Module 5: Data Preparation

Transcription

1 Module 5: Data Preparation This presentation was prepared by Adam Wagstaff and Caryn Bredenkamp 1 Which data? Which data? In what form?

2 WHICH VARIABLES? Minimum data requirements: Health lhoutcomes module Measure of living standards Inequalities in Inequities and Benefit health outcomes or decomposition incidence utilization analysis Health status variables Health care utilization variables Determinants of health care utilization NHA data, by provider Fees paid to providers, but not needed for CUS assumption

3 Minimum data requirements: Health lhfinancing module Consumption or income Out of pocket payments Non food consumption Poverty line Prepayments for health care NHA data on health financing mix Additional micro data Catastrophic Impoverishment Progressivity (optional) (optional) o A few words on the measurement of living standards All analyses require a measure of living i standards d to: rank households (e.g. inequalities and BIA) and/or capture financial well being (e.g. financial protection analysis) Measures of living standards d include: income asset index / wealth score expenditure consumption

4 WHICH DATA SOURCES HAVE THE NECESSARY VARIABLES? Data sources Household surveys Demographic andhealth Survey World Health Survey Multiple IndicatorClusterSurvey LSMS or multipurpose household survey Household budget survey / income and expenditure survey Other micro data National Health Accounts data Direct taxes: Information on tax rates andbrackets Indirect taxes: Information on tax rates, brackets and tax exemptions Poverty lines

5 Survey Household Surveys Health outcomes, utilization, and spending Living standards and health expenditure Other comments Demographic outcomes and utilization asset score is isoften collected and Health Survey (DHS) for women and children, limited info on men s health andncds NCDs, none on no consumption data no information on health expenditure regularly data are highly comparable across general care utilization countries and within countries over time World Health outcomes and utilization asset data consumption data Survey (WHS) (including adults and NCDs) consumption data (limited) income data health expenditure data are limited tends to be dated standardized Living Standards and Measurement type Surveys (LSMS) Household budget information on general care utilization limited i d information i on health outcomes and noncurative utilization limited information on health care outcomes or good consumption data often good health care expenditure data often ability to link utilization to expenditure excellent data on consumption collected at irregular intervals not a high h degree of standardization among non pure LSMS not standardized across countries NationalHealth Accounts data National Health Accounts data for BIA Expenditure data dt NHA country reports contain tables on government health hexpenditure for different levels l of care (e.g. primary outpatient, hospital outpatient, hospital inpatient) Or, some public expenditure reviews (PERs) can be used Or, data obtained directly from Ministries National Health Accounts share for progressivity analysis Financing data Only NHA shares are needed and can be found in NHA country tables at and in national NHA reports at

6 Other micro data Chiefly needed for progressivity analysis, these data are generated through combining household survey data with other information Indirect tax contributions to health Indirect taxes (e.g. VAT) needs to be imputed using information on official tax rates and exemptions of various goods and services listed in the consumption module Direct taxes Might appear in data Typically need to be assessed by applying li official i tax brackets to individual income / consumption SHI contributions Might appear in data More likely, need to be calculated by applying official SHI contribution rates to individual income or employment category Poverty lines National poverty lines International poverty lines $1.25 or $2.00 per day, appropriately expressed in local currency units. PPP conversion factors can be found at org/icpext/resources/ ICP_2011.html Poverty P t lines, rates and gaps can also be found at: tml

7 WHICH DATA STRUCTURE? Data structure ADePT reads STATA (.dta) or SPSS (.sav) files In the Health Outcome module, observations should be at the individual level In the Health hfinancing i modules, observations should be at the household level

8 Resources to help to get data into the correct shape ADePT ready dt datasetst DHS, WHS, MICS, LSMS and other datasets 82 and counting if you use them, please register on the relevant survey website STATA.do files for transforming publically available datasets into the above form dofiles for merging of files / modules dofiles for constructing indicators Do files can be modified to include different variables or construct indicators differently Example ofdofile

9 Relax! Making data ADePT ready is no more complicated than for any other software program Related materials Guide to methods: Analyzing Health Equity Using Household Survey Data ADePT Health Manual: Health Equity and Financial Protection Online video tutorials Health EquityandFinancial Protection reports (ongoing) Health Equity and Financial Protection datasheets (ongoing) BookAttacking Inequality in the Health Sector Training events org/povertyandhealth and org/adept

10 ANNEX: DATA REQUIREMENTS IN DETAIL Inequalities: concentration curve and concentration index Health outcomes and health hutilization, i for example: Child survival (e.g. under 1 mortality rate) Anthropometrics t (e.g. height for age hihtf z score ) Stunting prevalence Diagnosed withdiabetes Self assessed health (dichotomous, not categorical) Living standards measure continuous variable (e.g. Consumption, expenditure, income, asset index or score) Weights and survey settings relate to sample design information (sampling weight, cluster, strata) Household ID 20

11 Decomposition and inequities Health outcomes and health utilization variables Living standards measure continuous variable (e.g. Consumption, expenditure, income, asset index or score) Weights and survey settings relate to sample design information (sampling weight, cluster, strata) Household ID Standardizing variables for health outcomes (e.g. age and gender) Standardizing variables for health care utilization (e.g. age, gender andhealthneed need ) Control variables (e.g. income, education, please of residence and health insurance) Benefit incidence analysis Under constant unit subsidy bid assumption, need: Utilization by subsector (e.g. PHC, polyclinic, hospital) for each household member (or a representative household member) from household survey Living standards for each household (e.g. consumption of asset index) Total govt. expenditure on health (i.e. subsidies), broken down by subsector from a National Health Accountsreport (or directly from a Ministry) Under other assumptions, also need household dataon feespaidto publicprovidersproviders

12 Financial protection: Catastrophic and impoverishing expenditure Out of pocket spending on health h Total household consumption (or expenditure) For catastrophic payments: Total household non food consumption (or expenditure) For impoverishment: Poverty line(s) in local currency Weights and survey settings Household ID Progressivity Income or consumption variables ibl Health payment variables, e.g. out of pocket payments, private insurance, social healthinsurance contributions, and taxes Household size (if not already expressed on per capita basis) NHA health financing shares Indirect (consumption) tax rates if tax needs to be imputed Knowledge of structure of health insurance premiums (e.g. x% of income) if insurance premiums need to be imputed

13 Related materials Guide to methods: Analyzing Health Equity Using Household Survey Data ADePT Health Manual: Health Equity and Financial Protection Online video tutorials Health EquityandFinancial Protection reports (ongoing) Health Equity and Financial Protection datasheets (ongoing) BookAttacking Inequality in the Health Sector Training events org/povertyandhealth and org/adept