Measuring geographical disparities in waiting times for community-based specialist care - a novel statistical application

Gaps in provision of healthcare by geographical location have been demonstrated around the world. To detect disparities in waiting times (WT), methods are required for accurate comparison. This study sought to demonstrate a novel statistical application to assess geographical differences in WT and time trends for community dermatologists in Israel between 2019 and 2023. WT were measured using available appointments and actual visits for all dermatologists in all four Israeli Health Maintenance Organizations (HMO). Piecewise quantile regressions were conducted for two periods: 2019–2021 and 2022–2023. Additive and multiplicative models estimated absolute and relative differences in median WT between regions. Expected WT, time trends, and differences from the national level were calculated. Validation of the statistical model was conducted on 2024 data. Between 2019 and 2021, median national WT began at 17 days (95% CI 14–20), increasing annually by 2.5 days (+ 11%) until end of 2021. WT was shortest in Jerusalem (10 days) and longest in Tel Aviv (21 days). Increasing time trends were demonstrated in all regions except North and Haifa. The model successfully predicted 2024 national WT. By 2022–2023, national median WT began at 24.5 days (95% CI 21–28) and reached an annual increase of 8.5 days (+ 26%). WT was shortest in the North (14 days) and longest in Tel Aviv (40) and J&S (42 days). Increasing time trends were demonstrated in all regions except North and Haifa, with the greatest increase in Jerusalem and Centre regions (+ 12 days/year). WT for a dermatologist in the public healthcare system increased significantly between 2019 and 2023, particularly after the peak of the COVID pandemic, with increased disparities among regions. Advanced statistical modelling, such as piecewise quantile regression, highlights areas in need of intervention and can contribute to resource planning to help improve system efficiency. Future research should examine explanatory factors including physician-hours per population, patient scheduling preferences, as well as advancing tools to help direct patients to the most appropriate service to reduce burden on specialists.