The Effect of Behavioral Interventions on Enrollment and Adverse Selection in Health Insurance Marketplaces

Last registered on April 27, 2020
View Trial History

Pre-Trial

Trial Information

General Information

Title
The Effect of Behavioral Interventions on Enrollment and Adverse Selection in Health Insurance Marketplaces
RCT ID
AEARCTR-0003100
Initial registration date
June 29, 2018
Last updated
April 27, 2020, 4:36 PM EDT

Locations

Country
United States of America
Region
Massachusetts

Primary Investigator

Name
Timothy Layton
Affiliation
Harvard Medical School
Contact Primary Investigator

Other Primary Investigator(s)

PI Name
Adam Sacarny
PI Affiliation
Columbia Mailman School of Public Health
Contact Investigator
PI Name
Adrianna McIntyre
PI Affiliation
Harvard University
Contact Investigator
PI Name
Keith Ericson
PI Affiliation
Boston University Questrom School of Business
Contact Investigator

Additional Trial Information

Status
In development
Start date
2018-07-03
End date
2021-04-30
Keywords
Health
Additional Keywords
Health Insurance, Behavioral, Random Assignment
JEL code(s)
G22, H75, I18, I13
Secondary IDs
Abstract
Incomplete take-up of free and low-cost health insurance remains a puzzle. Failure to enroll in coverage has consequences for the uninsured as well as the health care providers and state budgets that bear the costs of uncompensated care. Moreover, if the marginal enrollee is healthier on average, increasing enrollment may improve competition and reduce premiums in the market by improving the risk pool. Research from other contexts suggests that behavioral frictions or mistakes may play an important role in determining whether households complete the enrollment process. We propose to conduct a randomized trial to test “nudges” (letters sent by postal mail) that could increase enrollment in the Massachusetts Health Connector, the state marketplace through which eligible residents can obtain subsidized private coverage. Nudges will target households that were determined eligible for financial assistance but fail to enroll in an insurance plan. Our study design will employ three treatment arms: a generic reminder letter, a personalized reminder letter, and a personalized reminder letter with a simplified (check-the-box) enrollment option. This will allow us to determine the mechanisms by which these nudges are (or aren’t) effective at (1) inducing enrollment among uninsured consumers and (2) shifting the risk pool.
External Link(s)

Registration Citation

Citation
Ericson, Keith et al. 2020. "The Effect of Behavioral Interventions on Enrollment and Adverse Selection in Health Insurance Marketplaces." AEA RCT Registry. April 27. https://doi.org/10.1257/rct.3100-2.1
Former Citation
Ericson, Keith et al. 2020. "The Effect of Behavioral Interventions on Enrollment and Adverse Selection in Health Insurance Marketplaces." AEA RCT Registry. April 27. https://www.socialscienceregistry.org/trials/3100/history/67050
Sponsors & Partners

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

Request Information
Experimental Details

Interventions

Intervention(s)
Treatment in this study will consist of letters sent (by postal mail) to households that have been determined eligible for marketplace coverage. Households in the study population will be randomized to one of four arms:

Arm 1, Control arm: No postal mail intervention.

Arm 2, Generic reminder letter: Households will receive letters that remind them of their eligibility for insurance through the Health Connector. These letters will not contain any personalized information about the specific plans or size of subsidies available to the household, nor a simplified enrollment option.

Arm 3, Personalized reminder letters: These letters will be generated to remind households of their insurance plan options (consisting of a maximum of only 4 options for the subsidized population we will target) through the Health Connector, including after-subsidy premium costs (and relevant cost-sharing information).

Arm 4, Streamlined-enrollment letter: These letters will be generated to reflect a household’s insurance plan options (consisting of a maximum of only 4 options for the subsidized population we will target), including after-subsidy premium costs (and relevant cost-sharing information). These letters would allow potential enrollees to enroll in coverage by simply checking a box for their preferred plan and returning the letter in an included prepaid envelope.
Intervention Start Date
2018-07-03
Intervention End Date
2019-07-01

Primary Outcomes

Primary Outcomes (end points)
Enrollment in coverage via the Connector.
Primary Outcomes (explanation)

Secondary Outcomes

Secondary Outcomes (end points)
Health risk scores and variables describing utilization and spending
Secondary Outcomes (explanation)

Experimental Design

Experimental Design
Inclusion criteria: To be included in this study, households must have been deemed eligible for subsidized coverage via the ConnectorCare program (the health insurance program for consumers below 300% FPL in Massachusetts) but who did not immediately effectuate coverage. The household must have only one adult eligible for ConnectorCare and must have a complete address on file with the Massachusetts Health Connector.

We include three types of intervention letters to distinguish between the mechanisms by which letters may change behavior. This randomized study design allows us to distinguish between the effect of any reminder letter (testing frictions related to memory and procrastination), a reminder letter with salient, personalized information about financial assistance (testing for information frictions), and a letter that provides a simplified enrollment option (testing for behavioral frictions caused by complexity of the enrollment process).

We will first use a regression framework to determine whether our interventions had an effect on our primary outcome, enrollment in health insurance coverage via the Massachusetts Health Connector. In addition to estimating average treatment effects for the full study sample, we will determine for which sub-groups the interventions were most effective.
Experimental Design Details
Randomization Method
Simple randomization done in office by a computer
Randomization Unit
Household
Was the treatment clustered?
No

Experiment Characteristics

Sample size: planned number of clusters
N/A, the treatment is not clustered.
Sample size: planned number of observations
Initially planned to enroll 43,000 households (final enrollment was 62,457 households)
Sample size (or number of clusters) by treatment arms
We intended to enroll approximately 10,750 households per arm (we actually enrolled approximately 15,600 per arm)
Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
1.4 percentage points (change in health insurance coverage)
Supporting Documents and Materials
IRB

Institutional Review Boards (IRBs)

IRB Name
Institutional Review Board (IRB) of the Harvard Faculty of Medicine
IRB Approval Date
2018-05-02
IRB Approval Number
IRB18-0461
Analysis Plan

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

Request Information

Post-Trial

Post Trial Information

Study Withdrawal

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

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