Achieving Universal Health Coverage with an Unenforceable Mandate: Evidence from the Government of Indonesia's JKN Mandiri Program

Last registered on December 28, 2017

Pre-Trial

Trial Information

General Information

Title
Achieving Universal Health Coverage with an Unenforceable Mandate: Evidence from the Government of Indonesia's JKN Mandiri Program
RCT ID
AEARCTR-0000815
Initial registration date
October 21, 2015

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published
October 21, 2015, 9:48 AM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Last updated
December 28, 2017, 10:55 AM EST

Last updated is the most recent time when changes to the trial's registration were published.

Locations

Region

Primary Investigator

Affiliation
MIT

Other Primary Investigator(s)

PI Affiliation
MIT
PI Affiliation
Harvard University
PI Affiliation
The SMERU Research Institute
PI Affiliation
Massachusetts Institute of Technology
PI Affiliation
Massachusetts Institute of Technology

Additional Trial Information

Status
On going
Start date
2014-11-03
End date
2018-12-31
Secondary IDs
Abstract
Law 40/2004 on the National Social Security System and Law 24/2011 on Social Security Carriers (BPJS) outlined Indonesia’s path towards the creation of a universal social security system. A primary goal is to achieve universal health coverage through a mandatory public health insurance scheme – Jaminan Kesehatan Nasional (JKN). Begun in January 2014, JKN aims to achieve total coverage by 2019.

JKN coverage is determined by household’s income and employment status. The Government of Indonesia (GoI) covers the insurance premia of the poor and near-poor, while wage workers premia are covered jointly by the employee and the employer. The toughest challenge will be the registration and retention of non-poor non-wage workers, who have to register for the program and make monthly payments.

Ensuring non-wage worker participation is especially difficult: not only is it harder for the government to locate these individuals and encourage them to sign up, but the fact that they have to bear the cost of their premia, and remit it directly to the government each monthly, decreases the probability that they would sign up.

The main objective of this project is to rigorously test potential strategies to increase the participation of non-poor informal workers into the insurance scheme. We focus on incentives to enroll at the household level and explore three main factors that could potentially be hindering participation: monetary costs, hassle costs and lack of information. Moreover, we test for adverse selection by looking at whether different incentives attract different types of people.
External Link(s)

Registration Citation

Citation
Banerjee, Abhijit et al. 2017. "Achieving Universal Health Coverage with an Unenforceable Mandate: Evidence from the Government of Indonesia's JKN Mandiri Program." AEA RCT Registry. December 28. https://doi.org/10.1257/rct.815-3.0
Former Citation
Banerjee, Abhijit et al. 2017. "Achieving Universal Health Coverage with an Unenforceable Mandate: Evidence from the Government of Indonesia's JKN Mandiri Program." AEA RCT Registry. December 28. https://www.socialscienceregistry.org/trials/815/history/24537
Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
Experimental Details

Interventions

Intervention(s)
Intervention Start Date
2014-11-03
Intervention End Date
2015-12-31

Primary Outcomes

Primary Outcomes (end points)
The first key outcome is enrollment or how each of the treatment affects the take-up of JKN. Relatedly, we will look at how different treatment affect the premium payment behavior. In addition, we will look at health care seeking behavior in the form of outpatient and inpatient claims from administrative data. Lastly, we will look at how pre-existing variables differ across treatments. The variables we will focus on are pre-trial health care system utilization together with demographic variables.
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
The trial consists of four different treatment arms, that will be all cross-cut with each other. Firstly, we will randomly assign households to different subsidy levels. Some households will receive a 50% subsidy in the form of a buy-one-get-one free bonus (if the family enrolls, they only have to pay the premium for half of the family members), some households will receive a full subsidy and some households will receive a no inpatient bonus (the household has to cover the full premium for all family members but, if they have no inpatient claim for a year, receive 50% of the premiums they paid back). The remaining households will receive no subsidy. Secondly, we will offer to half of the households the possibility to register on the spot as opposed to following the procedure that requires them to go to a central office. Thirdly, we will test three different information treatments focusing on health care costs of certain illnesses, a waiting period, and potential penalties for non-enrollment. Lastly, we will also vary the length of time for which the subsidy offer is valid (Medan only).
Experimental Design Details
Randomization Method
Randomization done in the field on a computer using CSPro
Randomization Unit
Households
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
2000 households in Medan and 4750 households in Bandung.
Sample size: planned number of observations
2000 households in Medan and 4750 households in Bandung.
Sample size (or number of clusters) by treatment arms
The study has 4 main treatments to test. All treatments will be cross-cut one to another. In Medan, all treatments had equal probability. In Bandung a higher number of households (2500) will be assigned to the treatment group; 1000 households will be assigned to the half subsidy and no inpatient bonus and 500 households will be assigned to the full subsidy treatment. All other treatments have equal probability.
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
The analysis for this study will be divided into two main parts. First, we estimate the effect that the different incentives schemes have on enrollment into JKN. This uses the entire sample. Second, we study whether different incentives schemes attract different individuals and differentially affect individuals’ health seeking behavior. To perform this analysis, we compare families which took up across different treatments. This part of the analysis restricts the sample to taker-uppers. As a result, we perform two sets of power calculations. Power calculations for enrollment use the entire sample and, in particular, the size of groups assigned to each treatment. Given the number of individuals assigned to each group, we calculate a predicted number of taker-uppers for each treatment group based on the enrollment rates that we saw in the previous pilot. We then use these group sizes for all power calculations related to adverse selection or health seeking behavior. All power calculations calculate MDEs based on control means and standard deviations from the previous pilot. We assume β=0.80 and α=0.05. All MDEs are expressed as fraction of control group standard deviation. Overall, for the Bandung pilot, MDEs for the enrollment regressions are in the order of 0.10 to 0.20 standard deviations, whereas MDEs for the adverse selection regressions are in the order of 0.20 to 0.40 standard deviations. The tables (uploaded separately) report the assumptions on take-up rates and MDEs for the different treatment arms and main outcomes of interest. Note that these MDEs are lower bounds for the onsite registration treatment and for the half-subsidy and full-subsidy treatment. These treatments are the same in the Medan and in the Bandung pilot, which means that we will be able to pool together the data from the two sites in the analysis.
Supporting Documents and Materials

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information
IRB

Institutional Review Boards (IRBs)

IRB Name
MIT COUHES
IRB Approval Date
2014-06-19
IRB Approval Number
1406006432

Post-Trial

Post Trial Information

Study Withdrawal

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Intervention

Is the intervention completed?
No
Data Collection Complete
Data Publication

Data Publication

Is public data available?
No

Program Files

Program Files
Reports, Papers & Other Materials

Relevant Paper(s)

Reports & Other Materials