Intervening with Consumers to Improve Choices on Health Insurance Marketplaces

How to cite

Sacarny, Adam; Kingsdale, Jon; Layton, Timothy; Ericson, Keith. 2016. "Intervening with Consumers to Improve Choices on Health Insurance Marketplaces." AEA RCT Registry. AEARCTR-0000982. https://doi.org/10.1257/rct.982-3.0

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Adam Sacarny* , Jon Kingsdale , Timothy Layton , Keith Ericson

* Primary investigator

Registry ID
AEARCTR-0000982
DOI
https://doi.org/10.1257/rct.982-3.0
First registered
Last updated
Status
Completed
Trial dates
2015-11-20 – 2016-12-31
Countries
United States of America (Colorado)
Registry record
http://www.socialscienceregistry.org/trials/982

Abstract

The Affordable Care Act dramatically expanded the use of regulated marketplaces to provide individuals with health insurance coverage. The efficiency of these marketplaces depends on the ability of consumers to choose plans that reflect their preferences, but inattention and information frictions may inhibit optimal choices. This study is a 3-arm randomized intervention to test for the presence of frictions in plan choice and to see if certain strategies to communicate with consumers can reduce these frictions. The implementing partner is the Colorado ACA marketplace, Connect for Health Colorado. In the study, we send simple, salient information about plan premiums to re-enrolling consumers. One study group receives e-mails and letters containing generic information about the potential premium savings from shopping, while another group is sent similar but personalized materials that project their savings based on their previous year’s income and household structure. A third group receives no intervention materials and acts as a control. We will track the effects of these letters on shopping behavior using administrative data from the marketplace. We will test the effect of the generic and personalized information on whether an individual switches her plan, how much she pays in premiums, and her elasticity of plan choice with respect to premiums. The results of this study will provide reduced form evidence on the size of behavioral frictions in health insurance relating to plan premiums. The study will also provide evidence on inexpensive, effective strategies for Marketplaces and policymakers to help consumers select health plans that more closely reflect their preferences.

Intervention

"As-usual" intervention: No additional materials were sent to the members.

Control Intervention: We send two generic letters and two e-mails telling individuals that they should actively shop for plans on the marketplace. The materials do not contain any information that is specific to the recipient, and will be based on standard communications templates that the marketplace already uses.

Premium Savings Intervention: We modify the generic letters and e-mails to include information about the maximum the individual can save compared to her current plan by switching plans on the marketplace, and send the customized communications. The savings calculation will be adjusted for the subsidies that the consumer can expect to receive based on the most recent information she has provided to the marketplace. The savings are presented as the difference in price relative to what the member would pay in her current plan, if she were to stay enrolled in it in 2016.

Experimental design

The set of marketplace members who were eligible for automatic re-enrollment for 2016 was identified by the marketplace. We then limited this sample to members who could potentially save on premiums by switching to another plan in their metal tier. The remaining members became the study sample, and they were randomly allocated to three arms:
1. No intervention
2. Control intervention
3. Premium Savings intervention

Design details

Randomization method
Randomization done by computer (in Stata)
Randomization unit
Connect for Health Colorado account (i.e. household) level
Clusters
15,000
Observations
15,000
Arms
5,000 per arm
Primary outcomes
Active shopping on the exchange Switching to another plan The realized premium of the member's 2016 plan The realized cost-sharing of the member's 2016 plan

Cited by

No citation data available for this registration yet.

Keywords