Key result
Random forest models outperform decision trees and regression, predicting colon surgery costs with ~21% error.
Why the study?
Hospitals face challenges managing cost-affecting treatment parameters, particularly when comorbidities hinder traditional prognoses, prompting investigation into whether machine learning algorithms can predict colon surgery costs.
Observational (n=347)
No
Absolute Event Rate: 21.4% vs 25.2%
A random forest machine learning model can predict the costs of colon surgery with reasonable accuracy (MAPE ~21.4%), highlighting the potential for ML to assist hospitals in financial planning and cost management.
May assist hospitals in forecasting colon surgery costs; leaves open prospective validation before routine use.
Hospitals are burdened with predicting, calculating, and managing various cost-affecting parameters regarding patients and their treatments. Accuracy in cost prediction is further affected when a patient suffers from other health issues that hinder the traditional prognosis. This can lead to an unavoidable deficit in the final revenue of medical centers. This study aims to determine whether machine learning (ML) algorithms can predict cost factors based on patients undergoing colon surgery. For the forecasting, multiple predictors will be taken into the model to provide a tool that can be helpful for hospitals to manage their costs, ultimately leading to operating more cost-efficiently. This proof of principle will lay the groundwork for an efficient ML-based prediction tool based on multicenter data from a range of international centers in the subsequent phases of the study. With a mean absolute percentage error result of 18%-25.6%, our model's prediction showed decent results in forecasting the costs regarding various diagnosed factors and surgical approaches. There is an urgent need for further studies on predicting cost factors, especially for cases with anastomotic leakage, to minimize unnecessary hospital costs.
No takes yet. Share an insight, caveat, or question.
Taha et al. (2022) conducted an observational in Colon surgery (n=347). Random forest machine learning model vs. Decision tree and generalized boosted regression models was evaluated on Mean absolute percentage error (MAPE) for predicting final costs (95% CI 17.2-26.8). A random forest machine learning model predicted the final costs of colon surgery with a mean absolute percentage error of 21.4%, outperforming decision tree and generalized boosted regression models.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: