Assessing private expenditures in HIV/AIDS: methods, tools, examples Dr Ravi P. Rannan-Eliya Targeted programs financing and budgeting for HIV/AIDS Seminar Yalta Hotel, Yalta, Crimea, Ukraine 19 July 2007
Overview Private expenditures Third party spending Household out-of-pocket Issues in estimating household expenditures for NHA Recommended approach Specific issues for HIV/AIDS Examples 1
Types of private expenditures Third-party spending Commercial insurance Collect data from insurance firms, claims databases Employer expenditures Employer surveys Household out-of-pocket spending Co-payments to public providers Government administrative data Payments to private providers?? 2
Issues in estimating household expenditures 3
Issues in estimating household expenditures Substantial share of overall expenditures: 25-75% outside OECD compared to 5-15% in OECD Accuracy more important than in OECD Important for policy because of impact on households Institutional records data often not available Problems with use of household survey data Challenges in estimating true level and final use of expenditures 4
Problems of Household Survey Data Example: Sri Lanka 5
Recall loss for hospitalizations 6
Household expenditures: National Accounts vs. Household Surveys 7
Global experience with using household survey data Suffer from significant non-sampling errors E.g.: recall loss, telescoping, etc. National accountants use mix of data sources to estimate household spending Sampling errors if complete population is not covered Surveys may lack sufficient detail or samples to meet NHA requirements Surveys expensive and are not available every year 8
Recommended Approach 9
Integrative Approach 1. Divide the problem into separate components based on type of spending and available data sources, eg: Purchase of medicines from pharmacies Payments to private doctors 2. Compile data from both sides: Household data Provider data 3. Reconcile data sources by triangulating data sources Take into account reliability and accuracy of data sources Adjust for potential biases Compare data across years 10
Integrative approach example Problem: To estimate private clinic doctors revenues 11
Integrative approach example Problem: To estimate private clinic doctors revenues 12
Integrative approach example Problem: To estimate private clinic doctors revenues 13
Integrative approach example Problem: To estimate private clinic doctors revenues 14
Integrative approach example Problem: To estimate private clinic doctors revenues 15
Integrative approach example Hong Kong DHA: Comparison of estimates of private doctors' revenues 10,000 9,000 8,000 7,000 6,000 5,000 4,000 3,000 2,000 1,000-1989 1990 1991 1992 1993 1994 1995 1996 1997 GHS 1992 GHS 1996 IMS and tax data HES 90/95 Proposed DHA estimate Example: Hong Kong 16
Household Spending in Hong Kong Public sector user fees Government hospital data Private hospitals Special hospital survey, drug sales data Western medicines Industry sales data, IMS Private medical practitioners Various data sources inc. drug sales data, tax returns, utilization surveys, doctor surveys 17
Household Spending in Canada Hospitals Canadian MIS Database Residential/nursing homes Survey of residential care facilities Pharmaceuticals Industry data from AC Nielsen Canada Medical supplies AC Nielsen Canada 18
Household Spending in Sri Lanka Public sector user fees Government hospital data Private hospitals Special hospital survey, drug sales data Western medicines Industry sales data, IMS Private medical practitioners Various data sources inc. drug sales data, utilization surveys, clinic surveys 19
Specific Issues for HIV/AIDS 20
Issues for HIV/AIDS HIV/AIDS affects only a small percentage of population > Samples too small in household or provider surveys Most PLWHA do not know their HIV status Stigma affects reporting Major treatment expenses related not to HIV, but consequences of HIV - opportunistic infections Treatment may not be at specialized HIV facilities but in general healthcare facilities 21
Suggestions Most personal medical treatment expenses for PLWHA is outcome of Opportunistic Infections (OI) In low prevalence countries (<3%), use targeted surveys of PLWHA to measure OI expenditures Use integrative approach to measure expenditures for retail goods 22
HIV/AIDS - Opportunistic infections Thailand NASA P * Q approach Expenditure = Price x Quantity Application Apply to Opportunistic infections Categorize OIs For each OI, obtain data on annual per capita frequency of OI and average household expenditure per episode Requires patient surveys/clinical studies, and advisable to stratify by HIV/ART status 23
HIV/AIDS - Others Safer sex practices Combine data using integrative approach: VCT Household expenditure surveys Retail sales data for condoms Production and import data for condoms Thailand NASA P * Q approach Expenditure = OOP payment x Quantity 24
HIV/AIDS - VCT Thailand NASA P * Q approach Expenditure = OOP payment x Quantity 25