Ventral hernia repair outcomes predicted by a 5-item modified frailty index using NSQIP variables

Frailty is a decrease in physiologic reserve that is separate from the normal aging process. Previously, an 11-item modified frailty index (mFI) using NSQIP variables predicted outcomes for surgical patients. We aim to validate a condensed 5-item mFI in ventral hernia patients and determine outcomes and the relative impact of each frailty variable.

The NSQIP database was queried from 2011 to 2016 for patients undergoing VHR. Spearman's rho correlation was used to determine the degree of correlation between 11-item and 5-item mFI raw frailty scores. Chi squared testing was used to determine odds ratios (95% CI) for accumulating frailty variables in both indices with regard to complications vs a baseline of zero variables present on the 11-item scale. Complications were defined by the Clavien-Dindo (CD) classification. Univariate and multivariate analyses were performed on each frailty variable to determine their relative weighted impacts on outcomes.

97,905 patients (99.45%) had all five frailty variables recorded. Only 11,549 patients (11.73%) had all variables from the 11-item mFI. No difference existed between groups for the five mutually shared frailty variables, BMI, emergent vs non-emergent procedures, operative time, or operative approach. For accumulating variables in both indices, the 5-item mFI predicts incidence of any complications, major complications, and discharge not to home similarly to the 11-item mFI. The most significantly weighted variable for complications and discharge not to home is functional status.

A 5-item mFI accurately predicts outcomes similar to the validated 11-item mFI and captures more patients for analysis.