A Time-series Analysis of Hospital Choice Behavior of Inpatients

本研究假设住院病人的医院选择行为是随着地区医疗服务水平、人口老龄化、生活水平提高对医疗服务需求的增加、交通便利等因素的变化而动态变化的。然而，时间序列分析表明，要求在同一地区进行至少两组医院选择调查。由于这一数据要求很少得到满足，过去对医院选择行为的研究大多是静态的，因此，1973年和1984年在广岛县进行的医院选择调查的数据可用性促使了本研究。对于每一个调查数据，准备一个住院病人的来源-目的地矩阵，并通过基于最小信息损失作为邻近度指标的聚类分析进行分析。在这两个日期之间，每1000人的平均每日患者日从8.96增加到10.78，增加了1.82。然而，这一增长的70%可以归因于人口老龄化因素。65岁以上患者的住所与医院之间的平均距离从7.71公里适度缩短到7.53公里，64至16岁患者的住所与医院之间的平均距离从11.00公里缩短到9.17公里，15岁以下患者的住所与医院之间的平均距离从18.44公里缩短到15.92公里。尽管各县的平均数有所改善，但偏远农村地区的数字却在恶化。在该省人口减少的农村地区，向老年病人提供保健服务日益成为一个关键问题。

This study hypothesizes that the hospital choice behavior of impatients is dynamically changing from one time to another in accordance with a combination of factors such as the level of medical services provided in the region, aging of the population, increasingly higher level of demand for medical services commensurate with the better standard of living, ease of transportation and so on. The time-series analysis, however, requires at least two sets of hospital choice surveys taken in the same region. Because this data requirement has been rarely satisfied, past studies of hospital choice behavior have been mostly static.Accordingly, the data availability of hospital choice surveys in 1973 and 1984 in Hiroshima Prefecture has prompted this study. For each survey data, an origin-destination matrix of in- patients is prepared and analyzed by means of a cluster analysis based on the minimum information loss as the indicator of proximity. Between the two dates, the mean daily patient-days per 1000 population increased from 8.96 to 10.78, an increase of 1.82. Seventy percent of this gain, however, can be accounted for by the aging factor of population. Average distances between residences of patients and locations of hospitals had been moderately shortened from 7.71 km to 7.53 km among patients 65 years of age or older, from 11.00 km to 9.17 km among patients between 64 and 16 years of age, and from 18.44 km to 15.92 km among patients 15 years or younger. In spite of the improvements in the prefectural averages, however, figures for the remote rural areas had deteriorated. Health-care provision for aged patients in the depopulating rural areas of the Prefecture is increasingly becoming a critical issue.