ALLOCATION MODEL INFORMING THE DISTRIBUTION OF AGING AT HOME FUNDS AT THE CENTRAL EAST LOCAL HEALTH INTEGRATION NETWORK

Transcription

1 POPULATION BASED ALLOCATION MODEL INFORMING THE DISTRIBUTION OF AGING AT HOME FUNDS AT THE CENTRAL EAST LOCAL HEALTH INTEGRATION NETWORK May 27, 2009 Prepared by the Centre for Research in Healthcare Engineering University of Toronto Michael W. Carter, PhD Ali Vahit Esensoy, MASc Agnita Pal, BASc Abstract

2 The Central East Local Health Integration Network (CE LHIN) identified the need for a quantitative methodology to assist them in the allocation of Aging at Home Strategy funds to address the service gaps in supportive housing (SH) and community support services (CSS) within their jurisdiction. The objective of this project was to develop a population-based allocation model that quantified the current demand and service gaps for SH and CSS in the CE LHIN area. Client groups were defined for the six priority services: SH, integrated adult day, adult day for cognitively impaired, respite, meals delivery, and transportation services. Available data was used to quantify the population rate that fit the client groups in different age-gender categories. Census data was used to project the CE LHIN population in client groups. The current average level of service was used to determine population-based demand. Population-based demand was compared with current service utilization data to quantify service gaps for each of the priority services. The model found unmet demand in all priority service areas except transportation. Transportation demand was expected to be an underestimate due to the limited data on current service utilization. The model found relatively high numbers of clients requiring SH and adult day services for cognitively impaired. Populationbased demand maps were developed to show the distribution of demand across the CE LHIN. Provider addresses were plotted on the map to show the location of current services. Future studies are required to gather information on provider catchment or client location in order to show the current capacity and service gaps on the CE LHIN maps. Once it becomes possible for both demand and supply data to be represented in a CE LHIN map, it would be possible to consider factors that are currently not quantified in the model, such as rurality, availability of health services (e.g. waitlists) and other public services (e.g. transportation), and chronic disease prevalence. In addition to facilitating appropriate investments in the right priority services, this would enhance local planning with providers on the optimum location for services.

3 About CRHE An initiative of the Department of Mechanical and Industrial Engineering at the University of Toronto, the CRHE is a response to the immediate and compelling desire for efficiency and quality improvements in the Canadian health care system. The Centre will provide both theoretical and practical advice and support for many of today's most pressing problems. CRHE research is focused on the application of Industrial/Systems Engineering techniques in relation to demand and capacity modelling and resource allocation issues in the health care industry. Our goals include creating quantitative decision support tools to help policy makers and industry leaders make better informed decisions. AAH TEAM MEMBERS Michael W. Carter, PhD Michael Carter is a professor in the Department of Mechanical and Industrial Engineering at the University of Toronto and is the Director of the Centre for Research in Healthcare Engineering (CRHE). Since 1989, his research focus has been in the area of health care resource modeling with a variety of projects in hospitals, homecare, rehab, long term care, medical labs and mental health institutions. He has supervised more than 160 engineering students in over 100 projects with healthcare institutions. He is an Adjunct Scientist with the Institute for Clinical Evaluative Sciences in Toronto. Ali Vahit Esensoy, MASc Ali Vahit is pursuing his PhD at CRHE. Currently, he is focusing on the applications of simulation modeling on longterm capacity planning on LHIN and Provincial levels as well as on the development of a population based allocation model for the AAH funds for the Central East LHIN. His past projects include Urgent Care Centre and St Mike s Surge Capacity Protocol Evaluation projects, hospital capacity forecasting for additional surgical volumes and the development of a benefits measurement framework for chemotherapy computerised physician order entry applications. Agnita Pal, BASc Agnita is a graduate of Industrial Engineering at University of Toronto. She has valuable experience in healthcare policy and engineering through her years in the industry at the Ontario Ministry of Health and Long-Term Care and Ministry of Community and Social Services. She is currently completing her Master s of Engineering degree at CRHE. Centre for Research in Healthcare Engineering Mechanical & Industrial Engineering, University of Toronto 5 King s College Road, Toronto, ON, M5S 3G8 Phone: (416) Fax: (416) Web: For inquiries regarding this document, please contact Ali Vahit Esensoy at esensoy@mie.utoronto.ca iii

4 TABLE OF CONTENTS 1. INTRODUCTION AGING AT HOME STRATEGY ALLOCATION DECISION REQUIREMENTS PROJECT OBJECTIVE BACKGROUND COMMUNITY SUPPORT SERVICES (CSS) AND SUPPORTIVE HOUSING (SH) CE LHIN POPULATION PROFILE CE LHIN SERVICES LITERATURE REVIEW OPERATIONS RESEARCH MODELLING TECHNIQUES IN HEALTHCARE COMMUNITY SERVICE & ELDERLY CARE MODELS CSS AND SH CLIENT CHARACTERISTICS METHODOLOGY OVERVIEW SOURCES OF DATA STEPS ASSUMPTIONS AND CONSTRAINTS ESTIMATING CURRENT DEMAND FRAMEWORK FOR DEFINITION OF CLIENT GROUPS CLIENT GROUPS BY PRIORITY SERVICES ANALYSIS OF PALS DATA ANALYSIS OF CENSUS DATA POPULATION DEMAND MAPS CURRENT SERVICE CAPACITY CURRENT SERVICE ESTIMATES CONCLUSIONS ABOUT SERVICE GAPS AREAS OF FUTURE STUDIES REFERENCES APPENDIX A: CENSUS DISTRIBUTION FIGURES APPENDIX B: DEMAND MAPS APPENDIX C: SAMPLE PROVIDER SURVEYS SAMPLE SURVEY FOR SUPPORTIVE HOUSING SAMPLE SURVEY FOR INTEGRATED ADULT DAY SERVICES APPENDIX D: EXPERT SERVICE PROVIDER PANEL MEMBERS...58 iv

5 List of Figures Figure 1: Population Density and % Over 65 by Sub-LHIN Planning Areas...10 Figure 2: Disability Rates in Elderly Ontarians, 2001 and Figure 3: Disability Rates by Type among Eldery Ontarians, Figure 4: Proportion of CSS and SH funding for Priority Services...13 Figure 5: Distribution of PALS Cases in Target Age Categories...32 Figure 6: Distribution of PALS Cases in SH Client Groups...33 Figure 7: Distribution of PALS Cases in Integrated Adult Day and Respite Services Client Groups...34 Figure 8: Distribution of PALS Cases in Adult Day Service for Cognitively Impaired Client Groups...35 Figure 9: Distribution of PALS Cases in Meals Delivery Client Groups...36 Figure 10: Distribution of PALS Cases in Transportation Client Groups...37 Figure 11: Distribution of CE LHIN Population Demand for Supportive Housing...38 Figure 12: Population Demand Map for Transportation...39 Figure 13: Distribution of CE LHIN Population Demand for Integrated Adult Day Service...46 Figure 14: Distribution of CE LHIN Population Demand for Adult Day Services for Cognitively Impaired...46 Figure 15: Distribution of CE LHIN Population Demand for Respite...46 Figure 16: Distribution of CE LHIN Population Demand for Meals Delivery...47 Figure 17: Distribution of CE LHIN Population Demand for Transportation...47 Figure 18: Demand Map for Supportive Housing...48 Figure 19: Demand Map for Adult Day Services...49 Figure 20: Demand Map for Respite...50 Figure 21: Demand Map for Meals Delivery Services...51 List of Tables Table 1: Community Support and Supportive Housing Provision within the CE LHIN...12 Table 2: Framework for Client Groups...24 Table 3: Client Groups for Supportive Housing...26 Table 4: Client Groups for Adult Day Services - Integrated...28 Table 5: Client Groups for Adult Day Services Cognitively Impaired...28 Table 6: Client Group for Respite...29 Table 7: Client Groups for Meals Delivery...29 Table 8: Client Groups for Transportation...30 Table 9: Population Rates for Supportive Housing...33 Table 10: Population Rates for Integrated Adult Day and Respite Services...34 Table 11: Population Rates for Adult Day Services for Cognitively Impaired...35 Table 12: Population Rates for Meals Delivery...36 Table 13: Population Rates for Transportation...37 Table 14: Population Based SH and CSS Demand in CE LHIN...38 Table 15: Current Capacity of Priority Services in the CE LHIN...40 Table 16: Priority Service Gaps in the CE LHIN...42 v

6 1. Introduction The Central East Local Health Integration Network (CE LHIN) identified the need for a quantitative methodology to assist them in allocation of Aging at Home (AAH) Strategy funds to address the service gaps in supportive housing (SH) and community support services (CSS) within their jurisdiction Aging At Home Strategy Developed by the Ministry of Health and Long-Term Care (MOHLTC), the AAH Strategy represents a $700 million provincial investment over three years enabling seniors to live independently in the comfort and dignity of their own homes [9]. The strategy aims to increase the overall supply of seniors services while relieving pressure on hospitals and long-term care homes. This would be achieved by meeting the needs of seniors and their caregivers with the appropriate supports, as well as avoiding crisis through proactive wellness approaches. In the public overview presentation of the AAH Strategy [9], the MOHLTC identified the following four goals for the AAH strategy: 1. Ensure that seniors homes support them - this involves increasing the residential options that support seniors, including improving safety and preventing injury in seniors home environments, and increasing the availability of assisted living and supportive housing services; 2. Supportive social environments this includes creating environments such as adult day centres and caregiver respite and relief services to decrease isolation among seniors and their caregivers; 3. Senior-centered care that is easy to access this involves providing easy access to a continuum of services from home care to supportive housing, community support services, long-term care homes and end-of-life care. This includes improved case coordination and transportation services to facilitate access; 4. Identify innovative solutions to keep seniors healthy this includes providing 20% or more of the funding for innovative approaches, such as partnerships with non-traditional providers that recognize informal services. Each LHIN was asked to manage the local processes for determining high priority areas and allocate funding based on these assessments. For the CE LHIN, the AAH Strategy represents a significant investment into the health and wellness of seniors. This initiative allows for improved use of available resources, enhanced access and integration of services, and improved client outcomes in the supportive housing and community support service settings Allocation Decision Requirements The CE LHIN required an allocation model for the distribution of the AAH strategy funds to meet the senior population s needs within its geographic area. As a result, this project was commissioned in early 2008 to 6

7 develop a population-based allocation model (PBAM). Development of the allocation model involved investigation of the following: 1. Definition of client profile for a focused set of services based on consultation with the LHIN and data available from existing sources and service providers; 2. Assessment of the CE LHIN population and health system characteristics to quantify demand for the services; 3. Measurement of current service capacity based on data from available sources; and 4. Determination of the gaps between population needs and service capacity to inform decisions on the increase to services required to meet the needs of the target population Project Objective The objective of this project was to develop a population-based allocation model (PBAM) that quantified the current demand and service gaps for supportive housing and community support services in the CE LHIN area. This information could then be used to support the allocation of AAH funds for target populations. 7

8 2. Background The AAH Strategy is expected to expand existing community support, home care and assisted living/supportive housing services. This is intended to serve more seniors with chronic conditions to help them remain at home, and avoid unnecessary hospital stays (e.g. waiting for community supports to be available), and inappropriate admissions to long-term care homes Community Support Services (CSS) and Supportive Housing (SH) Based on census 2006, Statistics Canada reported that the 65-and-over population (i.e seniors) was at a record high of 13.7% of the total population of Canada. In Ontario, seniors can access a variety of services that are provided in their home or in their community. Home and community support services include visiting health professional services, personal care and support, homemaking, and community support services. Such services are categorized and defined in the Planning, Funding & Accountability Manual (PFA Manual) developed by the MOHLTC. This project focused on supportive housing and community support services, which includes a wide variety of services from meals delivery, transportation, adult day services, caregiver relief or respite, and recreational services. The following sections briefly describe some of the key community services for the elderly Supportive Housing Services for seniors can be provided in residential care settings, such as long-term care homes, retirement homes or supportive housing units. Supportive housing accommodations are for seniors who require varied levels of personal care and supports to live independently. Such individuals require services to be accessible for up to 24 hours per day. Also known as assisted living services, supports provided in such settings include homemaking, personal care, and attendant services. Service units are measured in resident days. Accommodations usually consist of rental units within apartment buildings that are owned and operated by municipal governments or nonprofit groups [11]. Based on interviews with various directors from Community Supportive Housing units in the CE LHIN, it was apparent that these social housing units are in short supply, and have a growing demand. The level of staff supports and services varies to meet the needs of residents. The variety of services typically available in the building could include on-site personal care and support, such as routine hygiene, dressing and washing, daily visits or phone check-ins (CRHE, interviews, 2008). Although specialized supportive housing is also available for specific client groups, such as individuals with physical disabilities and acquired brain injury, the focus of this project was supportive housing available for the elderly. 8

9 Adult Day Services This support service provides supervised individual programming in a group setting for adults to assist them in achieving and maintaining their maximum level of functioning. Components of the service include planned social, recreational, and physical activities, as well as meals, personal support/attendant care and minor health care service (e.g. monitoring medication) [11]. Based on interviews with service providers this service is often considered to be a relief to live-in caregivers. The target population includes frail and elderly as well as individuals with Alzheimer Disease and other progressive cognitive disorders or dementias. Service providers acknowledged, however, that many of the group settings are not conducive to individuals with severe cognitive and mobility limitations. There are a few providers that specialize in delivering adult day care to specific complex client groups, such as those with severe cognitive impairment, and acquired brain injury (CRHE, discussions with CE LHIN service provider panel, February, 2009). Unit of service is measured in attendance days Respite These services provide in-home supports including homemaking, personal support, independence training, and home maintenance to eligible clients with live-in caregivers. Such services are intended to provide short and long-term relief to caregivers. Unit of service is measured by the number of hours of respite care delivered [11] Meals Delivery Services include delivering meals to seniors at their residence to meet their nutritional requirements. Such services are also known as meals-on-wheels. Unit of service is primarily measured by the number of meals delivered [11] Transportation This support service provides seniors escorted transportation to medical appointments, shopping and various social activities and programs. The target population includes seniors that are unable to access transportation due to a lack of public transportation or because they have short or long-term physical limitations or cognitive impairments. The key unit of service is the number of trips [11] CE LHIN Population Profile The Central East LHIN has several characteristics that make it a unique and diverse region. Home to almost 1.5 million individuals, the CE LHIN is the second-largest by population in Ontario, and is experiencing an above average growth rate. Figure 1, below, illustrates the population density of the CE LHIN by sub-lhin planning areas. The first four planning areas are the most densely populated areas in the LHIN. When considering the large geographical region encompassed by the LHIN, its services must be planned to meet the diverse needs of its urban population in the high density areas as well its rural residents. Further analysis of its seniors population shows that nearly 80% of the CE LHIN s seniors reside in the following five planning areas: Scarborough Cliffs - Scarborough Centre; 9

10 Scarborough Agincourt-Rouge; Durham West; Durham East; and Peterborough City and County. Figure 1 also illustrates that while the population may be more geographically dispersed in some of the sub-lhin areas, seniors, aged 65 and over, form a higher proportion of their population when compared to provincial or CE LHIN average rates of 12%. When allocating resources to promote utilization, the location and mobility of services form key considerations that enable fair and equitable access for both its rural and urban residents. As a result, it was important to map out population demand and current capacity to identify service gaps by geographic areas. Investments in supportive housing and adult day services could target meeting the needs of individuals in densely populated areas while investments in more mobile services such as meals delivery, respite, and transportation could target meeting the needs of its rural residents. Figure 1: Population Density and % Over 65 by Sub-LHIN Planning Areas Population Density 5,000 4, , ,000 14% 12% 1,000 - Scarb. Cliffs-Centre Scarb. Agincourt-Rouge Durham W. 8% 558 Durham E. 11% 276 Durham N./Cen. 14% 18% 18% 19% 24% Peterborough rth.-havelock Sub-LHIN Planning Areas Source: 2001 Census Kawartha Halib. Highlands 30% 24% 18% 12% 6% 0% % Sub-LHIN Population Over Living arrangements The type and intensity of community support services are often related to living arrangements. The majority of the elderly above 65 live in family households, and living alone is the second most common living arrangement. On average in Ontario, over 72% of seniors live in family households while 25.7% of seniors over 65 live alone (census, 2006). Investments in community support services must, therefore, not only target meeting the care needs of the seniors, but also provide relief to the caregivers living with high-needs seniors in the family households with respite and adult day services. 10

11 Disability rates The need for community support services is assessed by seniors ability to perform activities of daily living. Data on individuals that experience difficulties in performing such activities is available from the StatsCan Participation and Activity Limitation Survey (PALS). This survey targets individuals whose everyday activities are limited because of a condition or health problem. Due to the limited sample size of individuals surveyed in Ontario, provincial rates were used to approximate for disability rates in CE LHIN. Figure 2 shows the disability rates in two elderly age groups, as measured by the PALS survey data in 2001 and Among all Ontarians, aged 65 and over, the disability rates have increased from 44.9% in 2001 to 47.2% in 2006 [12]. It should be noted that disability rate for individuals aged 75 and over is substantially higher than those aged between 65 and 74. As a result, one might expect an increase in the use of support services by individuals, aged 75 and over, that are living independently (i.e. not in long-term care homes) in the community. Figure 2: Disability Rates in Elderly Ontarians, 2001 and 2006 The survey also provides data on the type and severity of disability. Types that may be of interest to this study are: Mobility limitations that may be used as an indicator of the need for transportation services; and Agility and memory limitations that may be used to assess needs at the population level for other community services such as adult day and supportive housing programs. Figure 3 shows provincial rates for various types of disabilities among the elderly population. As expected, the disability rates for all types are consistently higher for the 75 and over age-group compared to the age group. The figure also indicates a higher prevalence of mobility and agility limitations for the elderly compared to memory. Seniors with higher mobility and agility limitations would require more intensive services than the social/recreational supports offered in group settings in day service programs. 11

12 Figure 3: Disability Rates by Type among Eldery Ontarians, CE LHIN Services There are currently a variety of community support and supportive housing services offered throughout the CE LHIN. The cost of services is highly variable, ranging from $6.3 million to less than $5000, as seen on table 1 below. Table 1: Community Support and Supportive Housing Provision within the CE LHIN Service Name Total Expenses Total Providers Supportive Housing $6,298, Respite $4,766,028 3 Meals Delivery $3,871, Adult Day Services - Integrated $3,095, Supportive Housing - Disabled $2,721,703 3 Caregiver Support $2,424, Transportation $2,332, Pyschogeriatric Consulting Services $1,285,487 3 Adult Day Services - Cognitive $912,447 3 Home Help/Homemaking $762,158 3 Acquired Brain Injury Outreach $742,939 3 Friendly Visiting $738, Blind and Visually Impaired $510,600 2 Acquired Hearing $466,373 1 Client Intervention $320,043 3 Security Checks $313,695 9 Home Maintenance and Repair $219,017 8 Emergency Response Systems $189,956 3 Pain and Symptom Management $173,252 2 Public Education Coordinator $158,

13 Service Name Total Expenses Total Providers Foot Care $139,965 3 Extended Respite Service $130,923 1 Social Work $101,394 1 Aboriginal Service $64,245 1 Palliative Care $3, CE LHIN Priority Services Of these various service categories listed in the table above, the CE LHIN identified a few priority areas for future investments. As a result, the model focused on the following six services: 1. Supportive Housing; 2. Adult Day Services - Integrated; 3. Adult Day Services for Cognitively Impaired; 4. Respite; 5. Meals delivery; and 6. Transportation. As shown in Figure 4 below, these priority services account for about 65% of all the CE LHIN program funding. Figure 4: Proportion of CSS and SH funding for Priority Services 13

14 3. Literature Review 3.1. Operations Research Modelling Techniques in Healthcare Operations Research (OR) modelling approaches have been applied to health care over the past 40 years. Different approaches have been used to model patient flows for health service planning. These include Markov, semi-markov chain models, queuing models, and simulations. When compared with other modelling approaches, simulation is the preferred technique due to its ability to effectively deal with the complexity of a health system [3]. Simulation provides a powerful tool for modelling healthcare systems and testing the impact of policy on efficiency, effectiveness and equity without the risks and costs associated with a real-world implementation of such policy [5]. Discrete event simulations (DES) is one of the most widely used approaches in healthcare [1]. In their survey of literature, focusing on late 1970s to 1990s, Jun et al cited many studies that use DES to understand the inputs and outputs of the healthcare system. While DES is extensively used to highlight efficiency improvements in single healthcare clinics, very few articles used this technique to study complex, integrated, multi-facility systems [6]. For a complex, multi-agency community service system, DES approach may prove to be too expensive with its extensive data collection requirements. System Dynamics (SD) is used to model continuous processes and systems with explicit feedback loops.. Patients or clients are treated fluids that flow through the stocks in the system at rates that are controlled by valves. In SD, structure and relationships between elements of the system determines the behaviour of the system and provides an understanding of the complexity of the system as a whole. SD models can be used to model complex systems with larger populations and time periods without large quantities of high quality data. Unlike DES, however, SD models are deterministic and patients are not modelled individually but treated as a mass of objects flowing through the system [1]. With its strategic orientation, SD, however, is well suited to studies that involve flows of people and resources in a region. In the review of SD application in healthcare in Europe, Dangerfield highlights that studies which inherently involve patient flows and money flows at a strategic level of aggregation are obvious candidates for system dynamics framework to the analysis [2] Community Service & Elderly Care Models With growing proportion of elderly population in many (developed) countries, healthcare system planning must include community services to avoid policy decisions that lead to improvements to the acute care or hospital systems at the expense of community care [8]. Although a number of OR modelling techniques have been applied to community care, many of the models focus on the interface between acute and community care for elderly. Gray et al describes the development of an SD model for interactions between the acute system and community-based aged care services in Australia. Their plan is to construct the model at both national as well as regional levels to examine the impact of a number of policy scenarios (such as an increase in community places, etc.) over a 10-year time frame. The paper did not provide details on the system maps or the quantitative model. 14

15 A similar study in the UK aimed to investigate the care pathways of elderly post-discharge from the acute care system. This study attempted to use DES, but lack of data made it impossible to validate or run models for quantitative results. The study shifted its focus to description and mapping of services, measurement of capacity and development of a prototype model that gave illustrative results highlighting the potential value of the approach [7]. Another study in the UK used a Markov Reward model that included patient movement between hospitals and community and enabled the calculation of system wide costs. By assessing the complex relationship between hospital and community care, the aim of the model was to prevent policies that appear to generate savings in one part of the system at the expense of the other [8]. In a study for Hampshire Adult Services in UK, Desai et al used SD to model demand for elderly care that incorporated aging population projections as well as supply representing community based services and resources available to meet the increased demand. Services included community based day care, respite services, domiciliary care as well as residential care with supports and specialized nursing homes. Although SD does not allow tracking of individual characteristics, clients were grouped based on age categories, acute care referral, and assessed level of need to be modelled separately using arrays [4]. This approach could be used to model CE LHIN demand and its impact on services provided that it is possible to gather reliable data on population projections, service eligibility assessments, referral rates, and client flows through various services in the system. Prior models clearly demonstrate the complexity of interactions between community-based care services and other sectors in the healthcare system. In the CE LHIN, service accessibility and availability in other sectors, such as acute care, home-care, long-term care, and mental health can be expected to influence demand of community support services. Given the complexity of interactions, large populations, and in the absence of large quantities of quality data about the interfaces between the sectors, SD modelling is the only approach that may be considered. Based on national data (e.g. census), it may be possible to reasonably estimate population-based needs. SD models developed to date, however, require the following inputs about the service system that are not feasible within the scope of this study: Standardized eligibility considerations for access to services; Client flow model; and Defined referral rates and exit rates to/from services. In this study, alternate approaches were explored to develop a population based allocation model (PBAM) for the CE LHIN. The model included the definition of client profiles for a focused set of SH and CSS services, estimated demand for such services, and identified gaps in service resulting from comparing demand with current service capacity. The population-based demand and service gap information could be used by the CE LHIN to support investment of AAH funding CSS and SH Client Characteristics Prior studies that defined the profile of CSS and SH clients focused on the following key characteristics to categorize individuals on the long-term care waitlist into client profile groups: 15

16 Activities of daily living (ADL) impairment (e.g. eating, toileting, locomotion, and personal hygiene); Instrumental activities of daily living (IADL) impairment (e.g. managing medications, housekeeping, meal preparation, phone use, and transportation; Presence of a caregiver; and Confusion/cognitive impairment [14]. 16

17 4. Methodology 4.1. Overview A key driver for PBAM was to improve the availability of CSS and SH services based on the demand by the target population in sub-lhin areas. Data on current client characteristics and their service use was compared against population characteristics in sub-lhin areas to quantify service demand. Demand data was then compared to the current service capacity to compute service gaps in the LHIN. In order to simplify the data collection requirements, the model focused on the following six priority service areas identified by the CE LHIN: 1. Supportive Housing; 2. Adult Day Services - Integrated; 3. Adult Day Services for Cognitively Impaired; 4. Respite; 5. Meals delivery; and 6. Transportation Sources of Data The following available data sources were used for PBAM. The model outputs had to be adjusted depending on the reliability and granularity of the information available in each of the sources Provider surveys The following six surveys were developed to gather data on client characteristics and their current level of service use from CE LHIN service providers: 1. Agency and Service Profile Survey; 2. Supportive Housing Client Survey; 3. Adult Day Client (Integrated and Cognitively Impaired) Survey; 4. Respite Client Survey; 5. Meals Delivery Client Survey; and 6. Transportation Client Survey. Although the feasibility of the surveys was examined with volunteer service providers, due to time and resource constraints the surveys could not be deployed to obtain client data from the providers. An alternate framework was developed and validated with a service provider expert panel that is described in detail in section Participation and Activity Limitation Survey (PALS) 2001 Conducted by Statistics Canada, the Participation and Activity Limitation Survey (PALS) is a post-censal national survey of adults and children whose everyday activities are limited because of a condition or health 17

18 problem. PALS conceptualizes disability as activity limitations and participation restrictions associated with longterm physical or mental conditions or health-related conditions. The survey is designed to identify: Canadians with an activity limitation; the type and severity of activity limitations that they experience; the difficulties and barriers that they may face; and the degree to which persons with disabilities are able to obtain help and other supports they need at home, work, or school. The 2001 Census questionnaire included two general questions on activity limitations and a sample of of individuals who answered YES to the activity limitation questions were selected for PALS survey. The PALS sample was 43,000, consisting of approximately 35,000 adults and 8,000 children. PALS data is based on self assessment/reporting provided via telephone interviews to interviewers completing the questionnaires. The overall response rate was 82.5%. Although PALS survey for 2006 has been completed, the required data has not been released by Statistics Canada. As a result, 2001 results were used for this project Census 2006 The 2006 Census was an account of the Canadian population and its characteristics. May 16, 2006 and gathered data by geographic unit on the following areas relevant to this project: Age and gender; Households, including living arrangement; and Income/earnings. Census day was CE LHIN Provider Information 2006 One key source of provider information was the data table obtained from CE LHIN that included a comprehensive listing of the SH and CSS providers, their addresses, services, expenses and units of service from Information about CE LHIN annual average units of service by client for each of the priority service groups was obtained from this data table. Although the specific data on services was slightly dated, this table represented the most comprehensive listing available about the providers. This is because the providers are in the midst of transitioning to a new reporting system known as the Ontario Healthcare Reporting System (OHRS) that does not have complete information on the CE LHIN OHRS (MIS) Comparative Year-End Reports for Fiscal Year 2007/08 For some of the priority services, updated information was obtained from the CSS Ontario Healthcare Reporting System (OHRS) Comparative Year-End Reports from 2007/08. This report included provincial and LHIN-level summaries of expenditures, units of service and other key indicators for SH and CSS services. Since a number of SH and CSS providers have been transitioning to this reporting system over the past couple of years, the report did not include comprehensive data about the service system. Since CE LHIN data 18

19 represents a relatively small sample of providers for each priority service, provincial average level of service was used in the model to include a larger sampling of provider data. It should also be noted that service groupings have slightly altered from the previous system. Although it was possible to isolate the majority of the priority services, the report did not distinguish between the different types of adult day services. Additionally, due to the lack of specific service units, such as number of trips, the CE LHIN provider information was used for the average rate of transportation services despite its limitations Steps The methodology consists of the following steps: 1. Define client groups for priority services: The first step in constructing the model was to define profiles of typical clients accessing SH and CSS services. Due to the lack of a standardized client assessment data and map of client flows between service providers, the original plan was to use service provider surveys to define typical client profiles and corresponding service levels for SH and CSS services. Six provider surveys were developed to define client groups in each of the priority services. In order to ease the burden of data collection on providers, sampling targets for client surveys were recommended for each agency and service based on the minimum number of clients required to be a representative sample for the CE LHIN. As mentioned before, due to resource constraints the surveys could not be deployed to obtain client data from the providers. Given the unavailability of client data from service providers, a framework for defining client groups was developed based on characteristics identified in study by Williams et al [14] and the data available through the PALS 2001 survey. Such characteristics were primarily expressed as activity limitations, such as cognitive limitations. An expert panel with CE LHIN service providers described the dominant characteristics of their clients for each of the priority services using the dimensions identified in the framework. Refer to Appendix D for the list of names of panel members. 2. Quantify Prevalence of Profile Characteristics in different Age-Gender Categories: PALS data on activity limitations and help with everyday activities was mined to compute population rates in different age and gender categories with the key profile characteristics (e.g. 2% of year-old female PALS survey respondents identified with the dominant client characteristics for SH services). 3. Project CE LHIN Population in Client Groups: Census data was used to estimate the current population in each of the age and gender groups by census dissemination areas. Census data on living arrangements was used to estimate the current population living with caregivers. National rates of population with client profile characteristics in each of the age-gender groupings from PALS found in step 2 was applied to the current CE LHIN population estimates to quantify the proportion of the CE LHIN population that would belong in the client groups. 4. Define Service Demand by Client Group: Current level of service use data gathered from available data sources was used to compute annual average units of service required by clients in each of the groups. Data about current service capacity was obtained from the CE LHIN Provider Information. Updated 19

20 information on annual average units of service was obtained from the CSS OHRS Comparative Year-End Reports from 2007/ Quantify Population-Based Service Demand: Annual service units by client groups (from step 4) was multiplied by population projections in client groups (from step 3) to quantify population based service demand by census dissemination area. This geographic distribution of demand was visually represented on a CE LHIN map that clearly showed the varying density of service demand. 6. Obtain Service Capacity Data: Annual service units data by service categories were used to measure current service capacity in CE LHIN. In the absence of data about client locations or catchment area from CE LHIN, provider addresses were plotted on the map to show location of current services in the LHIN. Where provider site addresses and their units of service could be obtained, such as in supportive housing, current services capacity could be visually represented on a CE LHIN map. 7. Quantify Gaps in Each Service: Service gaps were then computed by subtracting current service supply obtained in step 6 from population-based service demand computed in step 5. The gaps were primarily computed at the overall LHIN level due to lack of specific data on catchment areas and client locations in current service information Assumptions and Constraints Overall Assumptions The specific steps in the methodology evolved as the project encountered constraints and limitations with availability of reliable and accurate data. Throughout the project, however, it was assumed that the CE LHIN and key representatives from the service provider community would be available to offer expert advice and validate assumptions to ensure accurate interpretation of available data and its constraints Client Characteristic Assumptions Building on prior work completed by Williams et al [14], the model focused on the following key characteristics to construct client profiles: Age; Gender; Activities of daily living (ADL) impairment (e.g. eating, toileting, locomotion, and personal hygiene); Instrumental activities of daily living (IADL) impairment (e.g. managing medications, housekeeping, meal preparation, phone use, and transportation); Cognitive impairment (e.g. memory); and Live-in caregiver support. This assumed that the above-mentioned dimensions would be sufficient to capture the dominant characteristics of priority service client groups. Williams et al [14] primarily focused on the clients waiting for 20

21 service in the long-term care waitlists. By provider account, such individuals represented a relatively small subset of the client group served by SH and CSS services, although the proportion was not quantified. It was also assumed that these characteristics would be sufficient to determine the level of service needed or accessed by client groups. Factors that were not being considered included other demographic variables such as income level, preferred language or other cultural characteristics, chronic disease prevalence, hospitalization rates, and long-term care waitlists. To compensate for these constraints, the model included adults aged 50 and over in both PALS and census data to consider clients that are experiencing some limitations due to chronic diseases or hospitalizations Client Profile Assumptions The original plan was to construct client profiles based on surveys of randomly selected sample of clients. Six provider surveys were developed using a survey tool to define client groups in each of the priority services. In order to ease the burden of data collection on providers, sampling targets for client surveys were recommended for each agency and service. These targets represented the minimum number of clients required to be a representative sample for the CE LHIN. Due to resource constraints, however, the surveys could not be deployed. As a result, the methodology had to be revised to rely solely on expert advice from service providers to define dominant characteristics of clients in the different priority services. The profile groups, therefore, included provider biases about client characteristics that could not be validated with data about current service system Service Level Assumptions The original plan was to also use the survey to define required level of service based on current utilization of services by the clients in the various groups. This would then be used to recommend future levels of service for estimated population in the profile group. Without the survey, however, the level of service was based on the provincial average level of service and did not include a range of service levels for the low to high needs client groups. This assumed that the average current level of service for the clients in priority services is sufficient and appropriate to meet the needs of all potential clients in the larger general population, including those that are not currently being served. The model could not distinguish where clients may be overusing certain services (perhaps due to lack of availability of the other services that are not in the priority list) or where the current level of service is unable to meet the level of client need due to capacity constraints in the current system. In the current service system, one individual may be seeking services from different providers due to provider capacity, service delivery models, etc. In the absence of unique identifiers across the service system, such an individual would be duplicated in the service provider data. As a result, current service capacity may be overstating the number of individuals that are actually served by the system. Current average units of service per client, however, would likely be an underestimate of the actual utilization of service by such an individual who is receiving services from multiple providers. When this understated average service level is applied to the potential clients in the population, they will result in lower estimates for demand in units of service. With higher estimates of individuals served in the current system and lower estimates for required service units (i.e. demand), the service 21

22 gap is likely to be an understatement of unmet needs. Without provider surveys, current waiting list information could not be incorporated into the service gap quantified by PBAM. Additional considerations such as availability and accessibility (i.e. with walkers, wheelchairs, etc) of public transportation in the sub-lhin areas that may be relevant were not incorporated in the model. As a result, demand estimates for transportation services, especially in the vast rural areas of the LHIN, where the population is not as dense, were likely to be an underestimate. This would also impact demand for other services where client mobility is an issue, such as in adult day services. Due to lack of client level data, however, such interactions between services were not captured in the PBAM PALS Data Limitations and Assumptions PALS 2006 survey has been completed but at the time of the analysis, complete 2006 responses could not be obtained from Statistics Canada. As a result, dated information from the 2001 survey was used in PBAM. PALS survey is limited to a sample of the population that has self-identified with a disability in the census. As a result, the data does not provide a complete account of the population needs. Due to the large sample size and high response rates, PALS survey is representative of a larger population with data from 28,193 participants, aged 50 and over. Due to limited sample size of the PALS participants from Ontario in the target age group, the rate of population that experience ADL/IADL/cognitive limitations was computed based on all surveyed clients regardless of their location (i.e. national ratios). It was assumed that these national rates are applicable to the CE LHIN population. Another limitation of PALS is that it is self-reported, which is not as reliable as clinician administered instruments such as Resident Assessment Instrument-Home Care (RAI-HC) that was used in Williams et al [14] to assess the level of need of clients waiting for long-term care services. For PBAM, the level of need (i.e. high/medium/low) was determined based on the level of help received from various sources of formal and informal care, such as live-in caregiver, friends, neighbours, agencies, and others. Due to unknown unmet needs, the selfreported level of service may not represent a sufficient level of need. In RAI-HC, ADL needs are assessed based on the Self-Performance Hierarchy Scale that includes questions on eating, personal hygiene, locomotion and toilet use. For PBAM, however, ADL needs assessment was limited by the few questions included in PALS about personal care and locomotion. Similarly, IADL needs were determined based on limited questions about meal preparation, housework, finances and transportation available in PALS and did not include areas such as medication management, and phone use. Additionally, PALS 2001 did not have specific questions on cognitive performance, such as, short term memory, cognitive skills for decision-making, expressive communication, etc. Instead PALS data on other limitations that included cognitive limitation was used for cognitive needs assessment Census Data Limitations Due to confidentiality reasons, individual census data included age, gender, living arrangement, and dissemination area, and did not include other characteristics, such as income level, ethnicity, etc. As a result, rate of service demand in PBAM was calculated based on age and gender groups with or without live-in caregivers. 22

23 These groups are likely to overestimate demand when applied across the population. PBAM is limited as a result of inability to include other population characteristics such as income levels, chronic diseases, ethnicity, etc. Such characteristics would be valuable to incorporate in future to determine a more comprehensive understanding of population demand Assumptions About LHIN Service Data Currently available data on CE-LHIN services are limited in both quality and granularity. As discussed in prior sections, since service providers were transitioning to the new OHRS reporting system, CE LHIN data was expected to have quality and completeness issues in the new system. As a result, historical data from 2006 provider information was used for some of the services to compute current supply and average levels of service. These would not reflect increases in service levels due to recent investments and changing client needs. It was assumed that service groupings in all available data were reasonable representations of the uniqueness of services and the current units of service were appropriate measures of the system capacity. Since the provider surveys could not be deployed, based on discussions with the expert provider panel, data mined from the two available sources was relied upon for an accurate representation of service supply in the CE LHIN. The current sources did not provide information on services delivered by client location. Service data was aggregated by providers with their head-office location. Provider location was plotted on the demand maps to show service locations in the CE LHIN. This assumed that current services are provided from the head-office address of the service provider. In many cases, however, providers may have satellite offices and/or travel to the location of their clients to provide services. Data on levels of service delivered by satellite offices and client locations, however, was not available to be factored into the model. The data did not include information on provider catchment areas. As a result, it was not possible to estimate service capacity by sub-lhin areas or show service distribution on the CE LHIN maps. 23