Predictive model for high cost and complications in Nephrolithiasis patients

Abstract

The average cost of treatment for patients with nephrolithiasis complications was twice that for those without (47,792.08 vs. 22,995.33 Thai Baht, THB). In-patient data for October 2014 to December 2018 were used to create predictive models in R. The C5.0 algorithm was used to create a decision tree to predict patients with high costs (>24,000 THB). The patient care team used the predictive model to revise their standard treatment procedures, and the model was evaluated after 8 months (May 2019–December 2019). In this period, 395 nephrolithiasis patients were treated using new treatment protocols. The average cost for patients with no complications did not change, but the average cost for patients with one or two complications decreased significantly (p < 0.05). Quality of care and patient safety were not affected, complication rates did not increase, but average length of stay increased from 3.72 days to 4.25 days.

References

AbdelRahman, S. E., Zhang, M., Bray, B. E., & Kawamoto, K. (2014). A three-step approach for the derivation and validation of high-performing predictive models using an operational dataset: Congestive heart failure readmission case study. BMC Medical Informatics and Decision Making, 14(41). doi:10.1186/1472-6947-14-41

Akolarikos, S., & de la Rosette, J. (2008). Prevention and treatment of complications following percutaneous nephroli thotomy. Current Opinion in Urology, 8(6), 229–234.

Bureau of Policy and Strategy, Ministry of Public Health, Thailand. (2009). Discharge Summarization Handbook for Medical Doctors. Nonthaburi, Thailand.

Bureau of Policy and Strategy. Bureau of Policy and Strategy, Ministry of Public Health, Thailand. (2015). 43 Standard Dataset for Health Data Center. Nonthaburi, Thailand. Bureau of Policy and Strategy.

El-Nahas, A. R., Eraky, I., Shokier, A. A., Shoma, A. A., El-Assmy, A. A., El-Taey, N. A., . . . El-Kenawy, M. R. (2012). Complications of percutaneous nephrolithomy for treatment of staghorn stone. Urology, 79(6), 1236–1241.

Hilbert, J. P., Zasadil, S., Keyser, D. J., & Peele, P. B. (2017). Using decision trees to manage hospital readmission risk for acute myocardial infarction, heart failure, and pneumonia. Applied Health Economics and Health Policy, 12(6), 573–585.

Khichi, M. H., Parsons, W., Winters-Miners, L. A. (2015). Predicting survival or mortality for patients with disseminated intravascular coagulation (DIC) and/or critical illnesses. In L. A. Winters-Miners (Ed.), Practical predictive analytics and decision support systems in medicine (pp. 558–602). San Diego: Elsevier.

Kamble, S. S., Gunasekaran, A., Gaswami, M., & Manda, J. (2019). A systematic perspective on the applications of big data analytics in healthcare management. International Journal of Healthcare Management, 12(3), 226–240.

Koch, S. (2019). Current trends in biomedical and health informatics. Methods of Information in Medicine, 58(02/03), 61–62.

Liu, B., Li, Y., Ghosh, S., Sun, Z., Ng, K., & Hu, J. (2018). Complication risk profiling in diabetes care: A Bayesian multi-task and feature relationship learning approach. IEEE Transaction of Knowledge and Data Engineering, 20(10), 1–12.

Stille, B., Chadler, J., & Reddy, S. (2016). Algorithm for data science. Cham: Springer International Publishing AG. Talaei-Khoei, A., Tavana, M., & Wilson, J. M. (2019). A predictive analytics framework for identifying patients at risk of developing multiple medical complications caused by chronic diseases. Artificial Intelligence in Medicine, 101. doi: 10.1016/j.artmed.2019.101750

Tanthanu, M., Apiwatgaroon, A., & Pripatnanont, C. (2005). Urinary tract calculi in southern Thailand. Journal of Medical Association of Thailand, 88(1), 80–85.

Thompson, S., Whitaker, J., Kohli, R., & Jones, C. (2019). Chronic disease management: how IT and analytics create healthcare value through the temporal displacement of care. MIS Quarterly, 20(1). Retrieved from https:// misq.org/misq/downloads/download/article/1405

Tortajada, S., Garcia-Gomez, J. M., Vicente, J., Sanjuan, J., de Frutos, R., Martin-Santos, R., . . . Robles, M. (2009). Prediction of postpartum depression using multilayer perceptrons and pruning. Methods of Information in Medicine, 48(03), 291–298.

Turney, B. W., Reynard, J. M., Noble, J., & Keoghane, S. R. (2011). Trends in urological stone disease. British Journal of Urology, 109, 1082–1087.

Varanasi, H. (2015). Predicting breast cancer diagnosis using support vector machines. In L. A. Winters-Miners (Ed.), Practical Predictive Analytics and Decision Support Systems in Medicine (pp. 850–865). San Diego: Elsevier.

World Health Organization. (2016). International Statisti[1]cal Classification of Diseases and Related Health Problems, 10th Revision. (5th ed.). Geneva, Switzerland. World Health Organization.

Yanagawa, M., Kawamura, J., Onishi, T., Soga, N., Kamed, K., Sriboonlue, P., . . . Borwornpadungkitti, S. (1997). Incidence of urolithiasis in northeast Thailand. International Journal of Urology, 4, 537–540.

Ziamba, J. B., & Matlaga, B. R. (2015). Guideline of guidelines: Kidney stones. British Journal of Urology, 116, 184–189