Table 4.

Multivariate logistic regression analysis of factors found to be significantly associated with postresidency practice in rural FM

VARIABLESTHOSE IN RURAL FM POSTRESIDENCYCRUDE OR (95% CI)ADJUSTED OR* (95% CI) EXCLUDING ENTRY CAREER CHOICEADJUSTED OR (95% CI) INCLUDING ENTRY CAREER CHOICE
Rural FM career interest at medical school entry, n (%)
  • Not considered16 (2.2)1.0NA1.0
  • First choice39 (21.2)11.92 (6.48–21.88)NA10.03 (5.29–19.03)
  • Considered but not first choice33 (5.2)2.43 (1.32–4.46)NA2.42 (1.30–4.52)
Mean (SD) age,§ y26.1 (5.0)1.14 (1.09–1.19)1.09 (1.03–1.14)1.07 (1.02–1.13)
Hometown population, n (%)
  • > 50 00051 (4.2)1.01.0NA
  • < 50 00037 (11.1)2.85 (1.83–4.44)2.37 (1.49–3.75)NA
Relationship status, n (%)
  • Single45 (4.1)1.01.0NA
  • Married or equivalent43 (9.6)2.48 (1.61–3.82)1.69 (1.05–2.73)NA
Factor 2: societal orientation,§ mean (SD) score3.7 (0.9)1.57 (1.22–2.03)1.31 (1.00–1.72)NA
Factor 5: varied scope of practice,§ mean (SD) score3.9 (1.0)1.69 (1.34–2.13)1.39 (1.09–1.77)NA
  • FM—family medicine, NA—not applicable, OR—odds ratio.

  • * In 67.9% of cases, the model correctly classified whether or not a student went on to practise rural FM after residency (specificity of 67.8% and sensitivity of 70.1%).

  • In 76.2% of cases, the model correctly classified whether or not a student went on to practise rural FM after residency (specificity of 76.9% and sensitivity of 65.5%).

  • Reference category.

  • § For continuous variables, each unit increase (year of age or unit on the Likert scale) is associated with the stated increase in the OR. The Likert scale ranged from 1 (no influence) to 5 (important influence).