Exposure notification system activity as a leading indicator for SARS-COV-2 caseload forecasting

PLOS ONE RESEARCH ARTICLE Exposure notification system activity as a leading indicator for SARS-COV-2 caseload forecasting Eliah Aronoff-Spencer ID1☯*, Sepideh Mazrouee ID1☯, Rishi Graham1, Mark S. Handcock ID2, Kevin Nguyen ID3,4, Camille Nebeker3, Mohsen Malekinejad5,6, Christopher A. Longhurst ID4 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 1 School of Medicine, Division of Infectious Diseases and Global Public Health, University of California San Diego, La Jolla, CA, United States of America, 2 University of California Los Angeles, Los Angeles, CA, United States of America, 3 Herbert Wertheim School of Public Health and Longevity Sciences, University of California San Diego, La Jolla, CA, United States of America, 4 University of California San Diego Health, San Diego, CA, United States of America, 5 California Department of Public Health, Sacramento, CA, United States of America, 6 Department of Epidemiology and Biostatistics, University of California, San Francisco, San Francisco, CA, United States of America ☯ These authors contributed equally to this work. * OPEN ACCESS Citation: Aronoff-Spencer E, Mazrouee S, Graham R, Handcock MS, Nguyen K, Nebeker C, et al. (2023) Exposure notification system activity as a leading indicator for SARS-COV-2 caseload forecasting. PLoS ONE 18(8): e0287368. https:// doi.org/10.1371/journal.pone.0287368 Editor: Jianguo Wang, China University of Mining and Technology, CHINA Received: July 15, 2022 Abstract Purpose Digital methods to augment traditional contact tracing approaches were developed and deployed globally during the COVID-19 pandemic. These “Exposure Notification (EN)” systems present new opportunities to support public health interventions. To date, there have been attempts to model the impact of such systems, yet no reports have explored the value of real-time system data for predictive epidemiological modeling. Accepted: May 29, 2023 Published: August 18, 2023 Methods Copyright: © 2023 Aronoff-Spencer et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. We investigated the potential to short-term forecast COVID-19 caseloads using data from California’s implementation of the Google Apple Exposure Notification (GAEN) platform, branded as CA Notify. CA Notify is a digital public health intervention leveraging resident’s smartphones for anonymous EN. We extended a published statistical model that uses prior case counts to investigate the possibility of predicting short-term future case counts and then added EN activity to test for improved forecast performance. Additional predictive value was assessed by comparing the pandemic forecasting models with and without EN activity to the actual reported caseloads from 1–7 days in the future. Data Availability Statement: Data cannot be shared publicly because of privacy issues. Data are available from the California Department of Health Institutional Data Access / Ethics Committee (https://canotify.ca.gov/) for researchers who meet the criteria for access to confidential data. The data underlying the results presented in the study are available from: https://www.cdph.ca.gov/ Funding: Funding statement update: California Department of Public Health (CDPH), AGREEMENT NUMBER 20-10777. The funders had no role in Results Observation of time series presents noticeable evidence for temporal association of system activity and caseloads. Incorporating earlier ENs in our model improved prediction of the caseload counts. Using Bayesian inference, we found nonzero influence of EN terms with probability one. Furthermore, we found a reduction in both the mean absolute percentage PLOS ONE | https://doi.org/10.1371/journal.pone.0287368 August 18, 2023 1/8 PLOS ONE study design, data collection and analysis, decision to publish, or preparation of the manuscript Competing interests: The authors have declared that no competing interests exist. Exposure notificaiton indicator for future caseload error and the mean squared prediction error, the latter of at least 5% and up to 32% when using ENs over the model without. Conclusions This preliminary investigation suggests smartphone based ENs can significantly improve the accuracy of short-term forecasting. These predictive models can be readily deployed as local early warning systems to triage resources and interventions. 1. Introduction In the 1530 rhymed account of epidemic disease, Giralamo Fracastoro gave name to the great pox “Syphilis sive morbus Gallicus” (Syphilis or the French Disease) and perhaps the earliest Western account of what is now referred to as contact tracing (CT) [1]. Even as many nations have implemented manual CT programs in response to COVID-19, SARS-COV-2 dynamics and global reach have quickly overwhelmed most traditional approaches, though there are significant bright-spots [2]. To meet this challenge, new solutions that synergize with CT have been developed employing surrogates of contact such as anonymous proximity notification utilizing smartphones; creating digital systems that emulate and potentially go beyond traditional public health practices [3]. These systems have now been deployed globally, yet there remain significant gaps in our understanding of impact. Further, there are no reports demonstrating that privacy preserving strategies that are foundational to the GAEN platform may still allow for epidemiological forecasting or spatiotemporal prediction models. In April 2020, Google and Apple jointly released the Google Apple Exposure Notification (GAEN) API built with the “Private Automated Contact Tracing” (PACT) protocols [4], to scale contact tracing through smartphone-based proximity sensing and ENS [5]. Early evidence supported epidemiological impact [6], leading to reports that modeled EN’s effect in the context of other public health interventions [7, 8]. California leveraged the GAEN platform for its exposure notification system, and launched CA Notify statewide on Dec 10, 2020 [9]. As of this writing, the system has been activated nearly 17million times. With over one year of data, we can now explore the potential contribution of EN data for predictive modeling. 1.2 Evidence before the study There have been significant advances in how non-exposure notification data are used for epidemiological forecasting, however, to date there are no reports employing EN data to predict future caseload counts. Numerous other forecasting models have been deployed that can be categorized into three different model types: i) Susceptible-Exposed-Infected-Recovered (SEIR) or Susceptible-Infected-Recovered (SIR) models [10, 11]; b) Agent-based models [12]; and c) Curve-fitting models [13–15]. The majority of the predictive models use parameters such as number of confirmed cases and deaths, masking guideli (...truncated)