Qiu Yue, Xiong Weiqing, Fang Xinyue, Li Pei, Conroy Simon, Maynou Laia, Rockwood Kenneth, Liu Xien, Wu Ji, Street Andrew
Tsinghua Medicine, Tsinghua University, Haidian District, Beijing, 100084, China.
Tsinghua Shenzhen International Graduate School, Tsinghua University, Shenzhen, Guangdong, 518055, China.
Eur Geriatr Med. 2025 May 2. doi: 10.1007/s41999-025-01212-0.
To validate the Hospital Frailty Risk Score (HFRS) in Chinese hospital settings, describing how patients are allocated to frailty risk groups and how frailty risk is associated with length of stay (LoS) and hospital costs.
Retrospective observational study.
Forty-eight hospitals in Lvliang City, Shanxi Province, China.
Patients aged 75 years or older hospitalised between 1 January 2022 and 31 December 2023 (n = 34,731).
A logistic regression model examined the association between long length of stay (LoS) and frailty risk. A generalised linear model assessed the association between hospital costs and frailty risk. Subgroup analyses of age group, sex, and hospital tiers were conducted.
22.2% of patients were categorised as having zero risk, 62.4% as low risk, 15.3% as intermediate risk, and 0.08% as high risk. Compared to the zero risk group: for those with low risk, the probability of long LoS was 1.92 (95% CI 1.79-2.06) times higher and hospital costs were ¥1926 (95% CI 1655-2197) higher; for those with intermediate risk, the probability of long LoS was 2.7 (95% CI 2.49-2.96) times higher and hospital costs were ¥4284 (95% CI 3916-4653) higher; and for those with high risk, the probability of long LoS was 6.7 (95% CI 3.06-14.43) times higher and hospital costs were ¥16,613 (95% CI 12,827-20,399) higher. The explanatory power of the HFRS held across subgroups.
Compared to patients aged 75 + elsewhere, those in China had lower frailty risk scores, likely reflecting a younger age structure and recording of fewer diagnosis codes. Even so, the HFRS is a powerful predictor of long length of stay and hospital costs in China.
在中国医院环境中验证医院衰弱风险评分（HFRS），描述患者如何被分配到衰弱风险组，以及衰弱风险与住院时间（LoS）和医院成本之间的关联。
回顾性观察研究。
中国山西省吕梁市的48家医院。
2022年1月1日至2023年12月31日期间住院的75岁及以上患者（n = 34731）。
采用逻辑回归模型检验长时间住院（LoS）与衰弱风险之间的关联。采用广义线性模型评估医院成本与衰弱风险之间的关联。进行了年龄组、性别和医院层级的亚组分析。
22.2%的患者被归类为零风险，62.4%为低风险，15.3%为中度风险，0.08%为高风险。与零风险组相比：低风险患者长时间住院的概率高1.92倍（95%可信区间1.79 - 2.06），医院成本高1926元（95%可信区间1655 - 2197）；中度风险患者长时间住院的概率高2.7倍（95%可信区间2.49 - 2.96），医院成本高4284元（95%可信区间3916 - 4653）；高风险患者长时间住院的概率高6.7倍（95%可信区间3.06 - 14.43），医院成本高16613元（95%可信区间12827 - 20399）。HFRS在各亚组中的解释力均成立。
与其他地方75岁及以上的患者相比，中国患者的衰弱风险评分较低，这可能反映了更年轻的年龄结构和较少的诊断编码记录。即便如此，HFRS在中国仍是长时间住院和医院成本的有力预测指标。
