Summary: Trump administration’s idea to make evidence-based decisions on interventions based on ranking counties by Coronavirus risk requires credible data. Such data does not exist today. Deploying rapid epidemiological survey techniques is the answer to generate such data.
I fully support the President’s effort to use data to assess risk and potentially relax social distancing requirements based on the evidence. See: https://www.wsj.com/articles/trump-administration-to-issue-guidelines-for-classifying-areas-by-coronavirus-risk-11585248715
I reside in Clermont County, Ohio (just east of Cincinnati), a mix of high density (urban) and low density (rural) sections. Experts suspect that high density areas are higher risk for COVID-19 than lower density areas — my county, like many others is heterogeneous. I follow the status of data on the COVID-19 epidemic nationally, state-wide and locally and have deep concerns that the currently available data are sufficient to do a credible county-based risk assessment, not to even mention a credible state-wide assessment.
In Ohio (according to health department officials) geographic location of the case is based on the location of the lab that did the testing, not the geographical residence of the case. Yes, we have a number indicating cases, but prevalence of the disease calculation is not credible due to selection bias of those chosen for testing. In addition, factors associated with its transmission, especially on asymptomatic cases, is not possible (and we know there is a long lag time between acquiring the virus and showing symptoms – during all that time the virus can be transmitted).
The lack of data could be alleviated if we did periodic sampling of county residents using rapid epidemiological survey (RES) techniques. The CDC has been using the RES for decades to great effect. These could be repeated periodically (i.e. bi-weekly) as needed to track individuals in the population.
In the RES people or households are selected using scientific sampling techniques (including stratification by, for example, high population density and low density areas). Trained workers can administer (depending on availability) the COVID-19 swab test, blood/serum antibody test, or, if none of those are available, even body temperature measurement. Data should be collected on demographic, social, and health factors. Data collected is quickly analyzed as required by RES – all could be accomplished in a week from start to finish.
We should start with a representative sample in the state or a region to get an accurate handle on this disease.
Deploying this technique to assess state, region, and county risk, would be better enable policy makers to make decisions on social distancing (and other interventions) based on credible data.
Thomas Wilson, PhD, DrPH; Epidemiologist & Founder; Trajectory® Healthcare, LLC