Evaluating the effect of Japan’s 2004 postgraduate training programme on the spatial distribution of physicians

Table 6 Results from likelihood ratio test (LRT) for all physicians

	R squared	LR statistic ^a	DF ^b	*P-value*	R squared	LR statistic ^a	DF ^b	*P-value*
Model only with control variables (#1)^c	0.625				0.648
#1 + Age-adjusted mortality^d	0.636	10.81	1	.001	0.652	4.75	1	.029
#1 + Physician density^e	0.624	0.15	1	.697	0.764	139.67	1	< .001
#1 + urban/rural status^f	0.627	4.24	2	.120	0.679	34.05	2	< .001
#1 + SES composite index^g	0.627	3.14	1	.077	0.701	57.18	1	< .001
Full model^h	0.640	18.95	5	.002	0.783	172.10	5	< .001

^aThe likelihood ratio test statistic.
^bDegree of freedom.
^cThe models included only control variables, which are total population, number of primary school students per number of primary schools, crime rate, discomfort index calculated by temperature and humidity, hospital beds per 1,000 population, and the presence or absence of medical schools.
^dThe models included control variables and age-adjusted mortality.
^eThe models included control variables and ratio of number of physicians to 1,000 population.
^fThe models included control variables and urban centre and suburban.
^gThe models included control variables and socioeconomic status (SES) composite index, which was created from the percent of the population with a college-level education, percent of white-collar workers, the unemployment rate, and per capita income.
^hThe models included all variables.

ISSN: 1478-4491