Population-based estimates of COVID-19 period prevalence and cumulative monthly incidence in New York City: A comparison of estimates from three surveys, July–August 2020

Abstract

Background Diagnosis-based surveillance of COVID-19 underestimates COVID-19 burden. Questions about COVID-19-consistent symptoms were added to three population-based surveys to obtain representative estimates of COVID-19 period prevalence and monthly cumulative incidence. Objective To evaluate if estimates of COVID-19 period prevalence and cumulative monthly incidence differed when collected from surveys with different sampling frames and modes of administration. Methods Data were collected from adult New York City (NYC) residents via the Community Health Survey (CHS) (sampling frame: random digit dial with dual landline and cellphone frame; mode: phone) and the Citywide Mobility Survey (CMS) (sampling frame: probabilistically selected panel; mode: online) in July 2020 and via CHS and Healthy NYC (sampling frame: probabilistically selected panel; mode: online and phone) in August 2020. Persons with COVID-19-like illness (CLI) were identified based on reported symptoms in the past 30 days. To obtain COVID-19 estimates, CLI estimates were adjusted by the proportion of laboratory-confirmed SARS-CoV-2 infections among citywide emergency department CLI visits in which patients received SARS-CoV-2 testing. We used t-tests to compare estimated CLI period prevalence in July 2020 between CHS and CMS and CLI period prevalence and cumulative monthly incidence in August 2020 between CHS and Healthy NYC. Results CLI period prevalence was similar between CHS and CMS during July (12.2% vs. 9.9%, respectively, p=0.511); COVID-19 period prevalence was 1.7% and 1.3%, respectively. In contrast, CLI period prevalence was higher per Healthy NYC during August 2020 than CHS (18.1% vs. 11.3%, p=0.014); COVID-19 period prevalence was 0.7% and 0.4%, respectively. CLI cumulative monthly incidence in August was similar (5.7% and 4.0%, respectively; p=0.246) in both surveys. Conclusions Because estimates of CLI were not consistently different by sampling frame or mode of administration, additional research to understand the cause of differences between CHS and Healthy NYC can support use of symptom-based surveillance to monitor COVID-19 trends.