Engagement With COVID-19 Public Health Measures in the United States: A Cross-sectional Social Media Analysis from June to November 2020

• DOI: 10.2196/26655
• Publication Date: 2021-04-16T22:49:51Z
• Authors: Daisy Massey; Chenxi Huang; Yuan Lu; Alina Cohen; Yahel Oren; Tali Moed; Pini Matzner; Shiwani Mahajan; César Caraballo; Navin Kumar; Yuchen Xue; Qinglan Ding; Rachel Dreyer; Brita Roy; Harlan Krumholz
• Keywords: Cross-sectional study; 2019-20 coronavirus outbreak; Pandemic

Background COVID-19 has continued to spread in the United States and globally. Closely monitoring public engagement perceptions of preventive measures using social media data could provide important information for understanding progress current interventions planning future programs. Objective The aim this study is measure public’s behaviors regarding its effects on daily life during 5 months pandemic. Methods Natural language processing (NLP) algorithms were used identify COVID-19–related unrelated topics over 300 million online sources from June 15 November 15, 2020. Posts sample geotagged by NetBase, a third-party provider, sensitivity positive predictive value both calculated validate classification posts. Each post may have included discussion multiple topics. prevalence these was measured time period compared case rates States. Results final size 9,065,733 posts, 70% which sourced In October November, including mentions related health did not increase as it had September, despite an cases beginning October. Additionally, more focused (n=6,210,255, 69%), with general (n=3,390,139, 37%) (n=1,836,200, 20%). Conclusions There decline mainly States, even increased highest rate since Targeted messaging be needed ensure prevention global vaccination efforts continue.