Abstract
Surveys can be a critical tool in monitoring public health during emergencies. Existing surveillance systems may provide timely reporting of cases and deaths associated with diseases. However, during COVID-19 they did not provide accurate information on the number of cases with virus-related symptoms, testing and treatment seeking. In fact, the surge of potentially infected individuals seeking diagnosis, testing and treatment represented a serious but largely unmeasured dimension of the crisis. This study aimed to evaluate the potential value of monitoring health, attitudinal, and behavioral dimensions that are not included in current U.S. disease surveillance systems during population health emergencies. Additionally, it seeks to demonstrate the feasibility of designing and implementing a low-cost, rapid-turnaround health and behavioral monitoring system when comparable data from existing surveillance systems are unavailable. From March through November 2020, we conducted national surveys with approximately 1,000 interviews each month with Census-balanced samples from a large national commercial panel. These surveys employed replicate national samples drawn from all 50 US states and the District of Columbia. A total of 9,200 interviews, averaging about 20 min in length, were completed over the course of the nine months of fielding. Nearly a quarter of respondents (22%) reported they had been sick for three days or longer since January with what might be COVID. Respondents were questioned about their symptoms, whether they had seen a doctor, had a confirmatory test for the COVID virus, and test results. Approximately one in ten respondents were currently experiencing COVID-like symptoms each month (95% CI: 10.7-12.0%). These numbers dwarf the 0.3% in April and 3.6% in November who had ever had a COVID positive test result. Moreover, 42% of these symptomatic adults sought medical care or testing, increasing strains on the health care system,. Although surveys may not be needed to estimate diagnosed cases, hospitalizations, or deaths, they can provide the missing data on symptomatic cases in the population, the proportion seeking medical care, ability to obtain a confirmatory test, and reasons for not seeking care or testing. This study demonstrates the ability of surveys to provide such information in a timely fashion, which could be replicated in other countries.