Public health officials in several states, including Florida, Georgia, Texas, Virginia, and Vermont, have been accused of altering data about coronavirus infection rates. Georgia, Florida, and Virginia were among the earliest states to begin easing lockdown restrictions.
In Florida, Rebekah Jones, a data scientist involved in the creation of the state’s coronavirus dashboard, says she was fired for refusing to manipulate data to garner support for the plan to reopen the state, according to the Associated Press.
On May 11, Georgia’s Department of Public Health published a graph showing new COVID-19 cases declining over time in heavily affected counties, but the data was arranged in descending rather than chronological order. The graph was later taken down. A spokesperson from Governor Brian Kemp’s office denied that the chart was an attempt to deceive the public, and said officials were instead trying to highlight which days had the highest rates of infections.
Georgia has also published graphs showcasing infections over time in which new cases are not listed by the date of the positive test result, which is standard practice, but rather on the day the patient first reported symptoms. This can shift the timeline of peak infections back.
Officials from other states have been accused of combining the rates of diagnostic testing, which show active infections, with antibody testing, which show past infections. This practice could potentially pad testing numbers and does not give a clear picture of how the virus is spreading. The Texas Department of Health acknowledged that it was including antibody tests in its case counts, as well as in calculating the positivity rate of the virus, or the ratio of positive cases to the total number of tests administered. Texas health officials have committed to ending the data combining practice.
The Departments of Health in both Vermont and Virginia have also promised to stop combining the two types of tests, although officials in Virginia claim that combining the numbers caused no significant difference in overall trends.
Jennifer Nuzzo, DrPH, of the Johns Hopkins Center for Health Security, said some cases of data combining are not necessarily the result of an attempt to mislead the public, but instead may be due to state officials lacking access to updated data systems that allow them to distinguish which data is from an antibody test compared to a viral test.