Dzando Gideon, Ward Paul R, Mwanri Lillian, Moussa Richard K, Aheto Justice Moses K, Ambagtsheer Rachel C
Research Centre for Public Health, Equity and Human Flourishing, Torrens University Australia, Wakefield Campus, Adelaide, Australia.
Ecole Nationale Supérieure de Statistique et d'Economie Appliquée (ENSEA), Abidjan, Côte d'Ivoire.
Commun Med (Lond). 2025 Jun 17;5(1):237. doi: 10.1038/s43856-025-00952-1.
Despite the increasing attention on frailty as a global public health concern, frailty screening among older people in Sub-Saharan Africa (SSA) continues to rely on instruments and thresholds from high-income countries. These instruments and thresholds may not be useful in SSA due to contextual differences. We explored the development of a frailty threshold for older people in SSA.
We utilized pooled cross-sectional data from four SSA countries (Kenya, Ghana, Uganda and Côte d'Ivoire) to determine a frailty index threshold for 5527 older people (50 years and above) using a two-step approach. The mean ages of the participants ranged from 62.13 (SD: 9.60) to 74.00 (SD: 9.40) years. The participants were mostly females across the four countries, ranging from 50.1% in Côte d'Ivoire to 65.3% in Kenya. Country-specific frailty thresholds were developed using the Receiver Operating Characteristics (ROC) method. The primary thresholds were further combined into a single threshold using random effects meta-analysis. Subgroup analyses and meta-regression were conducted to explore potential sources of heterogeneity in the pooled frailty threshold.
Here we show the Area Under the Curves from the ROC analyses ranging between 0.91 (CI: 0.89, 0.93) and 0.94 (CI: 0.92, 0.97), with sensitivities ranging from 0.83 to 0.94 and specificities from 0.72 to 0.87. An overall threshold of 0.29 (95% CI: 0.25, 0.33) was obtained after pooled analysis of the country-specific thresholds.
This work demonstrates that using context-specific data can yield valuable insights into frailty thresholds among older people in SSA, enabling more culturally relevant interventions. Effective frailty screening must account for population-level differences, including demographic, health, and socio-cultural factors.
尽管衰弱作为一个全球公共卫生问题日益受到关注，但撒哈拉以南非洲（SSA）地区老年人的衰弱筛查仍继续依赖高收入国家的工具和阈值。由于背景差异，这些工具和阈值在SSA可能并不适用。我们探索了为SSA地区老年人制定衰弱阈值的方法。
我们利用来自四个SSA国家（肯尼亚、加纳、乌干达和科特迪瓦）的汇总横断面数据，采用两步法为5527名50岁及以上的老年人确定衰弱指数阈值。参与者的平均年龄在62.13岁（标准差：9.60）至74.00岁（标准差：9.40）之间。这四个国家的参与者大多为女性，在科特迪瓦占50.1%，在肯尼亚占65.3%。使用受试者工作特征（ROC）方法制定了各国特定的衰弱阈值。使用随机效应荟萃分析将主要阈值进一步合并为一个单一阈值。进行亚组分析和荟萃回归以探索汇总衰弱阈值中异质性的潜在来源。
我们在此展示ROC分析的曲线下面积在0.91（置信区间：0.89，0.93）至0.94（置信区间：0.92，0.97）之间，灵敏度在0.83至0.94之间，特异性在0.72至0.87之间。对各国特定阈值进行汇总分析后，得出的总体阈值为0.29（95%置信区间：0.25，0.33）。
这项工作表明，使用特定背景的数据可以为SSA地区老年人的衰弱阈值提供有价值的见解，从而实现更具文化相关性的干预措施。有效的衰弱筛查必须考虑人群层面的差异，包括人口统计学、健康和社会文化因素。
