Incentivizing Cost Containment in Healthcare

How to cite

Costello, Anna; Eyring, Henry; Floyd, Eric. 2024. "Incentivizing Cost Containment in Healthcare." AEA RCT Registry. AEARCTR-0013418. https://doi.org/10.1257/rct.13418-1.2

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

Anna Costello* , Henry Eyring , Eric Floyd

* Primary investigator

Registry ID
AEARCTR-0013418
DOI
https://doi.org/10.1257/rct.13418-1.2
First registered
Last updated
Status
Completed
Trial dates
2024-04-23 – 2024-12-31
Countries
United States of America (Utah)
Registry record
http://www.socialscienceregistry.org/trials/13418

Abstract

In the context of the rising costs of healthcare, our study seeks to address a key component of input costs in hospital systems: the costs of supplies used in surgical procedures. Physicians are given substantial autonomy in the way they perform medical procedures including the choice of which supplies to use during treatment. This autonomy is an integral part of the organizational design of most hospitals, as it is thought that decentralization puts decision rights in the hands of the most informed agents. However, doctors are not typically incentivized to reduce costs and thus often select expensive supplies even when less expensive options exist. Our project uses several interventions to encourage doctors to select lower cost, but equal quality, supplies. Our interventions include (1) a message containing only information about lower cost alternatives (information-only), (2) information about lower cost alternatives with message indicating that cost savings will be shared with the patient (altruism), (3) information about lower cost alternatives combined with information about how the doctor ranks in terms of their supply costs (social comparison), and (4) information about lower cost alternatives with a message indicating that cost savings will be shared with the doctor and deposited in the departmental budget (financial incentives).

Our treatments are rolled out in a series of emails from each doctor's department head (e.g., surgery department chair or sub-group chair). Our outcomes will measure the doctor's responsiveness to the emails, and whether they click "accept" to the suggested changes. We will subsequently measure total cost supply cost savings realized by our treatments.

Intervention

1) We have one information intervention (which will be the baseline control group), who will get an email message about suggested supply changes that will result in $X of savings.

2) We will have an altruism intervention, which gets information about supply swaps in addition to the following message: As a reminder, much of the cost incurred in the <procedure_name> procedure is charged to your patient. By agreeing to make the changes below, you would pass an estimated savings of $x on to your patients each year.

3) We will have a peer comparison intervention, which gets information about supply swaps in addition to the following message: We performed a cost analysis of all surgeons performing the <procedure_name> procedure, and found that your current DPC is more expensive than ___ (e.g, 9) out of ___ (e.g., 12) of your peers who do this procedure at UUHC. By agreeing to make the changes below, you would improve your cost ranking by X points (e.g., you would move to a rank of 3 out of 12 doctors performing this procedure).

4) We will have a financial incentives intervention, which gets information about supply swaps in addition to the following message: We recognize that your effort in cost reduction is beneficial to the hospital and would like to share that benefit with you. By agreeing to make the changes below, we will pay 30% of the first year’s savings that you generate into your division’s budget.

Experimental design

Treatments will be randomized at the doctor level. Administrators in the given health care system (University of Utah Health Care, and University of Chicago Medicine) will send out the emails to doctors at the same system. The emails will describe the operating room stocking options. Doctors can choose to ignore the emails or to click a link to record whether they accept or decline the stocking option suggested in the email. If doctors do not wish to participate, they can decline to click on that link.

Design details

Randomization method
Randomization by computer.
Randomization unit
Individual (doctor)
Clusters
1 hospital system. Updated (as of a sample size significantly less than the original 100 we disclosed): Our final sample size will be determined by the number of participating hospitals and the participating departments within these hospitals. We have three hospitals we are collaborating with, and each of these hospitals has many departments. Our goal is to maximize the effective sample size we are able to achieve within these three hospitals while balancing institutional constraints; not all departments within a hospital will necessarily agree to participate in our study.
Observations
100 for first wave of treatment (more to follow)
Arms
25 control, 25 for each of three treatments
Primary outcomes
Whether the doctor clicks "accept" to the suggested supply swaps.
Secondary outcomes
Supply cost savings, measured several months after treatment.

Cited by

No citation data available for this registration yet.

Keywords