AboutSubmit my research

Health, Methodology

Research articles · Dec 31, 2015

Creating a synthetic database for use in microsimulation models to investigate alternative health care financing strategies in Canada

Deirdre Hennessy, Claudia Sanmartin, Sabha Eftekhary, Laurie Plager, Jennifer Jones, Kanecy Onate, Neil McEvoy, Chantal Hicks, Raisa Deber

Health Analysis Division, Statistics Canada, Canada; University of Toronto, Institute of Health Policy, Management and Evaluation, Canada; Social Analysis and Modelling Division (Modelling Group), Statistics Canada, Canada

Volume 8(3) Winter 2015

Cite as: D. Hennessy, C. Sanmartin, S. Eftekhary, L. Plager, J. Jones, K. Onate, N. McEvoy, C. Hicks, R. Deber; 2015; Creating a synthetic database for use in microsimulation models to investigate alternative health care financing strategies in Canada; International Journal of Microsimulation; 8(3); 41–74. doi: 10.34196/ijm.00121

Abstract

The objective of this study was to develop a synthetic database which would be used with a microsimulation model to evaluate alternative health care financing options in Canada. The Social Policy Simulation Database (SPSD), developed by Statistics Canada, was enhanced with population-based data on health status, disability, disease status, health service use and costs from the Canadian Community Health Survey (CCHS) and other relevant data sources. The Social Policy Simulation Database and Model (SPSD/M) is a microsimulation tool designed to investigate the financial interactions between governments and individuals in Canada. Typically the SPSD/M has been used to inform policy around income tax, commodity tax, and cash transfer reform. We supplemented the model’s database with health data to enable analysis of alternative health care financing options in Canada. Data on health status, disability, disease status, health service, medication use and out-of-pocket spending on health care were drawn together from population-based surveys (including the CCHS, the Canadian Health Measures Survey (CHMS), Survey of Household Spending (SHS)), health administrative data (the Discharge Abstract Database (DAD)) and estimates of health service use and cost available in the literature. Costs of health services and drugs were assigned to individuals in the CCHS 2009/2010 (host health dataset) through imputation. The resulting health dataset was merged with the SPSD. Models using the enhanced SPSD-Health will allow health policy makers and academics to “try-out” alternative health care financing options, and consider their monetary impacts on individuals and families in Canada.

This article contains 7 figures and 5 tables.

Figures

Figure 1. Map of data sources used to construct the SPSD.
Figure 2. Map of data sources used to supplement the SPSD with health information.
Figure 3. Validation of imputed hospitalization costs.
Figure 4. Validation of imputed medication costs.
Figure 5. Validation of imputed drug insurance status.
Figure 6. Validation of out-of-pocket costs for health services.
Figure 7. Schematic diagram of the variables used to statistically match the CCHS 2009/10 PUMF to the SPSD.

Tables

Table 1.
Cost type Data source Method Matching/ Imputation Variables
1. Cost per visit to physicians Schedule of costs form the NPD, publicly available from CIHI Simple multiplication of unit cost per-visit by number of physician visits reported in the CCHS 2009/10 PUMF n/a
2. Cos per visit allied health professionals Schedule of costs from Association of Workers Compensation Boards of Canada (AWCBC), publicly available from AWCBC Simple multiplication of unit cost per-visit by number and type of allied health professional visit reported in the CCHS 2009/10 PUMF n/a
3. Hospitalization costs CCHS 2007/08 linked to the CIHI DAD for the calendar year 2008–2009 Annual hospitalization costs were calculated for hospitalized respondents of the CCHS 2007/08linked to the DAD 2008–2009. Hospitalization costs were imputed onto the CCHS 2009/10 PUMF using “hot-deck” methods. Separate imputations for males and females.
Age, income, chronic disease and disability status
4. Prescription drug costs CHMS and British Colombia Drug Atlas 2nd edition. Reference population/period is the British Columbia population in 2006-
Publicly available from the University of British Columbia		 Average annual cost of drugs by class is assigned to respondents of the CHMS 2007 to 2011. Drug costs are then imputed onto CCHS 2009/10 PUMF, using “hot-deck” methods. Separate imputations for males and females.
Age, income and chronic disease status.
5. Prescription drug insurance CCHS 2007 Rapid Response Hot-deck Separate imputations for males and females. Age, income and chronic disease status.
6. Prescription drug insurance eligibility HIAD Assignment n/a
7. Out-of-pocket costs for health care SHS Hot-deck Income
8. Per-capita healthcare costs CIHI NHEX Assignment n/a

Abbreviations: NPD = Nation Physicians Database; CIHI = Canadian Institute of Health Information; DAD = Discharge Abstract Database; CCHS = Canadian Community Health Survey; PUMF = Public Use Microdata File; CHMS = Canadian Health Measures Survey; HIAD = Health Insurance Access Database; NHEX = National Health Expenditures report; SHS = Survey of Household Spending

Table 2.
Data source Data type Exclusions
CCHS Population-based survey • Children younger than 12 years
• Persons living on reserves and other Aboriginal settlements.
• Full-time members of the Canadian Armed Forces.
• Residents of institutions including institutions for the elderly and prisons.
• Some remote areas of Quebec
CHMS Population-based survey • Cycle 1 excluded children under 6 years
• Cycle 2 excluded children under 3 years
• Persons living in the 3 territories, Yukon, Northwest Territories and Nunavut.
• Persons living on reserves and other Aboriginal settlements.
• Full-time members of the Canadian Armed Forces.
• Residents of institutions including institutions for the elderly and prisons.
• Some remote areas of Canada.

Abbreviations: CCHS-Canadian Community Health Survey; CHMS-Canadian Health Measures Survey.

Table 3.
Variable description Data source Variable type Categories
Chronic health condition CCHS Numeric 0-None,
1-One,
2-Two +
Disability flag CCHS Categorical 0-No-
1-Yes
Imputed prescription drug insurance status CCHS 0-No-
1-Yes
Self-perceived health CCHS Categorical 0-poor-4-excellent,
5 not stated
Self-perceived mental health CCHS Categorical 0-poor-4-excellent,
5 not stated
Health Utility Index, category CCHS Categorical 0-None,
1-Mild,
2-Moderate,
3-Severe
Annual cost per capita for any doctor NHEX Numeric-dollars 0-max
Annual cost per capita of hospital stay NHEX Numeric-dollars 0-max
Annual cost per capita of other professionals NHEX Numeric-dollars 0-max
Annual cost per capita of drugs NHEX Numeric-dollars 0-max
Annual cost for visits to family doctor CCHS Numeric-dollars 0-max
Annual cost for visits to specialist doctor CCHS Numeric-dollars 0-max
Annual cost for visits to eye doctor CCHS Numeric-dollars 0-max
Annual cost for visits to physio CCHS Numeric-dollars 0-max
Annual cost for visits to SAOT CCHS Numeric-dollars 0-max
Annual cost for visits to PSYCH CCHS Numeric-dollars 0-max
Imputed annual cost of hospital stay DAD Numeric-dollars 0-max
Imputed annual cost of drugs CHMS Numeric-dollars 0-max
Annual OOP costs for prescribed drugs SHS Numeric-dollars 0-max
Annual OOP costs for other medicines and pharmaceutical products drugs SHS Numeric-dollars 0-max
Annual OOP costs for healthcare supplies SHS Numeric-dollars 0-max
Annual OOP costs for physician care SHS Numeric-dollars 0-max
Annual OOP costs for eye care goods and services SHS Numeric-dollars 0-max
Annual OOP costs for dental services SHS Numeric-dollars 0-max
Annual OOP costs for hospital care SHS Numeric-dollars 0-max
Annual OOP costs for healthcare practitioners SHS Numeric-dollars 0-max
Annual OOP costs for other medical services SHS Numeric-dollars 0-max

Abbreviations: CCHS-Canadian Community Health Survey; NHEX-Nation Health Expenditures Survey; DAD-Discharge Abstract Database; CHMS-Canadian Health Measures Survey; SHS-Survey of Household Spending; OOP-Out-of-pocket; SAOT-Speech language pathologist, audiologist or occupational therapist; PSYCH-Psychologist

Table 4.1.
Health care cost type % with family doctor cost % with specialist doctor cost % with eye doctor cost % with hospital cost % with medication cost
Data source CCHS SPSD-H CCHS SPSD-H CCHS SPSD-H CCHS SPSD-H CCHS SPSD-H
Weighted % (percent) 76.8 75.7 29.8 28.8 39.7 40.6 6.5 6.0 57.1 54.2
Average cost ($) 134 131 180 179 92 90 10,129 10,242 925 899
Median cost ($) 89 89 108 108 75 75 4,451 4,537 669 578
Minimum cost ($) 27 27 51 51 42 42 578 578 48 48
Maximum cost ($) 1,584 1,584 1,008 1,008 1181 1,181 613,342 537,324 5,375 5,375

Abbreviations: SPSD-H Social Policy Simulation Database-Health; CCHS-Canadian Community Health Survey

Table 4.2. Validation of match-merging of SPSD-Health to CCHS.
Health care cost type % with prescription drug insurance % with physiotherapy cost % with chiropractic cost % with speech language pathologist/audiologist/occupational therapist cost % with psychologist cost
Data source CCHS SPSD-H CCHS SPSD-H CCHS SPSD-H CCHS SPSD-H CCHS SPSD-H
Weighted % (percent) 73.3 72.6 9.6 9.2 11.5 11.3 2.5 2.5 3.5 3.7
Average cost ($) n/a n/a 347 340 287 281 142 150 659 681
Median cost ($) n/a n/a 216 215 189 189 79 79 433 519
Minimum cost ($) n/a n/a 31 31 27 27 33 33 86 86
Maximum cost ($) n/a n/a 1,992 1,992 1,627 1,627 1,075 1,075 2,433 2,433

Abbreviations: SPSD-H Social Policy Simulation Database-Health; CCHS-Canadian Community Health Survey