Preoperative CR-POSSUM Predicts Mortality but Not Length of Stay in Colorectal Cancer: A Retrospective Cohort Study at Prof. Dr. I.G.N.G. Ngoerah General Hospital, Bali, Indonesia

Abstract

Background: Preoperative risk stratification is central to perioperative internal-medicine practice, yet whether mortality-oriented surgical scores can be repurposed to predict hospital resource use is unclear. Colorectal cancer (CRC) is a leading cause of cancer death in Southeast Asia, and prolonged postoperative length of stay (PLOS) carries clinical and economic cost.

Objective: To examine whether the Colorectal-POSSUM (CR-POSSUM) score predicts prolonged PLOS, 30-day mortality and morbidity after CRC surgery.

Methods: This retrospective analytical cohort study, reported per STROBE, enrolled 53 consecutive patients undergoing colorectal resection at Prof. Dr. I.G.N.G. Ngoerah General Hospital, Denpasar, between January and March 2025. Proportions used Wilson 95% confidence intervals (CI); associations were tested by chi-square/Fisher test with relative risk (RR), odds ratio (OR) and Cramér's V, and upgraded with multivariable logistic regression, ROC analysis and number-needed-to-treat (NNT).

Results: Prolonged PLOS (≥6 days) occurred in 22 patients (41.5%; 95% CI 29.3–54.9). CR-POSSUM did not predict prolonged PLOS (RR 1.12, 95% CI 0.59–2.12; p=0.728; AUC 0.42) but strongly predicted 30-day mortality (13.2%; RR 6.72, 95% CI 0.87–52.05; p=0.035; AUC 0.93, 95% CI 0.79–1.00; Cohen's d 2.49). Preoperative anaemia (adjusted OR 5.20, 95% CI 1.45–18.69; p=0.012) and in-hospital morbidity (RR 2.11, 95% CI 1.20–3.70; p=0.045) independently predicted prolonged PLOS (Nagelkerke R²=0.276).

Conclusion: CR-POSSUM is a valid mortality predictor but should not be used to guide discharge planning; correcting anaemia and preventing complications are the actionable levers to shorten stay for internists managing colorectal-cancer surgery in Indonesian tertiary practice.