• Share
  • Email
  • Embed
  • Like
  • Save
  • Private Content

Loading…

Flash Player 9 (or above) is needed to view presentations.
We have detected that you do not have it on your computer. To install it, go here.

Like this presentation? Why not share!

Cirm presentation impact evaluation of health microinsurance

on

  • 1,140 views

 

Statistics

Views

Total Views
1,140
Views on SlideShare
1,057
Embed Views
83

Actions

Likes
0
Downloads
0
Comments
0

2 Embeds 83

http://ifmr.ac.in 82
http://www.onlydoo.com 1

Accessibility

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

    Cirm presentation impact evaluation of health microinsurance Cirm presentation impact evaluation of health microinsurance Presentation Transcript

    • Impact Evaluation of HMI through RCT Two interventions targeting better renewal rates and health outcomes Tilak Mukhopadhyay Akhil S. Behl* CIRM, IFMR Research September 30, 2010 September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • Centre for Insurance & Risk Management Safety Nets for All Established in 2006 as a specialized design and research centre at the Institue of Financial & Management Research [IFMR] Committed to undertaking product design & action research to facilitate greater market outreach of risk management solutions among vulnerable households. Focus Areas Verticals Product Innovation Agriculture Action Research Livestock Product Development Health Market Making Catastrophe Data Warehousing Life Training Term Life Policy Advocacy & Savings September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • Objectives Objectives of the paper Exhibit relevance of RCTs in research on HMI Demonstrate the Stratified Randomization and Instrumental Variables in Randomization techniques being used in live projects by CIRM September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • The RCT Method: Literature & Practice Literature & Practice of RCT Exists as ‘gold-standard’ for clinical medical research, having been developed since 1920 by the likes of Fisher and Hill. Lately become popular for evaluation of impact studies in developmental interventions. Been used to evaluate [amongst many others]: Behavioural changes due to financial education [Bayer et. al., 1996 and Cole & Shastry, 2009] Impact of providing incentives on immunization choice [Banerjee et. al., 2010] Impact of micro-credit on financial behavior and outcomes [Duflo et. al., 2009] Few attempts yet in the field of HMI, none in India. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • Calcutta Kids - Stratified Randomization Project Details Introduction - Calcutta Kids Non Governmental Organization Working in urban slum of Fakir Bagan, in Howrah, West Bengal. Mandate to deliver healthcare through use of Community Health Workers. Previous Work - Maternal & Young Child Health Initiatives, since 2005. Counselling and monitoring of pregnant women during pre- & post-natal phases. Present Work - Providing in-patient insurance to households in its catch- ment area since 2009. CK has also expanded into a new area of the slum with the introduction of the HMI scheme. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • Calcutta Kids - Stratified Randomization Project Details Hypotheses & Intervention ) Hypothesis ) Lack of understanding of insurance = Insurance perceived as a loss by non-claimants = Lack of demand for HMI ) Theory of Change Creating value for non-claimants = Increased demand [renewals]. Intervention Out-Patient Counselling Service to all the insured September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • Calcutta Kids - Stratified Randomization Project Details Out-Patient Counselling Service Insurance clients that access out-patient care at a network facility will be visited by the CHW within two days to: Ensure that the client is following the doctor’s prescription. Provide simple advice to the client regarding treatment. Record the client’s health status, and present this information to a doctor who will advise on follow-up options. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • Calcutta Kids - Stratified Randomization Research Methodology Learning Agenda 1. What is the effect of the OPCS on client retention (renewal rate)? 2. What is the effect of the OPCS on the claims expense? 3. Why, and how much do clients value the OPCS? 4. Why, and how much do clients value the insurance product as a whole? 5. Is the OPCS self-financing? Can it cross-subsidise the insurance premium, either in this scheme or elsewhere? September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • Calcutta Kids - Stratified Randomization Research Methodology Randomization - The Rationale Objective - Impact Evaluation. Other techniques suffer drawbacks - the confounding effect of variables such as ‘time’ and ‘behavior’. Time - Change in the number of empanelled hospitals, epidemic during the period of treatment etc. Behavior - ‘understanding of the risk’, ‘risk-mitigating behavior’, and ‘value of insurance’ etc. The ‘control’ and ‘treatment’ groups are statistically identical in behavioral attributes. The comparison of control and treatment, in the same time period, reveals impact. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • Calcutta Kids - Stratified Randomization Research Methodology Randomization - Design ) Level of Randomization Renewal decision made at the policy/family level = randomization at policy/family level. Stratification Randomization assumes the treatment and control are being drawn from the same population. Considerable variation in the population can be expected on two counts: Exposure to MYCHI - affects health-seeking and risk-mitigation behavior etc. Past presence of CK in the area - peer influence and community trust. Stratification for these two variables improves representativeness and re- duces sampling error. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • Calcutta Kids - Stratified Randomization Research Methodology Figure: Randomization Design for the CK experiment September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • Calcutta Kids - Stratified Randomization Research Methodology Challenges Contamination of treatment and control. Loss of treatment group Low enrolment, Lack of reporting, Receiving OP treatment outside the network, and Migration. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • Calcutta Kids - Stratified Randomization Research Methodology Way Forward Enrolment in three phases ending in May, October, December. Enrolment, baseline survey, and randomization have been completed for the first phase. Enrolment for the second phase is in progress. Renewal data shall be available for analysis by December 2011. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • CARE Foundation - Instrumental Variable in Randomization Project Details Introduction CARE - Network of hospitals in various Indian states. CARE Foundation - Conducting research, imparting health education and leveraging technology to provide healthcare for low income households. Present Project Located in Yavatmal district of Maharashtra, which is a drought prone district and largely poor. Healthcare delivered via a hub & spoke delivery model. Village Health Champions leveraging hand-held devices to deliver care at the doorstep. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • CARE Foundation - Instrumental Variable in Randomization Project Details Hypotheses H1. Financing primary care through out-patient insurance shall reduce out-of- pocket health expenditure and incidence of in-patient illnesses. H2. Easily accessible and quality healthcare shall accentuate the health-seeking behavior of those insured. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • CARE Foundation - Instrumental Variable in Randomization Project Details Interventions I1 Out-Patient Insurance through a hub and spoke model with the clinic at the backend, and Village Health Champions (VHC) (Community health worker of CARE) with hand-held devices at the frontend in the villages. I2 Preventive and promotive services at the village level for OP insured clients. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • CARE Foundation - Instrumental Variable in Randomization Research Methodology Learning Agenda L1 Is the technological intervention of introducing a hand held device cost- beneficial for the sustainability of the scheme? L2 Does purchasing the OP insurance package increase access to health care, enhance health outcomes, and mitigate the financial risk of the insured participants? L3 What is the likely impact of health interventions - preventive and promotive interventions - on health outcomes, and expenditures in the population of interest? L4 Does the CARE intervention improve the quality of health services to its users? September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • CARE Foundation - Instrumental Variable in Randomization Research Methodology Randomization - The Rationale As earlier, omitted variable bias in other techniques in assessing the impact on outcomes. However, also, self-selection bias since the ‘treatment’ [insurance] is a vol- untary decision. Vanilla randomization is not enough. Literature suggests Instrumental Variable. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • CARE Foundation - Instrumental Variable in Randomization Research Methodology Randomization - IV An instrument must possess the following properties in our context: Cov (insurance ; instrument) = 0 6 Cov (behavior ; instrument) = 0 The second condition is easily created by randomization. If can create, through randomization, a variable which is correlated with the uptake of insurance, we have an instrument. ) Price of insurance is correlated with uptake = Randomized Discount-Vouchers. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • CARE Foundation - Instrumental Variable in Randomization Research Methodology Randomization - IV ..continued IV Design: 30 HH each in 30 villages = 900 Group Description Strength Control One-time free check-up by VHC 10 x 30 = 300 Treatment 1 80% discount on Insurance 10 x 30 = 300 Treatment 2 80% discount on Insurance & PPP 10 x 30 = 300 September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • CARE Foundation - Instrumental Variable in Randomization Research Methodology Randomization - IV ..continued Instrument does away with self-selection. People buy insurance because of the heavy discount and not behavioral at- tributes. Drop those in the treatment groups who do not act on our ‘intent-to-treat’. Randomization does away with omitted variable bias. Pure differences in outcomes can now be observed. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • CARE Foundation - Instrumental Variable in Randomization Research Methodology Randomization - Design In this study the impact will be observed through three difference effects: 1. C - T2 : Impact of OP insurance on insured people vs. uninsured people. 2. C - T1: Impact of combined intervention (OP insurance + preventive and promotive package) on insured people vs. uninsured people 3. T1 - T2: Impact of combined intervention (OP insurance + preventive and promotive package) on insured people vs. people with only OP insurance. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • CARE Foundation - Instrumental Variable in Randomization Research Methodology Challenges Loss of Control & Treatment groups Migration. Insufficient use of vouchers - due to lack of knowledge and/or understanding. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • CARE Foundation - Instrumental Variable in Randomization Research Methodology Way Forward Baseline survey has been completed and data-entry in electronic format presently in progress. Baseline report to be completed by the end of November. Midline qualitative research to be conducted in January. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • Acknowledgements Acknowledgements The authors are extremely grateful to Dr. David Canning & Dr. Ajay Mahal (Harvard School of Public Health) and to Dr. Clara Delavallade (University of Cape Town, South Africa) for their invaluable suggestions and comments on the CARE project, and for designing the research plan for the Calcutta Kids project respectively. The authors express sincere gratitude to Ms. Rupalee Ruchismita (Executive director, CIRM) and colleagues at CIRM for their encouragement and support. Also, the authors extend their best wishes to the implementation partners - CARE Foundation and Calcutta Kids. Above all, the authors are indebted to, and wish well the inhabitants of the command area of the CARE project in Yavatmal, and that of the Calcutta Kids project in Fakir Bagan, Kolkata. September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research
    • Thank You Visit us at http://www.ifmr.ac.in/cirm http://www.ifmr.ac.in/cirm/blog September 30, 2010 Impact Evaluation of HMI through RCT CIRM, IFMR Research