Owing to shortages of primary-care physicians, increasing their numbers has been recognized as an urgent issue in Japan and other countries. However, it is unclear which factors in medical education influence the decision of residents to go into primary care. We investigated the factors associated with residents' choosing to practice primary care.
Of 281 randomly selected medical facilities designated as residency training hospitals, 137 facilities answered. Self-administered questionnaires were completed by 724 residents in the third or fourth postgraduate year. Responses were compared between residents who intended to choose a career in primary care (n=175, 24.2%) and residents who intended to choose a career in other specialties (n=549, 75.8%).
In addition, for residents who had intended during their undergraduate years to enter a non-primary-care specialty (n=442, 61.1%), responses were compared between those who now intended to go into primary care (n=33, 7.5%) and those who did not (n=409, 92.5%).
Residents who had planned during their undergraduate years to choose a career in primary care (adjusted odds ratio [95% confidence interval]: 9.85 [6.24-15.5]), residents who were working as primary-care physicians at the time of the survey (7.58 [4.92-11.7]), and residents who wanted to enter rural practices in the future (2.24 [1.36-3.68]) were significantly more likely to plan to choose a career as a primary-care physician in the future.
Residents who had worked at a rural practice during residency training were significantly more likely to change their career plans from other specialties to primary care (crude odds ratio [95% confidence interval]: 2.18 [1.05-4.49]). Exposure to a rural practice during residency training may affect residents' career plans.
Integrating rural primary-care practice into residency training may help increase the number of primary-care physicians in the future.

Background: When analyzing regional disparities in healthcare resources, hospital accessibility is given little consideration. We surveyed accessibility from residential districts to medical institutions using GIS (Geographic Information System) and estimated Gini coefficient for each hospital distribution.
Methods: The subjects were 2,688 census mesh blocks ( “Cho-cho-aza” ) and 109 hospitals in Tochigi prefecture. The number of hospitals located within the road distances of 5 km, 10 km and 15 km from the geometrical center of each block was calculated using GIS. The Gini coefficient of each hospital per 100 residents was calculated among the regions located within 5 km, 10 km and 15 km from the geometrical center of the census mesh block.
Results: The population of each block was 748±1,067 (mean±SD), and the road distance to the nearest hospital from the center of each block was 4.3±4.5 km. The number of census mesh blocks with distances from the center of each block to the nearest hospital within 5 km, 5-10 km, 10-15 km and more than 15 km were 1909 (71.0%), 561 (20.9%), 139 (5.2%) and 79 (2.9%) respectively. The number of hospitals located within 5 km, 10 km and 15 km were 3.3±4.7, 8.3±8.6 and 14.4±11.4. Gini coefficients were 0.65, 0.52 and 0.43.
Conclusion: When analyzing regional disparities in healthcare resources, it is necessary to take into account not only the number of physicians and beds, but also accessibility. Gini coefficient is useful to estimate geographical distributions, and can be used as an indicator for improvement projects for hospitals.

Objectives : To determine factors related to peoples' preference for visiting home-doctors when experiencing new health problems.
Method : A questionnaire survey was conducted of people receiving annual health checkups in municipalities in the vicinity of Jichi Medical University Hospital. We surveyed personal characteristics, test equipment, having of a home-doctor, and answers to an assumed scenario (asking about willingness to visit a home-doctor in case of getting certain health problems).
According to the responses to the scenario, we divided the subjects into two groups (a home-doctor group: visiting a home-doctor; and a specialist group: not visiting a home-doctor) and statistically compared the two groups.
Results : In the analytic sample of 1,829, the home-doctor group numbered 1,097 individuals (60%) and the specialist group numbered 732 individuals (40%). The home-doctor group statistically had more home-doctors than the specialist group (adjusted odds ratio, 95% confidence interval: 2.47, 2.00-3.05).
More home-doctors in the home-doctor group had test equipment than home-doctors in the specialist group: Gastrointestinal test equipment (gastroscopy, colonoscopy, or ultrasonography) (adjusted odds ratio, 95% confidence interval: 1.39, 1.06-1.83).
Conclusion : We revealed two factors relating to the preference for visiting home-doctors: First, those people had home-doctors, and, second, the home-doctors had test equipment.