The utilization of medical care services has been partly regionalized with the implementation of referral requirement by the government since July 1, 1989 when the health insurance coverage was extended to all the people. For the purpose of regionalization, the whole country has been primarily divided into tertiary care regions, and each of them again into secondary care regions. This study investigates the self-sufficiency for in-patient care services of secondary care regions focusing on why it varies among the regions. In doing so, analysis is performed to examine a model which embodies three sets of hypotheses as follows : 1) The regional self-sufficiency for medical care services would be subject to direct influences of regional characteristics, amount of available services and structural properties of regional medical care system ; 2) The regional characteristics would have indirect effects on the self-sufficiency which are mediated by medical care services ; and 3) The amount of available services would indirectly affect the self-sufficiency by influencing the structure of regional medical care system. The results of analysis were generally consistent with the model. The findings have some practical implications. The regional self-sufficiency for medical care services partly depends upon basic properties of each region which cannot be changed in a short period of time. Thus the self-sufficiency for medical care services can be improved mainly by health policy measures. In some of the regions the self-sufficiency for in-patient care services was much higher or lower than can be predicted from the bed-population ratio. Indication is that the allocation of health resources should be made considering a variety of factors bearing upon the supply of and demand for health care ; not on the basis of just a single criterion like the availability. The self-sufficiency of a certain region is related to not only its own characterstics but also the characteristics of neighboring regions. Therefore, attention should be also directed to the inter-regional relationships in health care when the needs for investment of health resources in a region are assessed. However, it should be noted that this study used the data collected before the referral requirement was imposed. A replication of this analysis using recent data would provide an evaluation of the impact on the self-sufficiency of the referral requirement as well as a confirmation of the findings of this study.