Subscribe to RSS
DOI: 10.1055/s-0039-1693731
Hyper-G: An Artificial Intelligence Tool for Optimal Decision-Making and Management of Blood Glucose Levels in Surgery Patients
Funding None.Publication History
13 March 2019
14 June 2019
Publication Date:
09 August 2019 (online)
![](https://www.thieme-connect.de/media/10.1055-s-00035037/20190203/lookinside/thumbnails/10-1055-s-0039-1693731_190030-1.jpg)
Abstract
Background Hyperglycemia or high blood glucose during surgery is associated with poor postoperative outcome. Knowing in advance which patients may develop hyperglycemia allows optimal assignment of resources and earlier initiation of glucose management plan.
Objective To develop predictive models to estimate peak glucose levels in surgical patients and to implement the best performing model as a point-of-care clinical tool to assist the surgical team to optimally manage glucose levels.
Methods Using a large perioperative dataset (6,579 patients) of patient- and surgery-specific parameters, we developed and validated linear regression and machine learning models (random forest, extreme gradient boosting [Xg Boost], classification and regression trees [CART], and neural network) to predict the peak glucose levels during surgery. The model performances were compared in terms of mean absolute percentage error (MAPE), logarithm of the ratio of the predicted to actual value (log ratio), median prediction error, and interquartile error range. The best performing model was implemented as part of a web-based application for optimal decision-making toward glucose management during surgery.
Results Accuracy of the machine learning models were higher (MAPE = 17%, log ratio = 0.029 for Xg Boost) when compared with that of the linear regression model (MAPE = 22%, log ratio = 0.041). The Xg Boost model had the smallest median prediction error (5.4 mg/dL) and the narrowest interquartile error range (−17 to 24 mg/dL) as compared with the other models. The best performing model, Xg Boost, was implemented as a web application, Hyper-G, which the perioperative providers can use at the point of care to estimate peak glucose levels during surgery.
Conclusions Machine learning models are able to accurately predict peak glucose levels during surgery. Implementation of such a model as a web-based application can facilitate optimal decision-making and advance planning of glucose management strategies.
-
References
- 1 Duncan AE. Hyperglycemia and perioperative glucose management. Curr Pharm Des 2012; 18 (38) 6195-6203
- 2 Sebranek JJ, Lugli AK, Coursin DB. Glycaemic control in the perioperative period. Br J Anaesth 2013; 111 (Suppl. 01) i18-i34
- 3 Kwon S, Thompson R, Dellinger P, Yanez D, Farrohki E, Flum D. Importance of perioperative glycemic control in general surgery: a report from the Surgical Care and Outcomes Assessment Program. Ann Surg 2013; 257 (01) 8-14
- 4 Kotagal M, Symons RG, Hirsch IB. , et al; SCOAP-CERTAIN Collaborative. Perioperative hyperglycemia and risk of adverse events among patients with and without diabetes. Ann Surg 2015; 261 (01) 97-103
- 5 Nair BG, Neradilek MB, Newman SF, Horibe M. Association between acute phase perioperative glucose parameters and postoperative outcomes in diabetic and non-diabetic patients undergoing non-cardiac surgery. Am J Surg 2018 (e-pub ahead of print). doi:10.1016/j.amjsurg.2018.10.024
- 6 Nair BG, Horibe M, Neradilek MB, Newman SF, Peterson GN. The effect of intraoperative blood glucose management on postoperative blood glucose levels in noncardiac surgery patients. Anesth Analg 2016; 122 (03) 893-902
- 7 Russel S, Norvig P. Artificial Intelligence: A Modern Approach. 3rd ed. Chennai: Pearson Education; 2015
- 8 Connor CW. Artificial intelligence and machine learning in anesthesiology. Anesthesiology 2019 (e-pub ahead of print). doi:10.1097/ALN.0000000000002694
- 9 Maier-Hein L, Vedula SS, Speidel S. , et al. Surgical data science for next-generation interventions. Nat Biomed Eng 2017; 1 (09) 691-696
- 10 Cheung DS, Grubenhoff JA. Machine learning in clinical medicine still finding its way. JAMA Netw Open 2019; 2 (01) e186926
- 11 Reymann MP, Dorschky E, Groh BH, Martindale C, Blank P, Eskofier BM. Blood glucose level prediction based on support vector regression using mobile platforms. Conf Proc IEEE Eng Med Biol Soc 2016; 2016: 2990-2993
- 12 Liszka-Hackzell JJ. Prediction of blood glucose levels in diabetic patients using a hybrid AI technique. Comput Biomed Res 1999; 32 (02) 132-144
- 13 Li J, Fernando C. Smartphone-based personalized blood glucose prediction. ICT Express 2016; 2: 150-154
- 14 Pappada SM, Cameron BD, Rosman PM. Development of a neural network for prediction of glucose concentration in type 1 diabetes patients. J Diabetes Sci Technol 2008; 2 (05) 792-801
- 15 Marling C, Wiley M, Bunescu RC, Shubrook J, Schwartz F. Emerging applications for intelligent diabetes management. AI Mag 2012; 33 (02) 67-78
- 16 Zhang Z. A mathematical model for predicting glucose levels in critically-ill patients: the PIGnOLI model. PeerJ 2015; 3: e1005
- 17 Pappada SM, Cameron BD, Tulman DB. , et al. Evaluation of a model for glycemic prediction in critically ill surgical patients. PLoS One 2013; 8 (07) e69475
- 18 Enders CK. Applied Missing Data Analysis. New York, NY: Guilford Press; 2010
- 19 Zhang Z. Variable selection with stepwise and best subset approaches. Ann Transl Med 2016; 4 (07) 136
- 20 Lesmeister C. Mastering Machine Learning with R. 2nd ed. Birmingham: PACKT Publishing; 2017
- 21 Grolemund G. Hands-On Programming with R. Boston, MA: O'Reilly Media; 2014
- 22 American Diabetes Association. 6. Glycemic targets: standards of medical care in diabetes-2018. Diabetes Care 2018; 41 (Suppl. 01) S55-S64