Real-time pandemic surveillance using hospital admissions and mobility data

• DOI: 10.1073/pnas.2111870119
• Publication Date: 2022-02-01T21:26:04Z
• Authors: Spencer J. Fox; Michael Lachmann; Mauricio Tec; Remy Pasco; Spencer Woody; Zhanwei Du; Xutong Wang; Tanvi A. Ingle; Emily Javan; Maytal Dahan; Kelly Gaither; Mark E. Escott; Stephen I. Adler; S. Claiborne John...; James G. Scott; Lauren Ancel Meyers
• Keywords: Pandemic; Lagging; Demographics; Public health surveillance

Forecasting the burden of COVID-19 has been impeded by limitations in data, with case reporting biased testing practices, death counts lagging far behind infections, and hospital census reflecting time-varying patient access, admission criteria, demographics. Here, we show that admissions coupled mobility data can reliably predict severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) transmission rates healthcare demand. Using a forecasting model guided mitigation policies Austin, TX, estimate local reproduction number had an initial 7-d average 5.8 (95% credible interval [CrI]: 3.6 to 7.9) reached low 0.65 CrI: 0.52 0.77) after summer 2020 surge. Estimated detection ranged from 17.2% 11.8 22.1%) at outset high 70% 64 80%) January 2021, infection prevalence remained above 0.1% between April March 1, peaking 0.8% (0.7-0.9%) early 2021. As precautionary behaviors increased safety public spaces, relationship weakened. We mobility-associated was 62% 52 68%) lower February 2021 compared 2020. In retrospective comparison, 95% CrIs our 2, 3 wk ahead forecasts contained 93.6%, 89.9%, 87.7% reported respectively. Developed task force including scientists, health officials, policy makers, executives, this project needs US cities.