Assessment of coding-based frailty algorithms for long-term outcome prediction among older people in community settings: a cohort study from the Shizuoka Kokuho Database

• DOI: 10.1093/ageing/afac009
• Publication Date: 2022-01-07T12:10:44Z
• Authors: Shiori Nishimura; Hiraku Kumamaru; Satoshi Shoji; Eiji Nakatani; Hiroyuki Yamamoto; Nao Ichihara; Yoshiki Miyachi; Alexander T Sandhu; Paul A Heidenreich; Keita Yamauchi; Michiko Watanabe; Hiroaki Miyata; Shun Kohsaka
• Keywords: Diagnosis code

To assess the applicability of Electronic Frailty Index (eFI) and Hospital Risk Score (HFRS) algorithms to Japanese administrative claims data evaluate their association with long-term outcomes.A cohort study using a regional government healthcare care (LTC) database in Japan 2014-18.Plan enrollees aged ≥50 years.We applied two assessed scores' distributions alongside enrollees' 4-year mortality initiation government-supported LTC. Using Cox regression Fine-Gray models, we evaluated between frailty scores outcomes as well models' discriminatory ability.Among 827,744 enrollees, 42.8% were categorised by eFI fit, 31.2% mild, 17.5% moderate 8.5% severe. For HFRS, 73.0% low, 24.3% intermediate 2.7% high risk; 35 36 predictors for eFI, 92 109 codes originally used HFRS available system. Relative lowest group, highest group had hazard ratios [95% confidence interval (CI)] 2.09 (1.98-2.21) 2.45 (2.28-2.63) LTC eFI; those 3.79 (3.56-4.03) 3.31 (2.87-3.82), respectively. The area under receiver operating characteristics curves unadjusted model at 48 months was 0.68 death 0.73 0.70, respectively, HFRS.The applicable system could contribute identifications risk or use.