How is increased selectivity of medical school admissions associated with physicians’ career choice? A Japanese experience

Table 5 Effects on medical facility choice: internal medicine vs other specialty

	7:1	DPC	Clinic	“Core”
			Owner	Primary care
	LOGIT	LOGIT	LOGIT	LOGIT
	(5)	(6)	(7)	(7)
Panel A. Internal medicine
D1 [55 < Rank ≤ 60]	1.511 8***	1.383 9***	0.860 1***	0.906 2*
	[1.327 4–1.721 9]	[1.226 3–1.561 7]	[0.775 3–0.954 2]	[0.807 4–1.016 9]
D2 [60 < Rank ≤ 65]	1.671 3***	1.492 0***	0.793 3***	0.889 3
	[1.415 1–1.973 9]	[1.275 9–1.744 7]	[0.686 7–0.916 4]	[0.740 8–1.067 5]
D3 [65 ≥ Rank]	1.689 0***	1.539 1***	0.795 4**	0.935 7
	[1.370 2–2.081 8]	[1.251 1–1.893 5]	[0.650 5–0.972 5]	[0.743 3–1.178 0]
Obs.	46 036	46 036	45 905	45 947
Mean of dep.	0.40	0.56	0.22	0.12
Panel B. Other specialty
D1 [55 < Rank ≤ 60]	1.515 2***	1.336 7***	0.832 8***	1.076 3
	[1.383 9–1.659 0]	[1.207 5–1.479 7]	[0.759 8–0.912 8]	[0.956 2–1.211 5]
D2 [60 < Rank ≤ 65]	1.663 3***	1.461 4***	0.768 7***	1.090 1
	[1.488 1–1.859 3]	[1.267 1–1.685 4]	[0.664 3–0.889 5]	[0.922 1–1.288 7]
D3 [65 ≥ Rank]	1.654 0***	1.435 5***	0.783 6**	1.336 1**
	[1.416 9–1.930 6]	[1.166 5–1.766 4]	[0.638 3–0.962 1]	[1.008 1–1.770 8]
Obs.	76 954	76 954	76 652	76 614
Mean of dep.	0.42	0.60	0.22	0.04

Note: In all equations, the polynomial function of physician’s age, gender, and fixed effects of medical school attended are controlled for. Standard errors are clustered at the level of medical school attended. Abbreviations: 7:1 hospital with a 7:1 basic reimbursement rate, DPC hospital that adopts a diagnosis procedure combination/per-diem payment system. “Obs” represents the number of observations. “Mean of dep.” represents the mean of dependent variables. “LOGIT” represents logistic regression. Odds ratio are reported. 95% confidence interval is in parentheses. ***p < 0.01, **p < 0.05, and *p < 0.1, respectively

ISSN: 1478-4491