Derivation and Validation of a Simplified Risk Prediction Model for Patients undergoing Major Abdominal Surgeries
Author(s): Shashank Agrawal, Pramod Kumar Mishra, Raghav Bansal, Amit Jain
Introduction: Pre-operative risk stratification is useful for resource allocation, for shared decision-making, and informed consent. Pre-operative risk prediction is not widely used due to the subjective nature of available models, the requirement of additional pre-operative investigations, or complex calculations.
Objective: We aimed to derive and internally validate a simple risk prediction model, which could address most possible aspects of patient assessment preoperatively and help clinicians to predict postoperative morbidity or 30 days mortality with reasonable accuracy in patients undergoing Laparotomy.
Methodology: A prospective and retrospective observational study was carried out in a tertiary care hospital. Two hundred retrospective and 101 prospective patients of age more than 18 years, who had undergone Emergency or Elective laparotomy were included. In 1st stage, Patient data like demographics, comorbidities, physiological data, laboratory results, and surgical details were collected in a retrospective manner from 2016-2019. The outcome of the patient including details of post-operative 30 days mortality and morbidities was recorded. A simplified risk prediction model was derived using regression analysis and considering the odds ratio. A 10-point score so derived labeled as “Simplified Max Score (SMS)” was validated in a prospective manner (2019-2020).
Results: Serum Urea, Serum Albumin, Neutrophils to Lymphocytes ratio, METS, and the presence of CVA, CLD, and COPD were the most significant predictors per the retrospective cohort (n=200). On this basis, a Simplified Max Score (SMS) was derived. The derived formula had an AUROC of 0.801 for morbidity and 0.935 for mortality in the retrospective cohort. Results were validated in the prospective cohort (n=101) which showed acceptable reproducibility with AUROC of 0.99 for morbidity and 0.64 for mortality. SMS showed good predictability with AUROC of 0.804 for morbidity and 0.86 for 30 days post-operative mortality, when applied to an entire cohort of 301 patients. SMS also performed better than the American Society of Anesthesiologist's score.
Conclusion: The 10-point SMS score gives a simplified prediction of both postoperative morbidity and mortality after laparotomy.