Predicting Complications in Head and Neck Surgery: Comparing Calculators to Surgeons

Area of Science:

• Surgical outcomes research
• Predictive analytics in medicine
• Healthcare quality assessment

Background:

• Surgical outcomes significantly influence hospital rankings, reputation, and funding.
• Objective surgical risk calculators (SRCs) are commonly used, but surgeons' own assessments are often overlooked.
• There is a need to compare the predictive accuracy of surgeons versus SRCs.

Purpose of the Study:

• To prospectively evaluate whether surgeons' preoperative risk assessments or objective surgical risk calculators (SRCs) are more accurate in predicting patient surgical outcomes.
• To compare the performance of surgeons and the American College of Surgeons (ACS) tool in predicting any complication and serious complication.

Main Methods:

• Prospective cohort study design involving 101 patients.
• Collected surgeons' preoperative risk assessments and calculated patient risk using an SRC.
• Compared predictions against actual patient outcomes and against each other using receiver operating characteristic (ROC) analysis.

Main Results:

• 36.6% of patients experienced any complication, and 17.8% experienced a serious complication.
• Preoperative smoking significantly increased overall complication rates (2.49 times higher).
• Neither surgeons (AUC 0.51) nor the ACS tool (AUC 0.58) accurately predicted any complication; performance for serious complications was also not statistically significant (surgeons AUC 0.55, ACS tool AUC 0.60).

Conclusions:

• Current validated risk calculators and surgeons demonstrate poor accuracy in predicting perioperative risk.
• Preoperative smoking is the most significant factor for improving complication prediction, alongside age and surgery type.
• Risk calculators may be inappropriate for assessing hospital performance; improved predictive tools, potentially incorporating AI, are needed.