To the Editor,Dr. Harris’ study, “COVID-19 Incidence and hospitalization during the delta surge were inversely related to vaccination coverage among the most populous U.S. Counties” [1] is timely and pertinent. However, his calculation of the impact of vaccination coverage on case incidence, hospitalization rate, and the% of admissions is incorrect. The loglinear regression coefficient, − βΔX, is used directly to represent the decrease in rates instead of using the correctly back-transformed equation, 1− exp (βΔX) [2]. This leads to an overestimation of the impact of vaccination coverage.Dr. Harris reports values of 28.3%, 44.9% and 16.6% for the decrease in incidence rate of cases, hospitalization rate and % of cases hospitalized, corresponding to a 10% difference in vaccination coverage. Using the correct equation however, the decreases are 24.6%, 36.2% and 15.3%. The magnitude of overestimation due to this error increases with increasing difference in vaccination coverage assessed and increasing magnitude of regression coefficient (β).While the essential finding of an inverse relationship between vaccination coverage and COVID-19 case incidence, hospitalizations and % of cases hospitalized remains valid, the magnitudes of decrease of these quantities have been overestimated. Estimation of the decrease of rates from a log linear model due to a change in vaccination coverage, ΔX, should be correctly calculated using the equation 1− exp (βΔX).
Author statement
Data used in this commentary were drawn directly from the Harris manuscript [1].