Socioeconomic bias in physician decision-making? An experimental design

How to cite

Dackehag, Margareta; Kåregård, Magnus; Wengström, Erik; Gerdtham, Ulf; Anell, Anders; Beckman, Anders; Dahlin, Sofia. 2019. "Socioeconomic bias in physician decision-making? An experimental design." AEA RCT Registry. AEARCTR-0004673. https://doi.org/10.1257/rct.4673-1.0

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

Margareta Dackehag* , Magnus Kåregård , Erik Wengström , Ulf Gerdtham , Anders Anell , Anders Beckman , Sofia Dahlin

* Primary investigator

Registry ID
AEARCTR-0004673
DOI
https://doi.org/10.1257/rct.4673-1.0
First registered
Last updated
Status
On going
Trial dates
2019-09-01 – 2020-07-01
Countries
Private
Registry record
http://www.socialscienceregistry.org/trials/4673

Abstract

Despite the central goal of providing medical treatment based on need there are well-known and persistent socioeconomic inequalities in all developed healthcare systems. Based on register and survey data, recent studies indicate that part of the problem relates to physician behaviour; physicians treat patients with equal care need differently depending on patients' socioeconomic status (SES). However, due to data limitations further analysis, using other methods, is required to establish a socioeconomic bias in healthcare provision and to study the underlying mechanisms. The objective of this project is to investigate whether and how patient SES influences physician decision-making, using economic experiments based on video-recordings of individuals describing their real health problems in a standardised, consultation-like setting. The “video patients” suffer from similar health problems but have different SES. The project consists of two economic experiments, where students of and graduates from medical professional programs participate. In the first experiment we investigate the alignment of SES-perceptions among a first group of experiment participants, who are asked to categorize patients according to SES after viewing individual videos. In the second experiment we investigate whether there is a socioeconomic bias in physician decision-making, by letting a second participant group assess the healthcare need of video patients.

Experimental design

The material (video recordings) for the experiments has been collected. All recruited “video patients” have problems with musculoskeletal pain. Administrative classification of SES and healthcare need is based on the information that the video patients registered when signing up for the study. They are classified according to SES and similar healthcare need has been established. The assessment of pain and healthcare need of the video patients is based on the Örebro Musculoskeletal Pain Screening Questionnaire (ÖMPSQ) evaluation template and the expert advice of the two GPs in the research group. There are 3-4 video patients per SES group (two SES groups: low income and low education, high income and high education), in total 6-8 video patients.

We recruit experiment participants among students of and graduates from medical professional programs. Both experiments will take place in computer laboratories. In both experiments we apply a within-subject design: every participant will watch the video recordings in random order and after each video make an assessment. In Experiment 1 the participant assigns the patient to a SES-group with respect to income and education. In addition, the participant rates the patients on bipolar attitude scales. In the second experiment the participant assesses health care need by allocating resources on a scale from 1 (watchful waiting) to 5 (intense and immediate treatment) for every patient. The main treatment in both experiments is a coordination game and rewards the participants for choosing the SES classification or medical treatment level they believe is the preferred choice of their (participant) peers, increasing the possibility to measure implicit socioeconomic bias.

Design details

Randomization method
Randomization done in office by a computer.
Randomization unit
Randomization of the order in which video recordings are shown to individual participants.
Clusters
60-120 students/graduates
Observations
60-120 students/graduates
Arms
Within-subject design: 60-120 students/graduates in both treatments
Primary outcomes
Experiment 1: assessments of patients’ socioeconomic status Experiment 2: assessments of patients’ healthcare need

Cited by

No citation data available for this registration yet.

Keywords