Public health officials had an unprecedented tool for near-instant, widespread communication during the COVID-19 pandemic and mpox epidemic: social media.
Now, one of the first studies of its kind, led by a health policy expert with the Texas A&M University School of Public Health , has found that X (Twitter at the time of the events) brought advantages — as well as disadvantages — in getting the word out.
"For the year that the COVID-19 pandemic and mpox outbreak overlapped, we wanted to understand how health authorities in Chicago used X: the nature of the posts, what information was shared, how it was presented to the public and how all of that impacted public engagement," said Matthew Boyce . "We also looked at the relationship between the number of reported mpox cases and the messages shared on social media."
For their content analysis, published in the Journal of Medical Internet Research, Boyce and colleagues from Georgetown University and CUNY Graduate School of Public Health & Health Policy analyzed 1,105 original, health-related X posts by the Chicago Department of Public Health (CDPH) from May 2022 to April 2023.
After giving each post an identifier, the team looked at the post's date, time, topic, media (image or video) and language (English, Spanish or bilingual), and noted whether the message was proactive or reactive. Health topics included communicable disease, environmental health and maternal and child health, among others.
On Dec. 15, 2022, X implemented a new policy allowing users to see the number of times communications posted after that date were viewed, and the researchers added this information to track public engagement. They then used a coding system to categorize health topics and the intent of the posts related to mpox, followed by two measurements of intercoder reliability to ensure consistency and accuracy.
To assess the relationship between mpox-related communications and reported mpox cases, the team got data on the weekly reported mpox cases from the CDPH website and used a timeframe of four weeks before the first mpox case was reported and four weeks of no reported cases.
"In terms of COVID-19, four broad findings stand out," Boyce said.
First, the analysis found that the CDPH — like most local and state health departments — used X to engage only in one-way communication during the study period. Two-way dialogues are the essence of social media, however, and this approach could not only enhance public trust levels in these health officials, Boyce noted, but also help counter the spread of misinformation and disinformation.
Next, findings suggested that communications with additional images or videos received significantly higher levels of engagement compared to communications without media. Boyce said health officials might want to include these additional media — although judiciously — for maximum public engagement.
Third, although this study was conducted during two public health emergencies, it found that communications related to COVID-19 or mpox were not the primary topics of X users. Instead, maternal and child health, mental health and substance use received significantly higher levels of public engagement, underscoring the need for strategies to combat "pandemic fatigue" during prolonged public health emergencies.
Finally, the CDPH X posts were only in English and Spanish, while major U.S. cities are home to speakers of many other languages. Boyce noted, for example, that the Houston Health Department's X posted the same messaging in Arabic, Chinese, English, Spanish and Vietnamese. Furthermore, the CPDH Spanish-language posts received significantly lower levels of public engagement, highlighting the need for outreach efforts in languages other than English.
The mpox analysis found moderate, positive correlations between mpox-related communications and reported mpox cases. It also found that a majority of the CDPH communications related to mpox sought to inform the public, as opposed to manipulating or coercing people to change behaviors.
"The most notable result was that — while all correlations were positive, of moderate strength and statistically significant — the strongest correlation existed when communications and cases occurred at the same time," Boyce said. "One would expect communications to increase following increases in cases, and the CDPH should be commended for providing accurate information during a critical period of time in the outbreak response."
Boyce noted four limitations in the research. First, it looked only at text in image-based — not video-based — media. Second, ever-changing algorithms make social media difficult to study. Third, the team did not assess the influence of these communications on public behavior. And finally, the findings might not apply to other public health departments, social media platforms or time periods.
"Still, the bottom line is that social media platforms are valuable tools — but should not be the only tool," Boyce said. "Other communications tools and approaches might be more appropriate for communicating nuanced information, convincing people to change their behavior and reaching certain demographic groups during public health emergencies."
By Ann Kellett, Texas A&M University School of Public Health
