Randomized controlled trials in the shoulder arthroplasty literature rarely include key demographic and socioeconomic patient data

Abstract

Background: It is well known that socioeconomic and demographic variables can greatly affect health outcomes. Previous studies across medical and surgical subspecialties have demonstrated that these variables are frequently under-reported in randomized controlled trials (RCTs). No such study currently exists that examines rates of reporting of sociodemographic variables in shoulder arthroplasty RCTs. This study aims to present these rates and explore the impact of failing to report socioeconomic variables in shoulder arthroplasty RCTs.

Methods: The PubMed database was queried for the term shoulder arthroplasty. Sixty-five RCTs from the past 10 years were identified for inclusion from 5 high-impact orthopedic surgery journals. Each RCT was analyzed for patient age and sex or gender as well as the following sociodemographic variables: race, ethnicity, insurance status, income, work status, and education. It was also noted whether each of the above variable was mentioned in the results section of the article. Data were presented in a descriptive fashion as well as analyzed using χ² and Fisher exact tests where appropriate.

Results: From 2014 to 2023, the 65 shoulder arthroplasty RCTs published reported age in 40 of 65 (61.5%) in their results sections and 61 of 65 (93.8%) in any section. Sex or gender was reported in 27 of 65 (41.5%) in their results sections and 61 of 65 (93.8%) in any section. No articles included any sociodemographic variables in the results section. Reporting rates for sociodemographic variables in any section were as follows: race, 6 of 65 (9.2%); ethnicity, 5 of 65 (7.7%); work status, 4 of 65 (6.2%); and insurance status, 1 of 65 (1.5%). No studies included income or education of the enrolled patients. There was no difference in reporting sociodemographic variables by journal (P = .45) or by year of publication (P = .57). However, no study prior to 2020 included any sociodemographic variable (0 of 27, 0%), whereas from 2020 onward 6 studies included at least 1 (6 of 38, 15.8%). Sociodemographic variables were reported significantly less frequently than age and sex or gender (P = .001).

Discussion: Our study found sociodemographic variables are rarely reported in shoulder arthroplasty RCTs, whereas age and sex or gender are reported with great frequency. In order to understand the results of shoulder arthroplasty RCTs, apply their findings to the care of our patients, and address health disparities, we must ensure these studies include patient sociodemographic data.

Keywords: Shoulder; income; insurance; race; randomized controlled trial; shoulder arthroplasty; work.