PT - JOURNAL ARTICLE AU - Kaneko, Makoto AU - Ikeda, Takaaki AU - Inoue, Machiko AU - Sugiyama, Kemmyo AU - Saito, Manabu AU - Ohta, Ryuichi AU - Cooray, Upul AU - Vingilis, Evelyn AU - Freeman, Thomas R AU - Mathews, Maria TI - Development and validation of a rurality index for healthcare research in Japan: a modified Delphi study AID - 10.1136/bmjopen-2022-068800 DP - 2023 Jun 01 TA - BMJ Open PG - e068800 VI - 13 IP - 6 4099 - http://bmjopen.bmj.com/content/13/6/e068800.short 4100 - http://bmjopen.bmj.com/content/13/6/e068800.full SO - BMJ Open2023 Jun 01; 13 AB - Objectives Rural–urban healthcare disparities exist globally. Various countries have used a rurality index for evaluating the disparities. Although Japan has many remote islands and rural areas, no rurality index exists. This study aimed to develop and validate a Rurality Index for Japan (RIJ) for healthcare research.Design We employed a modified Delphi method to determine the factors of the RIJ and assessed the validity. The study developed an Expert Panel including healthcare professionals and a patient who had expertise in rural healthcare.Setting The panel members were recruited from across Japan including remote islands, mountain areas and heavy snow areas. The panel recruited survey participants whom the panel considered to have expertise.Participants The initial survey recruited 100 people, including rural healthcare providers, local government staff and residents.Primary outcome measures Factors to include in the RIJ were identified by the Expert Panel and survey participants. We also conducted an exploratory factor analysis on the selected factors to determine the factor structure. Convergent validity was examined by calculating the correlation between the index for physician distribution and the RIJ. Criterion-related validity was assessed by calculating the correlation with average life expectancy.Results The response rate of the final survey round was 84.8%. From the Delphi surveys, four factors were selected for the RIJ: population density, direct distance to the nearest hospital, remote islands and whether weather influences access to the nearest hospital. We employed the factor loadings as the weight of each factor. The average RIJ of every zip code was 50.5. The correlation coefficient with the index for physician distribution was −0.45 (p<0.001), and the correlation coefficients with the life expectancies of men and women were −0.35 (p<0.001) and −0.12 (p<0.001), respectively.Conclusion This study developed the RIJ using a modified Delphi method. The index showed good validity.Data are available upon reasonable request. The data sets are available from the corresponding author on reasonable request.