Statistical analyses
Descriptive analyses were performed on payment data. Per-physician
payments were calculated only among endocrinologists receiving payments
each year, as in previous studies (20-22).
Regarding general payments, this
study performed descriptive analyses, including calculation of
per-physician payments and total amounts, for general payments excluding
ownership interests and royalties and licenses (hereby GPEOR), because
only several endocrinologists received large amounts of ownership
interests and royalties and licenses (Supplemental Material 3). For
associated research payments, per-physician payments were calculated as
total payments divided by the number of principal investigators in a
research payment, and total research payments were calculated by the
number of endocrinologists times the payment per principal investigator
because some research payments included principal investigators with
other specialties. The author evaluated the share of payments by
specific proportions of endocrinologists and the Gini index. The Gini
index ranges from 0 (complete equality) to 1 (complete inequality)
(21,23). Payment data were also analyzed by categories (20,24,25).
The study evaluated trends in GPEOR and research payments with
interrupted time series analysis, using generalized estimating equations
(GEE) of annual per-physician payment clustering individual
endocrinologists, to adjust for the COVID-19 pandemic’s influence since
2020. As the payment distribution was skewed, the study employed the
negative binomial regression GEE model for changes in per-physician
payments and modified linear regression GEE model with Poisson
distribution for changes in the proportion of endocrinologists receiving
payments from 2014-2022 (22,23,25,26).
Relative annual mean percentage
change was reported. The study
excluded from the GEE trend analyses of GPEOR those payments for
acquisitions, debt forgiveness, and long-term medical supply or device
loans, because general payments for acquisitions, debt forgiveness, and
long-term medical supply or device loans were only available for
2021-2021 data. Inflation was adjusted to 2022-dollar values by dividing
the payment values each year by relative consumer price index (CPI)
compared to 2022 from the US Bureau of Labor Statistics CPI Inflation
Calculator (Supplemental Material 4).