Accuracy and Feasibility of Clinically Applied Frailty Instruments before Surgery: A Systematic Review and Meta-analysis.

BACKGROUND: A barrier to routine preoperative frailty assessment is the large number of frailty instruments described. Previous systematic reviews estimate the association of frailty with outcomes, but none have evaluated outcomes at the individual instrument level or specific to clinical assessment of frailty, which must combine accuracy with feasibility to support clinical practice. METHODS: The authors conducted a preregistered systematic review (CRD42019107551) of studies prospectively applying a frailty instrument in a clinical setting before surgery. Medline, Excerpta Medica Database, Cochrane Library and the Comprehensive Index to Nursing and Allied Health Literature, and Cochrane databases were searched using a peer-reviewed strategy. All stages of the review were completed in duplicate. The primary outcome was mortality and secondary outcomes reflected routinely collected and patient-centered measures; feasibility measures were also collected. Effect estimates were pooled using random-effects models or narratively synthesized. Risk of bias was assessed. RESULTS: Seventy studies were included; 45 contributed to meta-analyses. Frailty was defined using 35 different instruments; five were meta-analyzed, with the Fried Phenotype having the largest number of studies. Most strongly associated with: mortality and nonfavorable discharge was the Clinical Frailty Scale (odds ratio, 4.89; 95% CI, 1.83 to 13.05 and odds ratio, 6.31; 95% CI, 4.00 to 9.94, respectively); complications was associated with the Edmonton Frail Scale (odds ratio, 2.93; 95% CI, 1.52 to 5.65); and delirium was associated with the Frailty Phenotype (odds ratio, 3.79; 95% CI, 1.75 to 8.22). The Clinical Frailty Scale had the highest reported measures of feasibility. CONCLUSIONS: Clinicians should consider accuracy and feasibility when choosing a frailty instrument. Strong evidence in both domains support the Clinical Frailty Scale, while the Fried Phenotype may require a trade-off of accuracy with lower feasibility.