Health Programme Evaluation by Propensity Score Matching: Accounting for Treatment Intensity and Health Externalities with an Application to Brazil (HEDG.

Slides:



Advertisements
Similar presentations
Impact analysis and counterfactuals in practise: the case of Structural Funds support for enterprise Gerhard Untiedt GEFRA-Münster,Germany Conference:
Advertisements

REGRESSION, IV, MATCHING Treatment effect Boualem RABTA Center for World Food Studies (SOW-VU) Vrije Universiteit - Amsterdam.
Presented by Malte Lierl (Yale University).  How do we measure program impact when random assignment is not possible ?  e.g. universal take-up  non-excludable.
Evaluation of the impact of the Natural Forest Protection Programme on rural household incomes Katrina Mullan Department of Land Economy University of.
The World Bank Human Development Network Spanish Impact Evaluation Fund.
BUSINESS AND FINANCIAL LITERACY FOR YOUNG ENTREPRENEURS: EVIDENCE FROM BOSNIA-HERZEGOVINA Miriam Bruhn and Bilal Zia (World Bank, DECFP)
PHSSR IG CyberSeminar Introductory Remarks Bryan Dowd Division of Health Policy and Management School of Public Health University of Minnesota.
Review of Barrier Free Approach and Additional Analysis of MEPS Data Related to ‘Potential’ vs. ‘Experienced’ Barriers.
WHY LARGE-SCALE RANDOMIZED CONTROL TRIALS? David Myers Senior Vice President IES 2006 Research Conference David Myers Senior Vice President IES 2006 Research.
What is a sample? Epidemiology matters: a new introduction to methodological foundations Chapter 4.
A METHODOLOGY FOR MEASURING THE COST- UTILITY OF EARLY CHILDHOOD DEVELOPMENTAL INTERVENTIONS Quality of improved life opportunities (QILO)
Impact Evaluation: The case of Bogotá’s concession schools Felipe Barrera-Osorio World Bank 1 October 2010.
Impact of property rights on poor households’ investment decisions: a treatment evaluation of a titling programme in Peru Oswaldo Molina July 1, 2008.
1 Photo: bbsweb.net Child Migration. 2 Migration’s changing age structure: a field-building opportunity Project to synthesize existing information, explore.
Palestinian Central Bureau of Statistics (PCBS) Palestine Poverty Maps 2009 March
Types of Evaluation.
The Economic Impact of Intensive Case Management on Costly Uninsured Patients in Emergency Departments: An Evaluation of New Mexico’s Care One Program.
The new HBS Chisinau, 26 October Outline 1.How the HBS changed 2.Assessment of data quality 3.Data comparability 4.Conclusions.
Chapter 8 Introduction to Hypothesis Testing
The Impact of Court Decentralization on Domestic Violence Against Women Raúl Andrade Jimena Montenegro March 2009.
Matching Methods. Matching: Overview  The ideal comparison group is selected such that matches the treatment group using either a comprehensive baseline.
Impact Evaluation in the Real World One non-experimental design for evaluating behavioral HIV prevention campaigns.
Quasi Experimental Methods I Nethra Palaniswamy Development Strategy and Governance International Food Policy Research Institute.
A Latent Class Call-back Model for Survey Nonresponse Paul P. Biemer RTI International and UNC-CH Michael W. Link Centers for Disease Control and Prevention.
Slide 1 Estimating Performance Below the National Level Applying Simulation Methods to TIMSS Fourth Annual IES Research Conference Dan Sherman, Ph.D. American.
Welfare Reform and Lone Parents Employment in the UK Paul Gregg and Susan Harkness.
Family Health Program Brazil Coverage and access Aluísio J D Barros Andréa D Bertoldi Juraci Cesar Cesar G Victora Epidemiologic Research Center, UFPel.
Targeted Interventions in Health Care: The case of PROMIN Sebastian Galiani Mercedes Fernandez Ernesto Schargrodsky.
Beyond surveys: the research frontier moves to the use of administrative data to evaluate R&D grants Oliver Herrmann Ministry of Business, Innovation.
Impact Evaluation Designs for Male Circumcision Sandi McCoy University of California, Berkeley Male Circumcision Evaluation Workshop and Operations Meeting.
EFFECTS OF HOUSEHOLD LIFE CYCLE CHANGES ON TRAVEL BEHAVIOR EVIDENCE FROM MICHIGAN STATEWIDE HOUSEHOLD TRAVEL SURVEYS 13th TRB National Transportation Planning.
AFRICA IMPACT EVALUATION INITIATIVE, AFTRL Africa Program for Education Impact Evaluation David Evans Impact Evaluation Cluster, AFTRL Slides by Paul J.
Nigeria Impact Evaluation Community of Practice Abuja, Nigeria, April 2, 2014 Measuring Program Impacts Through Randomization David Evans (World Bank)
Generalizing Observational Study Results Applying Propensity Score Methods to Complex Surveys Megan Schuler Eva DuGoff Elizabeth Stuart National Conference.
Applying impact evaluation tools A hypothetical fertilizer project.
Non-experimental methods Markus Goldstein The World Bank DECRG & AFTPM.
Measuring Impact 1 Non-experimental methods 2 Experiments
Randomized controlled trials and the evaluation of development programs Chris Elbers VU University and AIID 11 November 2015.
Critical Appraisal (CA) I Prepared by Dr. Hoda Abd El Azim.
Targeting of Public Spending Menno Pradhan Senior Poverty Economist The World Bank office, Jakarta.
Africa Program for Education Impact Evaluation Dakar, Senegal December 15-19, 2008 Experimental Methods Muna Meky Economist Africa Impact Evaluation Initiative.
Using Propensity Score Matching in Observational Services Research Neal Wallace, Ph.D. Portland State University February
Chile’s Supplier Development Program Irani Arráiz May 2015.
Randomized Assignment Difference-in-Differences
Overview of evaluation of SME policy – Why and How.
1 Validating Ex Ante Impact Evaluation Models: An Example from Mexico Francisco H.G. Ferreira Phillippe G. Leite Emmanuel Skoufias The World Bank PREM.
Differences-in- Differences. Identifying Assumption Whatever happened to the control group over time is what would have happened to the treatment group.
What can a CIE tell us about the origins of negative treatment effects of a training programme Miroslav Štefánik miroslav.stefanik(at)savba.sk INCLUSIVE.
Prof. (FH) Dr. Alexandra Caspari Rigorous Impact Evaluation What It Is About and How It Can Be.
1 Joint meeting of ESF Evaluation Partnership and DG REGIO Evaluation Network in Gdańsk (Poland) on 8 July 2011 The Use of Counterfactual Impact Evaluation.
Types of Studies. Aim of epidemiological studies To determine distribution of disease To examine determinants of a disease To judge whether a given exposure.
Methodology: IV to control for endogeneity of the measures of innovation. Results (only for regions with extreme values) Table 2. Effects from the 2SLS.
Africa Impact Evaluation Program on AIDS (AIM-AIDS) Cape Town, South Africa March 8 – 13, Randomization.
Workshop on MDG, Bangkok, Jan.2009 MDG 3.2: Share of women in wage employment in the non-agricultural sector National and global data.
Effects of migration and remittances on poverty and inequality A comparison between Burkina Faso, Kenya, Nigeria, Senegal, South Africa, and Uganda Y.
Do European Social Fund labour market interventions work? Counterfactual evidence from the Czech Republic. Vladimir Kváča, Czech Ministry of Labour and.
Segmentation of Brazilian labor market and manpower allocation between agricultural and non-agricultural sectors from 2004 to 2009 Priscila Casari Carlos.
1 Assessment of Potential Bias in the National Immunization Survey (NIS) from the Increasing Prevalence of Households Without Landline Telephones Meena.
Assessing the Impact of Informality on Wages in Tanzania: Is There a Penalty for Women? Pablo Suárez Robles (University Paris-Est Créteil) 1.
Patricia Gonzalez, OSEP June 14, The purpose of annual performance reporting is to demonstrate that IDEA funds are being used to improve or benefit.
Measuring Results and Impact Evaluation: From Promises into Evidence
Impact evaluation: The quantitative methods with applications
Matching Methods & Propensity Scores
Matching Methods & Propensity Scores
Jeremiah Maller Partner Organization: Operation Smile
Matching Methods & Propensity Scores
Evaluating Impacts: An Overview of Quantitative Methods
Sampling for Impact Evaluation -theory and application-
Analysing RWE for HTA: Challenges, methods and critique
Alternative Scenarios and Related Techniques
Presentation transcript:

Health Programme Evaluation by Propensity Score Matching: Accounting for Treatment Intensity and Health Externalities with an Application to Brazil (HEDG Working Paper 09/05, March, 2009) Rodrigo Moreno-Serra Centre for Health Economics & Department of Economics University of York AfrEA-NONIE-3ie Conference, Cairo, 2009

Introduction & Motivation Main programme evaluation challenge: ex-post construction of an adequate comparison group, often within a non- experimental setting, to obtain average treatment effects Key assumption: Values of treated and untreated outcomes for a given individual are not influenced by the treatment status of other individuals Usually unrealistic for health programmes: externalities can lead to underestimation of total programme impacts (cf. e.g., Miguel and Kremer, 2004)

Introduction & Motivation Possible solution in non-experimental settings: use availability of a health programme in a given area as the treatment variable of interest Treated individuals defined as those who live in areas where programme is in place (treated areas) Methodology already used, normally through an indicator variable for presence/absence of the programme One mean programme impact that accounts for health externalities to individuals in the treatment areas who did not directly receive the intervention Yet magnitude of externalities within each locality (and thus the associated average treatment effect) is likely to depend on the number of individuals who actually receive the programme’s services there: intensity of treatment

Suggested empirical methodology at a glance Use a measure of the programme’s population coverage across areas as the treatment variable of interest Estimate average treatment effects through comparisons between the health impacts of alternative coverage levels vs. reference level (e.g., zero) Need panel-data or repeated cross-sections (before and after) on coverage levels (phased-in programme) and individual variables Compare change in outcomes for individuals living in an area with coverage level l (a treatment area) to the change in health outcomes for similar individuals living in the area with coverage level 0 (the comparison area), for a number of l > 0 Implementation: propensity score matching estimators adapted to the case of multiple treatments (Imbens, 2000; Lechner, 2000), coupled with a difference-in-differences approach (PSDD)

Suggested empirical methodology at a glance PSDD estimator of the average treatment effect on the treated (ATT) with repeated cross-sections (Blundell and Costa-Dias, 2000) : Construction of comparison groups through propensity score matching for multiple treatments (generalized propensity score): more than one active treatment, i.e. coverage levels ATT computed by difference-in-differences Key assumption: bias stability ATT of living in area l vs. living in area 0 takes into account health benefits to individuals living in l who did not receive programme services themselves

An example: the Brazilian PSF Family Health Programme (PSF): MoH initiative with the stated aim of “improving the health status of covered families” Family Health Teams have to be formed by family doctor, nurse, assistant-nurse and 4-6 community health agents Monthly household visits: preventive and health promotion actions for all the individuals in a family (adults, children, seniors) Municipalities make PSF adoption decision: individuals mandatorily covered (visited) Broadest health programme ever launched in Brazil: 80 million people (2006), yet important variations across regions Average health gain for resident of a given region likely to increase (non-linearly?) according to coverage, also due to externalities

Data Porto Alegre is the comparison region (~ zero coverage, 98-03) Household survey data (1998 & 2003): >127,000 individuals (adults and children) and 34,000 households per wave Individual matching variables include household living conditions, demographics, education, labour and income characteristics

Evaluation question What are the average health impacts of being exposed to each of the eight observed PSF coverage levels from 1998 to 2003, compared to living in the comparison region during the same period (the “no-programme” benchmark)? Health outcomes: (1) self-assessed health; (2) bed due to illness; and (3) inability to perform usual activities due to illness One ATT is estimated for each of the eight relevant pairwise comparisons: being exposed to the PSF coverage level observed in region 1, 2, 3… vs. not exposed to the PSF in Porto Alegre Specification tests: matching successful for construction of similar comparison groups (both samples, adults and children)

Main results Overall, positive levels of PSF coverage in a region tend to lead to improvements in individual health outcomes (small effects for adults, larger estimated impacts for children) Largest ATT tend to be found for residents of the regions with the three highest median PSF coverage levels during (Belo Horizonte—16%, Recife—23%, Fortaleza—24%) E.g., children in Fortaleza vs. Porto Alegre: (1) 5-8p.p. higher prob of good SAH; (2) 3p.p. lower prob of bed episode; (3) 4-5p.p. lower prob of inability to perform usual activities due to illness But no clear pattern of increasing health benefits according to higher coverage levels: too few/low coverage levels available

Concluding remarks Health programme evaluation strategy that can be applied when information on actual impacts is needed to guide resource allocation and roll-out strategies, but only limited (routine, non-experimental) data are available Of course, validity of assumptions of the PSDD estimator with multiple treatments needs to be assessed case-by-case Impact estimates account for (i) different health endowments, and the potentially substantial (ii) treatment non-linearities and (iii) externalities from different levels of population coverage Comprehensive account of the health benefits generated by an intervention—not only its effects on actually “treated” individuals: relevance for policy-makers