Answers are grounded in this paper and cite their sources.
To identify preoperative factors associated with moderate-to-severe pain during the first 24 h following peripheral nerve block in older patients undergoing diabetic foot surgery and to describe internal performance of a prediction model. In this single-center retrospective cohort, 412 patients were screened, 121 were excluded, and 291 were analyzed, according to the study flow diagram. The primary outcome was a maximum Numeric Rating Scale (NRS) score > 3 at scheduled assessments in the first postoperative 24 h. Five preoperative factors (Wagner grade, vibration perception threshold [VPT], serum vascular cell adhesion molecule-1 [VCAM-1], interleukin-8 [IL-8], and pain-sensation grade) were entered into a logistic regression model; repeated stratified fivefold cross-validation was reported. Of the 291 patients, 240 (82.5%; 95% CI, 77.7%-86.4%) met the pain outcome. The reported cross-validation AUC was 0.911, Brier score 0.075, calibration intercept − 0.031, and calibration slope 1.031. A previously reported bootstrap interval (0.861–0.956) cannot be treated as a validated confidence interval for the cross-validated AUC because its resampling procedure does not establish out-of-sample coverage. Internal results were promising in a selected cohort, but the high outcome frequency, possible exclusion-related bias, incompletely verified uncertainty estimates, unavailable supplementary diagnostics, and absence of external validation preclude clinical deployment. Further validation across centers, surgical procedures, and analgesic regimens is required.
No takes yet. Share an insight, caveat, or question.
Yaqiong et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: