Improving Risk Assessment and Treatment Choice in Medical Decision Making

How to cite

Pérez, Maria Hofbauer; Winter, Joachim; Wuppermann, Amelie; Hartung, Corinna; Manski, Charles. 2022. "Improving Risk Assessment and Treatment Choice in Medical Decision Making." AEA RCT Registry. AEARCTR-0009589. https://doi.org/10.1257/rct.9589-1.0

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

Maria Hofbauer Pérez* , Joachim Winter , Amelie Wuppermann , Corinna Hartung , Charles Manski

* Primary investigator

Registry ID
AEARCTR-0009589
DOI
https://doi.org/10.1257/rct.9589-1.0
First registered
Last updated
Status
In development
Trial dates
2022-07-01 – 2022-07-31
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/9589

Abstract

Physicians often face situations in which they must predict patients’ health outcomes under uncertainty. Existing decision tools are based on prediction models derived from clinical data, but typically include only a small set of known risk factors. In clinical practice, physicians often observe additional risk factors and patient characteristics which may improve their risk assessment. By performing an online experiment, we
explore a new risk assessment tool that allows physicians to combine their clinical experience and their assessment of a patient’s risk factors with the predictions of a decision tool in a structured way. In doing so, we aim to investigate the behavior of physicians in risk assessment and treatment selection.

Experimental design

Physicians (cardiologists and general practitioners) are randomly drawn into one of four treatment groups that differ with respect to the provided decision tool. After subjects are randomly assigned to a treatment group, the survey starts with a questionnaire eliciting information on subjects’ socio-demographic, educational and job characteristics. In the subsequent part of the survey, subjects first assess a vignette’s risk and then, make treatment recommendations. The exact procedure depends on the treatment group. The survey ends with a short questionnaire asking subjects to report about prior knowledge and use of a specific risk assessment tool partly provided in the survey.

Design details

Randomization method
randomization is done within cardiologists and general practicioners (different surveys); it is done automatically by the survey software
Randomization unit
Individual physicians
Clusters
400 physicians (200 cardiologists, 200 general practicioners)
Observations
400 physicians (200 cardiologists, 200 general practicioners)
Arms
100 per treatment group (4 treatment groups)
Primary outcomes
accuracy in risk prediction, confidence in risk assessment and treatment choice

Cited by

No citation data available for this registration yet.

Keywords