Chicago Public Health Department Social Media Communications on Twitter During the COVID-19 Pandemic and the Mpox Epidemic: Cross-Sectional Content Analysis.

Background: Protecting public health depends on the effective communication of health-related information to the public, especially during public health emergencies. Health communication campaigns traditionally relied on mass media outlets but increasingly incorporate social media platforms. This paper presents a content analysis of original communications posted to X (formerly Twitter) by the Chicago Department of Public Health (CDPH) from May 1, 2022, to April 30, 2023, a time characterized by the concurrent COVID-19 pandemic and mpox epidemic public health emergencies. Objective: This paper aims to investigate: (1) what information was being discussed by CDPH, (2) how information was presented, (3) the nature of communications, and (4) the impact of communication attributes on engagement. Secondary objectives included investigating the correlation between mpox cases and mpox-related communications and using a bioethical risk communication framework to characterize the intent of mpox-related communications. Methods: Original communications posted by the CDPH from May 1, 2022, to April 30, 2023, were collected. Communication attributes including the date and time of posting, the communication text, accompanying media, text in image-based accompanying media, and the language of the text were extracted at the time of collection. A total of 2 researchers independently reviewed the communications using a coding schema that was developed to codify the health topics and the bioethical framework to codify the intent of mpox-related communications. Percent agreement and Cohen kappa were used to establish intercoder reliability. Negative binomial regressions were used to investigate the impact of attributes on public engagement. Spearman rank correlation coefficients were used to measure the strength and direction of the correlation between the weekly number of mpox cases and the number of weekly mpox-related communications. Results: A total of 1105 original communications were posted, a majority of which discussed communicable diseases (n=539, 51.8%), were posted in English (n=801, 72.5%), during the standard workday (n=1003, 90.8%), and with additional media (n=839, 75.9%). All communications were proactive in nature, and none directly responded to other accounts. Regression analysis suggested that communications posted during the workday (event rate ratio [ERR]=1.25) and those with images (ERR=2.59) or videos (ERR=2.40) received significantly higher levels of engagement, as did those discussing maternal and child health (ERR=2.35), mental health (ERR=1.48), and substance use (ERR=1.61). Communications discussing communicable diseases were not among the health topics with higher levels of engagement. Communications posted exclusively in Spanish received significantly lower levels of engagement (ERR=0.67). In addition, mpox-related communications were positively correlated with reported mpox cases at a significant level, and most mpox-related communications sought to inform the public (n=60, 60.6%), as opposed to influence behavior (n=39, 39.4%). Conclusions: Social media platforms can represent valuable tools for risk communication during public health emergencies but should supplement other dissemination vehicles that may be more appropriate for communicating nuanced information, achieving behavior change, and reaching certain demographic groups.