“…As the payments were highly skewed, negative binomial regression GEE model for the payments per physician and linear log-linked regression GEE model with Poisson distribution for the number of physicians with payments were employed. 3,4 As the national emergency concerning the COVID-19 pandemic was declared in the United States on March 13, 2020, we divided the study period into before (January 2014 to February 2020) and during the pandemic (March to December 2020). To adjust the seasonality of the payments, we included the month variable in the ITS models.…”