Modeling the Role of PSWs in the Production of Care in the Ontario Nursing Home Sector (March 2012)

E-mail Print
Title: Modeling the Role of Personal Support Workers in the Production of Care in the Ontario Nursing Home Sector
Principal ApplicantTitle(s)Institutional AffiliationFunding AmountAnticipated CIHR Submission Date
Dr. Audrey Laporte Associate Professor Health Economics, Dept. of Health Policy, Management and Evaluation, Faculty of Medicine at University of Toronto $10,000 March 2012
Co-Applicant(s)Titles(s)Institutional Affiliation(s)
Dr. Raisa Deber Professor Dept. of Health Policy, Management and Evaluation at University of Toronto
Dr. Andrea Baumann

Associate Vice-President,

Scientific Director

Faculty of Health Sciences, Global Health,

Nursing Health Services Research Unit at McMaster University

Dr. Brenda Gamble Assistant Professor Faculty of Health Sciences, University of Ontario Institute of Technology
Dr. Whitney Berta

Assistant Professor,


Health Services Organization and Management,

MSc/PhD Health Services Research Program, Dept. of Health Policy, Management and Evaluation at University of Toronto


Description of Proposed Project for Submission to CIHR


To investigate the mix of health human resources being used in the production of nursing home care, with a focus on the extent to which personal support workers (PSWs) are used as substitutes or complements to other types of care providers (e.g. RNs, LPNs, therapists) in that setting over time. We aim to investigate whether and in what manner the use of PSWs varies by: type of nursing home (e.g. for-profit, not-for-profit/charitable, public), rural location, and patient mix and whether these patterns have changed over time. This longitudinal view will also allow us to examine the impact of two policy initiatives that may have affected the use of PSWs in nursing homes: the requirement from 2005 that RNs have a bachelor degree and the recently announced Ontario PSW registry (effective 2012).


PSWs represent a significant component of the health care labour force and have tended to be concentrated in the nursing home and home care sectors. PSWs also tend to work in teams with RNs and LPNs. While there is little detailed empirical data which fully characterizes the PSW labour force, there are indications that these workers are often drawn either from the ranks of recent immigrants who may or may not have had some health professional training in their country of origin or more generally from the pool of lesser educated female segment of the labour market (Lum et al, 2010). Unionization amongst these workers is relatively low compared to professionalized workers in health care. It is estimated that there are roughly 90,000 personal support workers in Ontario, of whom 57,000 work in long term care (OMOHLTC web site). As nursing wages have increased, in part to deal with the need to retain these skilled workers, there have been incentives to substitute lower cost workers. Nursing homes have been characterized by a shift from RNs towards LPNs and PSWs.

Another policy initiative which we plan to monitor is the implications of the new registry for PSWs which Ontario has planned to implement by the summer of 2012; this registry will collect and make publicly available "contact information, current employment, educational background and years of experience"(Ontario Newsroom 2011). An aim of this initiative in part appears to be increased transparency, so that patients and potential employers are able to better match PSW training and experience with health care service needs. In aggregate, the ability to distinguish higher skill from lower skill workers raises the possibility that higher skill workers would then expect to earn higher wages commensurate with their added skill since they too will observe where they are in the skill distribution of PSWs in Ontario. With the seed grant, we also hope to explore whether we will be able to access this data to see whether there is an association between skill levels, settings where PSWs work, PSW pay levels, and the mix of care givers delivering resident care. Since this policy has not yet been implemented, we have designed the study so that the potential registry is not an essential component of the research.


We will use staffing data (FTEs) for nursing homes derived from the Residential care Facility Survey (RCFS), an annual census of all Canadian nursing homes(with data available for analysis from 1994) of all nursing homes and homes for the aged that receive funding support from their provincial governments. The survey is mandatory and includes detailed information about the FTEs of PSWs, RNs, LPNs, therapists, administrative, dietary and maintenance staff used. Information about the number of patients, care days, age-sex morbidity mix, and mortality rates of the patients in each facility as well as its location and profit-status are collected. Facilities must also report how much money they spent on each category of worker and how much they spent on materials and supplies including medications. The RCFS is accessible through the Statistics Canada Data Centre located at Robarts' library and these data have been used extensively by the research team. We will also use the seed money to explore the potential for extending the analysis nationally, including a focus group component.


We will estimate a production function on longitudinal data (1994 onward) in which morbidity adjusted patient days are the output of interest and the inputs will include the FTEs of each of PSWs and other worker types, number of beds (to account for nursing home size) and material and drug inputs. Age-sex distribution of the patient population, profit-status, chain ownership, and rural location will be included as control variables. In economics, the production function allows us to model the relationship between various inputs and our output of interest and allows us to estimate the impact of changes in particular inputs (greater reliance on PSWs for example) on the number of morbidity adjusted patient days of care produced. It also allows us to estimate the degree of complementarity and substitutability of one input for another, i.e. among PSWs, LPNs, and RNs over time and whether these relationships have changed markedly in response to the two policy changes.

Policy Implications

We will investigate how the requirement since 2005 that RNs have a university degree has affected the number and mix of PSWs employed in relation to RNs and LPNs. Here we wish to assess whether a policy (higher educational requirement) that would tend to make RN labour more expensive shifted care provision toward greater reliance on PSWs. Based on the most recent years of available data (i.e. 2012 and beyond) we will monitor the impact of the registration requirement for PSWs on their use compared to other care providers (complementarity and substitutability) and whether the amount of money spent on PSWs changes significantly (i.e. has there been a shift toward the lower cost or higher cost PSWs) adjusting for FTEs and holding constant the age-sex morbidity distribution of the patient population within each nursing home.