Background a Frailty Index (FI) calculated by the accumulation of deficits is often used to quantify the extent of frailty in individuals in specific settings. This study aimed to derive a FI that can be applied across three standardised international Residential Assessment Instrument assessments (interRAI), used at different stages of ageing and the corresponding increase in support needs. Methods deficit items common to the interRAI Contact Assessment (CA), Home Care (HC) or Long-Term Care Facilities assessment (LTCF) were identified and recoded to form a cumulative deficit FI. The index was validated using a large dataset of needs assessments of older people in New Zealand against mortality prediction using Kaplan Meier curves and logistic regression models. The index was further validated by comparing its performance with a previously validated index in the HC cohort. Results the index comprised 15 questions across seven domains. The assessment cohort and their mean frailty (SD) were: 89,506 CA with 0.26 (0.15), 151,270 HC with 0.36 (0.15) and 83,473 LTCF with 0.41 (0.17). The index predicted 1-year mortality for each of the CA, HC and LTCF, cohorts with area under the receiver operating characteristic curves (AUCs) of 0.741 (95% confidence interval, CI: 0.718–0.762), 0.687 (95%CI: 0.684–0.690) and 0.674 (95%CI: 0.670–0.678), respectively. Conclusions the results for this multi-instrument FI are congruent with the differences in frailty expected for people in the target settings for these instruments and appropriately associated with mortality at each stage of the journey of progressive ageing.

中文翻译：

---

与三种 interRAI 评估工具兼容的脆弱指数的开发和验证

背景 通过赤字累积计算的虚弱指数 (FI) 通常用于量化特定环境中个体的虚弱程度。本研究旨在得出一个 FI，该 FI 可应用于三个标准化的国际住宅评估工具评估 (interRAI)，用于不同的老龄化阶段和相应的支持需求增加。方法 interRAI 接触评估 (CA)、家庭护理 (HC) 或长期护理设施评估 (LTCF) 共有的缺陷项目被识别并重新编码以形成累积缺陷 FI。该指数通过使用卡普兰迈尔曲线和逻辑回归模型对新西兰老年人进行死亡率预测的大型需求评估数据集进行验证。通过将其性能与 HC 队列中先前验证的指数进行比较，进一步验证了该指数。结果 该索引包含七个领域的 15 个问题。评估队列及其平均虚弱度 (SD) 为：89,506 CA 为 0.26 (0.15)，151,270 HC 为 0.36 (0.15) 和 83,473 LTCF 为 0.41 (0.17)。该指数预测 CA、HC 和 LTCF 队列的 1 年死亡率，受试者工作特征曲线下面积 (AUC) 分别为 0.741（95% 置信区间，CI：0.718–0.762）、0.687（95%CI： 0.684–0.690) 和 0.674 (95%CI: 0.670–0.678)。结论 这种多仪器 FI 的结果与这些仪器的目标设置中人们预期的虚弱差异一致，并且与渐进老龄化过程的每个阶段的死亡率适当相关。