Risk prediction models for frailty in older adults: A systematic review and critical appraisal.

Area of Science:

• Gerontology
• Clinical Epidemiology
• Biostatistics

Background:

• Frailty in older adults increases adverse health outcomes.
• Risk prediction models can guide frailty prevention strategies.
• The quality and clinical applicability of existing frailty prediction models are largely unknown.

Purpose of the Study:

• To systematically review risk prediction models for frailty in older adults.
• To critically appraise the quality and validation of these models.

Main Methods:

• Searched PubMed, Embase, CINAHL, and Cochrane Library (up to June 2024).
• Included studies developing or validating frailty risk prediction models.
• Assessed model quality using the Prediction Model Risk of Bias Assessment Tool.

Main Results:

• Included 19 studies with 22 frailty risk prediction models.
• Common predictors: age, cognitive function, self-rated health, sex, ADLs, depression.
• Models showed good internal discrimination (AUC/c-index 0.707-0.920) and calibration, but external validation was limited (AUC/c-index 0.612-0.889).
• High risk of bias and concerns regarding applicability were noted in most models.

Conclusions:

• Current frailty risk prediction models for older adults exhibit inadequate validation and evaluation.
• Future research needs to enhance existing models and develop new ones with robust methodologies for improved clinical implementation.