Study on Advance Care Directives

How to cite

Rameas, Peyton; Viscusi, Kip; Barrett, Mackenzi; Cronin, Colton; Schoonover, Sydney. 2026. "Study on Advance Care Directives." AEA RCT Registry. AEARCTR-0017148. https://doi.org/10.1257/rct.17148-1.0

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

Peyton Rameas* , Kip Viscusi , Mackenzi Barrett , Colton Cronin , Sydney Schoonover

* Primary investigator

Registry ID
AEARCTR-0017148
DOI
https://doi.org/10.1257/rct.17148-1.0
First registered
Last updated
Status
In development
Trial dates
2026-01-28 – 2026-02-04
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/17148

Abstract

Advance care directives are written legal documents in which individuals specify what types of medical care they would like to receive in the future should they lose the ability to make decisions for themselves. Behavioral economics suggests that individuals’ responses are susceptible to framing effects--or the manner in which information is presented--rather than the objective content of the information. It is possible that individuals’ responses to advance care directives are impacted by ordering effects, meaning that individuals’ responses vary with changes in the specific sequence in which questions and options are presented. However, there is no experimental evidence that isolates the role of question framing and ordering in the advance directive context, leaving a gap in the understanding of how question ordering may impact responses on these forms.

The purpose of this study is to examine whether individuals’ preferences for medical care and treatment as communicated on an advance directive form are impacted by the ordering of options and conditions. The primary objective is to determine whether participants are more likely to select that they would not prefer to receive certain treatments if the possible conditions are presented in differing orders of seriousness. To that end, we test whether individuals exhibit differences in medical care and treatment preferences, as identified on an advance care directive form, based on the order in which conditions and treatment options are presented to them. By varying the order in which options are presented to participants while holding the content of the questions constant, we seek to provide evidence directly relevant to improving individuals’ communication of their healthcare preferences.

The survey leverages four variations where participants will be randomly assigned one of the four variations. First, participants will be asked whether certain health conditions are acceptable or unacceptable. Half of the variations include an ordering where end-stage illness is the first condition while the other half lists permanent unconsciousness first. Next, participants will view treatments and be able to choose which treatments they would like to be given if they end up in a condition that is unacceptable to them. Half of the variations will allow a piece-wise choice, and the other half applies to all unacceptable conditions. After that section, there is a rationality test to see if the preferences for conditions are transitive.

We predict that respondents are more likely to indicate that they would not prefer to receive treatment if the possible conditions are presented in descending order of seriousness. Additionally, respondents are likely to have variable preferences for treatment when presented with treatment options for each condition as compared to when treatment options are presented in a single question that pertains to all conditions.

Experimental design

The experimental design is a survey with four variations. Participants will be randomly assigned to one of the four variations. Because participants are assigned to only one condition, this is a between-subjects design. The variations are the result of a 2x2 matrix with respect to advanced healthcare directives–that is, there are two treatment conditions that are being varied. Specifically, we are comparing the results of variations 1/2 to variations 3/4 and the results of variation 1 to variation 3 and of variation 2 to variation 4. Participants do not know to which of the four treatment variations they are assigned.

Design details

Randomization method
Randomization done in the survey software by a computer
Randomization unit
Individual
Clusters
1,000 adults
Observations
1,000 adults
Arms
250 adults
Primary outcomes
We are collecting three key outcome variables: preferences over different health conditions, preferences over different treatment options, and preferences for allowing healthcare providers to choose treatments based on a benefit-cost analysis. We will analyze these outcomes by calculating the frequencies and means of respondents’ preferences and by using regression analysis to identify relevant determinants of participants preferences.

Cited by

No citation data available for this registration yet.

Keywords