# Development of a Medical Social Media Ethics Scale and Assessment of #IRad, #CardioTwitter, and #MedTwitter Posts: Mixed Methods Study

## TL;DR

This study created a scale to evaluate ethical issues in medical social media posts and found that while rare, these issues vary across different medical communities.

## Contribution

The paper introduces a new scale for assessing ethical issues in medical social media and applies it to three distinct medical Twitter communities.

## Key findings

- The developed scale showed good interrater agreement and correlation.
- Ethical issues were rare overall but varied by hashtag, with #IRad showing higher conflict of interest concerns.
- Specialty-specific hashtags had more patient privacy and conflict of interest issues compared to #MedTwitter.

## Abstract

Social media posts by clinicians are not bound by the same rules as peer-reviewed publications, raising ethical concerns that have not been extensively characterized or quantified.

We aim to develop a scale to assess ethical issues on medical social media (SoMe) and use it to determine the prevalence of these issues among posts with 3 different hashtags: #MedTwitter, #IRad, and #CardioTwitter.

A scale was developed based on previous descriptions of professionalism and validated via semistructured cognitive interviewing with a sample of 11 clinicians and trainees, interrater agreement, and correlation of 100 posts. The final scale assessed social media posts in 6 domains. This was used to analyze 1500 Twitter posts, 500 each from the 3 hashtags. Analysis of posts was limited to original Twitter posts in English made by health care professionals in North America. The prevalence of potential issues was determined using descriptive statistics and compared across hashtags using the Fisher exact and χ2 tests with Yates correction.

The final scale was considered reflective of potential ethical issues of SoMe by participants. There was good interrater agreement (Cohen κ=0.620, P<.01) and moderate to strong positive interrater correlation (=0.602, P<.001). The 6 scale domains showed minimal to no interrelation (Cronbach α=0.206). Ethical concerns across all hashtags had a prevalence of 1.5% or less except the conflict of interest concerns on #IRad, which had a prevalence of 3.6% (n=18). Compared to #MedTwitter, posts with specialty-specific hashtags had more patient privacy and conflict of interest concerns.

The SoMe professionalism scale we developed reliably reflects potential ethical issues. Ethical issues on SoMe are rare but important and vary in prevalence across medical communities.