Calibration and Discrimination of POSSUM and P-POSSUM Scores in Elective General Surgery: A Prospective Observational Study

Introduction: Risk stratification is central to perioperative decision-making and audit. The Physiological and Operative Severity Score for the Enumeration of Mortality and Morbidity (POSSUM) and its Portsmouth Modification (P-POSSUM) are widely used. Yet performance can vary by population and patient-procedure profile. Aim: To evaluate the calibration and discrimination of POSSUM and P-POSSUM in elective general surgery at a tertiary Indian centre. Materials and Methods: The present prospective observational study was conducted in the Department of General Surgery at Dr. D. Y. Patil Medical College, Hospital and Research Centre, Pune, Maharashtra, India, between January 2023 and March 2025. A total of 200 consecutive adults undergoing elective, non-cardiac general surgery were included. Primary endpoints were postoperative morbidity and 30-day all-cause mortality. Performance was assessed using Observed versus Expected (O:E) rates, Hosmer-Lemeshow goodness-of-fit, and Area Under the Receiver Operating Characteristic curve (AUROC). Diagnostic metrics {sensitivity, specificity, Positive Predictive Value (PPV), Negative Predictive Value (NPV)} were calculated. Results: The mean age of the cohort was 52.4±14.8 years, with a predominance of males (male-to-female ratio 1.67:1). Observed morbidity was 52/200 (26.0%; 95% Confidence Interval (CI) 20.0-32.7) and 30-day mortality 8/200 (4.0%; 95% CI 1.8-7.8). POSSUM-predicted morbidity was 67/200 (33.5%; 95% CI 27.0-40.3) and mortality 12/200 (6%; 95% CI 3.1-10.0), giving O:E values of 0.78 (morbidity) and 0.68 (mortality). P-POSSUM predicted mortality 8/200 (4%; 95% CI 1.7-7.3) with an O:E of 0.96. Calibration by Hosmer-Lemeshow was acceptable for all models with p>0.85. P-POSSUM demonstrated superior discrimination for mortality compared with POSSUM (AUC 0.893, 95% CI 0.812-0.974 vs 0.841, 95% CI 0.732-0.950; p=0.042). For morbidity, the original POSSUM morbidity model showed moderate performance (AUC 0.772, 95% CI 0.689-0.855). Conclusion: In elective general surgery, POSSUM remains adequate for morbidity estimation but overpredicts mortality. P-POSSUM shows near-unity calibration and better discrimination for mortality and should be preferred for perioperative risk stratification and audit in similar populations.