The perioperative frailty index derived from the Chinese hospital information system: a validation study

Abstract

Background: There are various frailty assessment tools in the world, and the application choice of frailty assessment tools for the elderly perioperative population varies. It remains unclear which frailty assessment tool is more suitable for the perioperative population in China. To validate the Perioperative Frailty Index (FI-32) derived from the Chinese Hospital Information System by investigating the impact of preoperative frailty on postoperative outcomes, and ascertain the diagnostic value of FI-32 for predicting postoperative complications through comparing with the FRAIL scale and the modified Frailty Index (mFI-11).

Methods: A prospective cohort study was conducted in a tertiary hospital. Elderly patients who were 60 years or older and underwent selective operation were included. The FI-32, FRAIL scale, and mFI-11 were assessed. Demographic, surgical variables and outcome variables were extracted from medical records. The data of readmission and mortality within 30 days and 90 days of surgery were ascertained by Telephone follow-up by professionally trained researchers. Multiple logistic regression was used to examine the association between frailty and complications. Receiver operating characteristic curves(ROC) were used to compare FI-32 with mFI-11 and FRAIL, to explore the predictive ability of frailty.

Results: 335 patients qualified for the inclusion criteria and were enrolled in the study, and among them, 201 (60.0%) were females, and the Median(P₂₅, P₇₅)age at surgery was 69 (65,74) years. The prevalence of frailty in the study population was 16.4% (assessed by FI-32). After adjusting for concomitant variables including demographic characteristics (such as gender, BMI, smoking, drinking, average monthly income and educational level) and surgical factors (such as surgical approach, surgical site, anesthesia method, operation time, intraoperative bleeding, and intraoperative fluid intake), there was a statistically significant association between frailty and the development of postoperative complication after surgery (OR = 3.051, 95% CI:1.460-6.378, P = 0.003). There were also significant differences in mortality within 30 days of surgery, the length of hospital stay (LOS) and the hospitalization costs. FI-32, FRAIL and mFI-11 showed a moderate predictive ability for postoperative complications, the Area Under Curves (AUCs) were 0.582, 0.566 and 0.531, respectively. With adjusting concomitant variables associated with postoperative complications, the AUCs of FI-32, FRAIL and mFI-11 in the adjusted prediction models were 0.824, 0.827 and 0.820 respectively.

Conclusions: The FI-32 has a predictive effect on postoperative adverse outcomes in elderly Chinese patients. Compared to FRAIL and mFI-11, the FI-32 had the same ability to predict postoperative complications, and FI-32 can be extracted directly from HIS, which greatly saves the time for clinical medical staff to evaluate perioperative frailty.

Keywords: Frailty; Perioperative complication; Perioperative frailty index( FI-32); Postoperative outcomes.

Fig. 1

An ROC curve of FI-32, FRAIL and mFI-11 predicting the postoperative complications

Fig. 2

An ROC curve of the adjusted prediction models of FI-32, FRAIL and mFI-11 predicting the postoperative complications. Notes: * indicates: adjusted by demographic characteristic (gender, BMI, smoking, drinking, average monthly income and educational level) + surgical factors (surgical approach, anesthesia method, operation time, intraoperative bleeding, total intraoperative intake)