Marketing Digital Falcón & Solís Consultores. GRADO CERO PRESS
Rebecca Thornton, University of Michigan April 2, 2009 Laurel Hatt & Mursaleena Islam, Abt; Barbara...
-
Upload
job-spencer -
Category
Documents
-
view
213 -
download
0
Transcript of Rebecca Thornton, University of Michigan April 2, 2009 Laurel Hatt & Mursaleena Islam, Abt; Barbara...
Rebecca Thornton, University of Michigan
April 2, 2009
Laurel Hatt & Mursaleena Islam, Abt; Barbara Magnoni, EA Consultants; Freddy Solís, ALVA
Consultores; Erica Field, Harvard
Health Insurance in Nicaragua: A Randomized Evaluation
Motivation
Health insurance can mitigate shocks faced by informal workers
Low rates of enrolling in health insurance How does having health insurance assist informal
workers?
This paper
What are the determinants of health insurance enrollment? Is information enough? How important are cost factors or convenience? Is there adverse selection? Can local MFIs assist in the delivery?
What is the impact of health insurance? What happens to utilization and expenditures?
What are the determination of retention? How does MFI status/cooperation affect retention?
Econometric Challenges
Most studies use cross-sectional or longitudinal data; typical strategy: Compare demographic data and health status/utilization
between those with and without insurance or compare individuals before and after having health insurance
Limitations Difficult to quantify the most important determinants of uptake Choice to have health insurance is endogenous
More health conscious or higher income upward bias Less healthy downward bias
Experimental Design
“Ideal experiment” to study the impact of health insurance: randomly force individuals to have health insurance. Then measure the effects
Problem Can’t force individuals Couldn’t measure what affects health insurance take-up
Encouragement design: Randomly encourage some to get health insurance
Insurance Program
Prior to January 2007, only formal sector workers had insurance Social Security Institute (INSS) extended health insurance to informal
workers Partner with 3 MFIs to facilitate subscription Cost
$18 first two months (no coverage); $15 monthly in subsequent months pre-paid plan, no reimbursement or co-pay
Coverage Spouse eligible for maternity services Dependent children under 12 eligible Comprehensive: preventive, diagnostic, and curative
Baseline Survey
Representative sample in largest Managua markets (~June 2007) Eligibility
Business owners ages 18-55 Had government ID card No formal insurance
2 Rounds of Baseline survey to calibrate the experiment Round 1: Randomly select eligible respondents from a sample frame Round 2: Interview all eligible respondents from a market census
Num %Away 1,591 0.21Refused 518 0.07Not Eligible 1189 0.16Other 292 0.04Completed 4,002 0.53
Randomization
Encouragement design: Lottery at end of the survey
Round 1 Num %
Pure Control 125 0.17
Brochure 117 0.16
2 month at INSS 123 0.17
2 month at MFI 123 0.17
6 month at INSS 126 0.17
6 month at MFI 114 0.16
Total 728 100
Round 2
Brochure 1,157 0.35
6 month at INSS 1,066 0.33
6 month at MFI 1,050 0.32
Total 3,273 100
Signing up: Government ID, photoTravel to an institution
Randomization
Convenience experiment Randomly selected individuals for door to door affiliation in one market Conditional on not signing up already – not randomly selected at the
beginning
Door to door, 6 month subsidy 112 0.140
Door to door, Brochure 63 0.079
6 month subsidy 326 0.406
Brochure 301 0.375Total 802
Affiliation
Measure uptake of insurance Download data from INSS Merge with survey and lottery data – government ID
Follow-up Survey
Conduct a follow-up survey in 2008 of a subsample of respondents Only interview in the 3 largest markets Exclude those offered 2 month subsidies
93% completion rate Rates of attrition
Number PercentNothing 119 0.93Information 1138 0.936 month at INSS 702 0.946 month at MFI 673 0.93Door to Door + 6 month 112 0.92Door to Door + Information 62 0.98Total 2806 0.93
Empirical Strategy
Uptake of insurance
Effects of having insurance
2608 respondents interviewed at baseline and follow-up
Retention (among those who enrolled)
iiii XVInsurance
iiii XV Retained
iiiii XYInsuranceY 12
Results: Insurance Take-up
Informational brochure alone had no impact on take-up Cost an important issue: large impact of subsidies Other factors important: many people still do not sign up
Results: Insurance Take-up
Individuals significantly more likely to go to the INSS
Insurance Take-up: Door to door
Door-to-door offer more effective than 2 months subsidy Adding door-to-door almost doubles take-up for subsidy winners
IV Results: Utilization and Expenditures
Visited Number of Visits Amount spentCoeff. SD Coeff. SD Coeff. SD
Pharmacy -0.026 [0.049] 0.3 [0.354] -0.705 [0.374]Private doctor 0.011 [0.021] 0.034 [0.043] 0.149 [0.103]Laboratory -0.068* [0.019] -0.047 [0.192] -1.011* [0.287]EMP 0.376** [0.060] 1.334* [0.357]Private clinic/hospital -0.111* [0.030] -0.498** [0.105] -0.762** [0.151]Public (MINSA) clinic -0.05 [0.033] -0.442 [0.375] 0.012 [0.046]Public (MINSA) hospital -0.094 [0.047] -0.05 [0.292] -0.007 [0.031]All/Any -0.015 [0.016] 1.059*** [0.053] -0.32 [0.195]
iiiii XYInsuranceY 12
Other Effects
No impact on Utilization of preventative care such as: General
check up, Prenatal test, Pap Smear, Blood pressure, Blood test, Urine test
Sick days Wait before going to MD # times sick?
Caveats Few people insured Only insured for 6 months
Results: Retention
Summary of Results
Enrollment Low rates of signing up Informational brochure not main barrier to signing up Costs and convenience important
No health effects, only substitution Switch out of public clinic and costly private clinics to EMPs
Large savings in out of pocket expenditures Savings especially large for costs of laboratory, pharmacy and private
clinics with insurance Cost of health insurance ($178) higher than baseline health
expenditures ($87) Savings did not “make up” for cost of insurance
Retention rates: less than 10%
Conclusions
Lessons learned Cost and convenience important in insurance enrollment Retention a challenge Determinants of enrollment and retention include quality,
convenience, trust/reliability, and satisfaction
Thinking about other programs Strong institutional capacity and partnerships is essential Quality is essential to ensure satisfaction MFIs face challenges in delivering insurance
ADDITIONAL SLIDES
Baseline Health Behavior
CostsTotal HH health care costs 88.27Total health care costs for R 41.61Last visit cost for R 18.24
HealthEver sick in 2007 0.80Times sick in 2007 2.57Days waited to see doctor 4.32
Dizziness 0.17Diabetes 0.06Hypertension 0.19Heart Problems 0.06Respiratory Problems 0.14Sight/Hearing Limitations 0.25Stress 0.15Kidney Problems 0.25Headaches 0.40Skin Problems 0.08Allergies 0.15
Flu 0.60Fever 0.32Vomitting 0.08Diahrea 0.10Cough 0.19Chest pains 0.15Back pains 0.26
Baseline Health Behavior
Visited Num Visits Expenditure
Mean SD Mean SD Mean SD Median
Pharmacy 0.72 0.45 3.21 4.15 42.66 91.22 16.04
Private doctor 0.28 0.45 2.18 2.64 0.91 2.36 0.34
Laboratory 0.26 0.44 2.09 2.78 25.67 61.93 13.40
Social Security Clinic (EMP) 0.01 0.08 3.88 5.45 14.88 47.25 28.90
Private clinic/hospital 0.14 0.35 2.31 2.17 24.04 73.63 7.50
Public (MINSA) health center 0.16 0.36 2.80 3.10 0.26 2.00 2.40
Public (MINSA) hospital 0.09 0.28 2.79 5.20 2.33 16.22 12.00
Many self-medicate and visit pharmacies Visits to both private and public providers
Where did individuals enroll?
Obs INSS MFI
Control 1 1.00 0.00
Information 17 0.88 0.12
6 Month - INSS 279 0.98 0.02
6 Month - MFI 233 0.01 0.99
Results: Insurance Take-up
(N=530)
Qualitative Methods
Individual Interviews (2007) 40 randomly selected subsidy winners in Huembes market who
had not signed up for insurance. 20 survey participants who had registered for insurance
Focus Group Discussions (2008) 9 focus group discussions stratified across those who got
subsidies and did not and those who signed up and did not (73 people)
Sampled respondents by phone and visited their market stall Discussions held at locations near each market. All participants paid approximately 5 dollars.
Information -0.041 -0.048
6 month –INSS 0.328*** 0.330***
6 months –MFI 0.277*** 0.274***
Door to door 0.172** 0.191**
Door to door * 6 month 0.139 0.154*
Male -0.025 -0.025*
Education 0.001 0.002
Married 0.012 0.022
MFI client 0.033 0.029
Number of children -0.014 -0.011
Has children under 12 0.033 0.037**
Log income 0.009
Has any savings 0.013
Smokes 0.009
Any chronic disease 0.044*
Ever sick -0.009
Log costs of healthcare 0.00
Number of visits to provider 0.000
Any visit to provider 0.031
Observations 2608 2215
6 month subsidyAll All MFI Client Non-MFI Client
Information -0.632* -0.637*6 month –INSS -0.910*** -0.927***6 months –MFI -0.895*** -0.924*** 0.006 0.058** -0.03Door to door -0.267 -0.282Door to door * 6 month 0.193 0.218Male 0.001 0.016 0.012 -0.099* 0.055Years of Education 0.01 0.009 0.007 0.015 0.001Married 0.026 0.027 0.031 0.008 0.033MFI client 0.016 -0.003 -0.007Number of children -0.007 -0.006 -0.006 -0.002 -0.005Has children under 12 -0.018 0.004 -0.018 0.019 -0.038Log income -0.003 0.007 0.036 0.004Has any savings 0.019 0.014 -0.045 0.047Smokes -0.035* -0.03 0.101** -0.056Any chronic disease -0.032**Ever sick 0.002Log costs of healthcare 0Number of visits to provider 0.002Any visit to provider 0.038Observations 530 470 387 157 230
What do we learn?
How do we contextualize the empirical findings? Qualitative research Won the subsidy but did not sign up
Most said intended to sign up; half said they didn’t have the time Trouble obtaining required documentation (photos, copy of ID
card, children’s birth certificates, and subscription forms) Did not fully understand the affiliation process.
Confusion about the insurance product Legitimacy that this was a “real” product; proof of enrollment; getting
trapped paying MFIs vs. INSS EMPs