Abstract
<jats:p>Background/Objectives: Serum creatinine testing was not a statutory component of routine occupational health examinations in Japan during the study period. Because it will become a statutory examination item from April 2027, we evaluated baseline facility-level variation in recorded creatinine results and whether patient- and facility-level characteristics accounted for this variation. Methods: We analyzed 832 women from 151 facilities. The outcome was the presence of a numerical serum creatinine result in the health examination dataset. Between-facility variation was assessed using mixed-effects logistic regression with a facility random intercept. Intraclass correlation coefficients (ICCs) and 95% confidence intervals (CIs) were estimated using the latent-variable approach and parametric bootstrap. The adjusted model included available patient-level characteristics. Facility-level associations were examined using grouped quasibinomial regression and analyses restricted to facilities with at least five or ten examinees. Results: Creatinine results were recorded for 581 women (69.8%). Facility-specific rates ranged from 0% to 100%; 25 facilities (16.6%) had no recorded results and 59 (39.1%) had results for all examinees. The unadjusted ICC was 0.093 (95% CI, 0.016–0.164; p = 0.0056). Among 779 complete cases from 147 facilities, the adjusted ICC was 0.085 (95% CI, 0.006–0.159; p = 0.0084). Facility size, municipality population, and ordinance-designated city status were not statistically associated with recording rates; conclusions were unchanged in restricted analyses. Conclusions: Modest but statistically significant facility-level variation persisted after adjustment for available patient-level characteristics. These findings provide a pre-implementation baseline for evaluating the nationwide regulatory change.</jats:p>