Table 3.

Effects of rural-urban location and malpractice burden on providers’ likelihood of discontinuing obstetric care

	Model 1^a	Model 2^b	Model 3^c
Effect of Claim Experience
*Family Physicians*
Ever had a malpractice claim filed	3.80 (1.46-9.86)	1.78 (0.61-5.14)	2.23 (0.69-7.19)
Primary practice in rural (versus urban) Michigan	2.98 (1.20-7.38)	3.22 (1.21-8.57)	4.01 (1.26-12.76)
Interaction term: claim experience × rural practice	0.65 (0.18-2.39)	0.59 (0.15-2.33)	0.69 (0.15-3.07)
Sample size	226	226	225
Max-rescaled R-Square	0.54	0.75	0.89
*Obstetrician-Gynecologists*
Ever had a malpractice claim filed	2.67 (0.74-9.57)	0.70 (0.16-3.08)	0.83 (0.17-3.98)
Primary practice in rural (versus urban) Michigan	2.31 (0.40-13.27)	3.35 (0.58-19.34)	5.34 (0.79-36.13)
Interaction term: claim experience × rural practice	0.42 (0.06-2.89)	0.34 (0.04-2.69)	0.31 (0.04-2.84)
Sample size	269	269	260
Max-rescaled R-Square	0.05	0.53	0.63
Effect of Malpractice Payment
*Family Physicians*
Ever made malpractice payment ≥$30,000	1.70 (0.57-5.07)	0.56 (0.17-1.86)	0.76 (0.20-2.90)
Primary practice in rural (versus urban) Michigan	2.26 (1.11-4.57)	2.62 (1.15-5.99)	3.24 (1.23-8.58)
Interaction term: malpractice payment ≥$30,000 × rural practice	0.61 (0.14-2.74)	0.53 (0.11-2.51)	0.51 (0.10-2.54)
Sample size	224	224	223
Max-rescaled R-Square	0.19	0.76	0.88
*Obstetrician-Gynecologists*
Ever made malpractice payment ≥$30,000	1.49 (0.69-3.21)	0.59 (0.25-1.37)	0.58 (0.24-1.43)
Primary practice in rural (versus urban) Michigan	0.94 (0.33-2.69)	1.09 (0.32-3.73)	1.67 (0.42-6.63)
Interaction term: malpractice payment ≥$30,000 × rural practice	1.43 (0.32-6.36)	1.47 (0.24-8.93)	1.24 (0.19-8.13)
Sample size	264	264	255
Max-rescaled R-Square	0.03	0.53	0.63

Estimates were reported in odds ratios (95% confidence intervals).

Model 1 = controlling for rural-urban location of the physician’s primary practice, malpractice burden, and the interaction term between these two variables.

Physician personal characteristics, including age, gender, race/ethnicity, medical school (whether graduated from a medical school in another country), and board certification, were considered as candidate explanatory variables to be added to Model 2. The exact list of covariates varied across the models for family physicians and obstetrician-gynecologists based on the significance level of their association with discontinuation of obstetric care found in bivariate analysis.

For family physicians: Model 2 = Model 1 + age.

For obstetrician-gynecologists: Model 2 = Model 1 + age, gender, and board certification

Model 3 = Model 2 + type of practice (primarily solo practice, non-office-based practice, versus office-based non-solo practice), and the average number of hours per week spent on direct patient care.