Social Media as an Early Proxy for Social Distancing Indicated by the COVID-19 Reproduction Number: Observational Study (Preprint)

• DOI: 10.2196/preprints.21340
• Publication Date: 2020-06-12T14:58:01Z
• Authors: Joseph Younis; Harvy Freitag; Jeremy S Ruthberg; Jonathan P Romanes; Craig Nielsen; Neil Mehta
• Keywords: Social distance; Preprint

<sec> <title>BACKGROUND</title> The magnitude and time course of the COVID-19 epidemic in United States depends on early interventions to reduce basic reproductive number below 1. It is imperative, then, develop methods actively assess where quarantine measures such as social distancing may be deficient suppress those potential resurgence nodes possible. </sec> <title>OBJECTIVE</title> We ask if media an indicator public by investigating its correlation with time-varying reproduction (R&lt;sub&gt;t&lt;/sub&gt;) compared mobility estimates reported from Google Apple Maps. <title>METHODS</title> In this observational study, estimated R&lt;sub&gt;t&lt;/sub&gt; was obtained for period between March 5 April 5, 2020, using EpiEstim package. Social activity assessed queries “social distancing” or “#socialdistancing” Trends, Instagram, Twitter, Maps data. Cross-correlations were performed States. used Pearson correlations coefficient determination (ρ) significance set &lt;i&gt;P&lt;/i&gt;&amp;lt;.05. <title>RESULTS</title> Negative found search interest (&lt;i&gt;P&lt;/i&gt;&amp;lt;.001), state-specific 9 states highest cases (&lt;i&gt;P&lt;/i&gt;&amp;lt;.001); most experienced a delay varying 3-8 days before reaching significance. A negative seen at 4-day start Instagram hashtag 6 Twitter (&lt;i&gt;P&lt;/i&gt;&amp;lt;.001). Significant manifest earlier Maps, peaks –6 –4 days. Meanwhile, changes correlated best –2 +1 day workplace grocery/pharmacy, respectively. <title>CONCLUSIONS</title> Our study demonstrates use epidemiological tools assessment during pandemic. Their rise peak correlative strength when provide proactive insight into whether efforts are sufficiently enacted. Whether proves valuable creation more accurate assessments uncertain due limitations. These limitations include biased sample that internet literate access, which covary socioeconomic status, education, geography, age, subtotal mentions distancing. Future studies should focus how reactions change epidemic, well conversion behavior actual physical behavior.