Work

Fallacy of residence-based estimates of rehabilitation demand: a nationwide study linking residence, preceding hospitalization, and rehabilitation in Korean National Health Insurance claims

Year: 2026
Type: preprint
Study design: Observational Study
Abstract: Background: Residence-based estimates of rehabilitation demand do not indicate where patients receive inpatient care before rehabilitation. We examined where patients with out-of-region rehabilitation had their preceding admission, and how regional patient distributions varied depending on where the patients were counted (attribution points). Methods: In this nationwide retrospective study, we used National Health Insurance Service claims of all 37,480 adults with their first inpatient claim under Korea's designated rehabilitation program between March 2020 and December 2024. We linked registered residences, preceding-inpatient regions, and rehabilitation regions among Korea's 17 regions. The preceding admission was the most recent eligible episode ending on or within 90 days before the index admission date. We classified out-of-region rehabilitation by preceding-inpatient region and compared regional distributions using the index of dissimilarity. Results: All three regions were identified in 36,984 patients (98.7%), of whom 10,870 (29.4%) underwent out-of-region rehabilitation. Among the patients who underwent out-of-region rehabilitation, 65.5% (n=7,121) had an out-of-region preceding admission. Of the 10,870 patients who underwent out-of-region rehabilitation, 51.2% had their preceding admission in the rehabilitation region and 14.3% in a third region. By residence region, the proportion of patients with an out-of-region preceding admission ranged from 29.9% in Seoul to 85.3% in Chungnam. The regional distributions also differed by attribution point. The index of dissimilarity between residence-based and preceding-admission-based counts was 13.91% (N=36,984). Among the 37,477 patients with residence and rehabilitation regions identified, the corresponding index for residence-based and rehabilitation-based counts was 16.69%. Relative to residence-based counts, Seoul's share was higher when patients were counted at the preceding admission and lower when they were counted at the rehabilitation institution. Conclusions: The proportion of out-of-region rehabilitation alone cannot be interpreted directly as a local shortage of rehabilitation beds. The regional patient distribution varies by attribution point; therefore, residence-based counts alone are insufficient for regional bed planning. Regional reports should state the attribution point and distinguish out-of-region rehabilitation according to the preceding-inpatient region.
Language: English
Open Access status: green
OpenAlex ID: W7225871406
Cites: 23
Cited by:
Related to: 10