Appendix Table B.

Associations between Primary Care Physician Characteristics and Potentially Inappropriate Medication Prescription Rate*

CoefficientSE
Age0.016(0.002)
Female−0.708(0.051)
DO degree0.141(0.037)
Medical school cohort
    Pre to 1980 graduateReference Category
    1980 to 1989 graduate−0.136(0.037)
    1990 to 1999 graduate−0.050(0.046)
    2000 to 2009 graduate0.075(0.066)
    2010 to present graduate0.335(0.173)
Primary care specialty
    Family medicineReference Category
    Internal medicine0.253(0.033)
    General practice0.740(0.111)
    Geriatric medicine−0.082(0.122)
Practice region
    NortheastReference Category
    Midwest0.107(0.275)
    South0.875(0.313)
    West0.305(0.759)
Practice size
    Quartile 1 (1–5)Reference Category
    Quartile 2 (6–65)−0.350(0.040)
    Quartile 3 (66–300)−0.368(0.044)
    Quartile 4 (300+)−0.384(0.051)
Patient panel size (age 65+)
    <50Reference Category
    50–199−2.808(0.228)
    200–499−3.690(0.241)
    ≥500−3.714(0.247)
  • The model was adjusted for patient panel characteristics, year, and hospital referral region fixed effects. Patient panel characteristics included average age of the panel, proportion of female, proportion of Medicare patients who were under 65 years of age, average CMS-Hierarchial Condition Categories risk score of the panel, and whether the panel had higher than sample average proportion of White and dual-eligible patients.

  • * Data obtained from 2013 to 2015 Medicare Part D Public Use File and 2015 American Medical Association Masterfile.

  • P < .001.

  • P < .05.