 'inflow' as proxy measures of the relative attractiveness of various sub-sectors of nursing employment. Social Science & Medicine 63: Birch, S. et al. (2007). Human resources planning and the production of health: A needs-based analytical framework. Canadian Public Policy 33, supplement. Canada. Office of the Auditor General of Ontario Annual Report of the Office of the Auditor General of Ontario Web. April 7 th, CNA (2009). Tested Solutions for Eliminating Canada's Registered Nurse Shortage, report prepared by Tomblin Murphy, G. et al., Canadian Nurses Association, Ottawa Hamilton, Lynelle. PSW Educational Program Accreditation. Web. April 7 th, < HWA (2012). National Health Workforce Productivity Modelling - Final Report, Health Workforce Australia, Adelaide. Lum, Janet. (2013) Ontario PSW Registry Data Analysis Results. Canadian Research Network for Care in the Community. April 7 th, < Ministry of Health and Long-Term Care (2012). Long-term Care in Ontario: Sector Overview. Health Analytics Branch. Murphy et al (2012). Eliminating the shortage of registered nurses in Canada: An exercise in applied needs-based planning. Health Policy 105: OANHSS (2015). The Need Is Now: Addressing Understaffing in Long Term Care. Singh, D. et al. (2010), Ontario Population Needs-Based Physician Simulation Model, Ministry of Health and Long-Term Care and the Ontario Medical Association, Toronto. Wodchis et al (2015). Factors Associated with Personal Support Worker Turnover in Ontario LTC Homes. Accessed online August 19,

45 Extra Slides 45

46 Base Case Gap of PSW s 46 PSW's (Concert) FTE's by LHIN (Predicted) Central Central East Central West Champlain Erie St. Clair Hamilton Niagara Haldimand Brant Mississauga Halton North East North Simcoe Muskoka North West South East South West Toronto Central Waterloo Wellington

47 Base Case Gap of RN s 47 RN s FTE's by LHIN (Predicted) Central Central East Central West Champlain Erie St. Clair Hamilton Niagara Haldimand Brant Mississauga Halton North East North Simcoe Muskoka North West South East South West Toronto Central Waterloo Wellington

48 Base Case Gap of RPN s 48 RPN's FTE's by LHIN (Predicted) Central Central East Central West Champlain Erie St. Clair Hamilton Niagara Haldimand Brant Mississauga Halton North East North Simcoe Muskoka North West South East South West Toronto Central Waterloo Wellington

49 50% cared for in HCC PSW s 49 PSW's (Concert) FTE's by LHIN (Predicted) Central Central East Central West Champlain Erie St. Clair Hamilton Niagara Haldimand Brant Mississauga Halton North East North Simcoe Muskoka North West South East South West Toronto Central Waterloo Wellington

50 50% cared for in HCC RN s 50 RN's FTE's by LHIN (Predicted) Central Central East Central West Champlain Erie St. Clair Hamilton Niagara Haldimand Brant Mississauga Halton North East North Simcoe Muskoka North West South East South West Toronto Central Waterloo Wellington

51 50% cared for in HCC RPN s 51 RPN's FTE's by LHIN (Predicted) Central Central East Central West Champlain Erie St. Clair Hamilton Niagara Haldimand Brant Mississauga Halton North East North Simcoe Muskoka North West South East South West Toronto Central Waterloo Wellington

52 14% cared for in HCC PSW s 52 PSW's (Concert) FTE's by LHIN (Predicted) Central Central East Central West Champlain Erie St. Clair Hamilton Niagara Haldimand Brant Mississauga Halton North East North Simcoe Muskoka North West South East South West Toronto Central Waterloo Wellington

53 14% cared for in HCC RN s 53 RN's FTE's by LHIN (Predicted) Central Central East Central West Champlain Erie St. Clair Hamilton Niagara Haldimand Brant Mississauga Halton North East North Simcoe Muskoka North West South East South West Toronto Central Waterloo Wellington

54 14% cared for in HCC RPN s 54 RPN's FTE's by LHIN (Predicted) Central Central East Central West Champlain Erie St. Clair Hamilton Niagara Haldimand Brant Mississauga Halton North East North Simcoe Muskoka North West South East South West Toronto Central Waterloo Wellington

55 Wait List + Staffing Ratios PSW s 55 PSW's (Concert) FTE's by LHIN (Predicted) Central Central East Central West Champlain Erie St. Clair Hamilton Niagara Haldimand Brant Mississauga Halton North East North Simcoe Muskoka North West South East South West Toronto Central Waterloo Wellington

56 Wait List + Staffing Ratios RN s 56 RN's FTE's by LHIN (Predicted) Central Central East Central West Champlain Erie St. Clair Hamilton Niagara Haldimand Brant Mississauga Halton North East North Simcoe Muskoka North West South East South West Toronto Central Waterloo Wellington

57 Wait List + Staffing Ratios RPN s 57 RPN's FTE's by LHIN (Predicted) Central Central East Central West Champlain Erie St. Clair Hamilton Niagara Haldimand Brant Mississauga Halton North East North Simcoe Muskoka North West South East South West Toronto Central Waterloo Wellington