Hospital Access Restrictions and Conditional Admission in Japanese Long-Term Care Facilities During COVID-19: A Nationwide Facility Survey of Outcomes

Background We quantified the prevalence and forms of hospital access restrictions in Japanese long-term care facilities (LTCFs) during COVID-19 and described preference-confirmation practices. Secondarily, we tested whether a simple facility-level cumulative score tracked adverse outcomes. Methods Nationwide cross-sectional mailed survey of a random sample of 1,000 LTCFs (February–March 2025). Responses were obtained from 413 facilities (41.3%) across all prefectures. Items captured admission difficulty for COVID-related and non-COVID conditions; conditional admission requirements that conflicted with resident/family preferences (e.g., admission only if ICU-level treatments/ventilation were forgone, or only if a do-not-resuscitate [DNR] order was preaccepted); outpatient refusal; preference-confirmation pathways; and a facility-level outcome indicating death or clinical deterioration requiring later admission among residents for whom hospitalization was medically indicated but initially unavailable. A four-item cumulative score (0–4) was examined; variants added or substituted outpatient-refusal items. Logistic regression estimated the association between the score and adverse outcomes in the primary cohort. Results Admission difficulty was common: 270/413 (65.4%) for COVID-related and 135/413 (32.7%) for non-COVID conditions. Conditional admission—contrary to resident/family preferences—occurred in 21/413 (5.1%; forgoing ICU-level treatments) and 25/413 (6.1%; preaccepting DNR). Preference confirmation was mainly by families/agents (287/413, 69.5%); resident 8/413 (1.9%); none 58/413 (14.0%). Facility capacity was limited (on-site night-shift nursing 4/413 [1.0%]; full-time physicians 7/413 [1.7%]). In the primary cohort (310/413), the adverse-outcome proportion rose stepwise with the score; unadjusted logistic regression showed an odds ratio of 1.71 per point (95% CI 1.31–2.22). Sensitivity analyses—including variants adding outpatient refusal, a fever-specific item, and bounding assumptions—were directionally consistent. Conclusions Japanese LTCFs frequently faced hospital access barriers, including conditional admission, and preference confirmation was largely mediated by families/agents. A simple cumulative measure tracked worse outcomes, highlighting decision points vulnerable during surges and underscoring the need to integrate LTCFs into regional escalation/transfer frameworks and to standardize resident-centered confirmation processes distinct from triage.