Abstract
This study quantitatively analyzed healthcare administration studies in Japan using text mining, focusing on articles published during 1994-2021 in the Journal of the Japan Society for Healthcare Administration (prior to 2008, the journal was called Hospital Administration). Both the co-occurrence network and the correspondence analysis (these are extracted words that refer to the two systems) demonstrate two major changes: (1) the introduction of the long-term care insurance system, which was enacted in 1997 and came into effect in 2000, and (2) the introduction of the late-stage medical care system for the elderly in 2008, both of which had a significant impact on the Japanese public health and welfare system. Co-occurrence network and correspondence analysis were conducted to understand changes in research interests. The analysis used two time periods following a change in the journal's name in 2008. To readily comprehend changing research trends, 10-year segments were considered, resulting in three time periods. The research features and trends during the aforementioned periods were examined using correspondence analysis. Configuration figures derived from this analysis plotted time transition (first dimension) against certain abstract/concrete situations (second dimension). The extracted words displayed in the configuration maps at the axes' intersection were patient, survey, and evaluation. They revealed no distinctive features compared with other words and were commonly used in article titles within this journal during each period. The following results were obtained from the correspondence analysis: first, changes in the geriatric care system of public medical insurance and the introduction of the long-term care insurance system in 2000 were expressed in the characteristics of the extracted words; second, in the 14 years after the journal's name changed, published studies frequently referred to the roles of doctors, nurses, and other healthcare professionals. A chi-squared test on these extracted words and the period classification confirmed a statistically significant relationship between them.