Elsevier

Decision Support Systems

Volume 52, Issue 3, February 2012, Pages 698-705
Decision Support Systems

An analytic approach to better understanding and management of coronary surgeries

https://doi.org/10.1016/j.dss.2011.11.004Get rights and content

Abstract

Demand for high-quality, affordable healthcare services increasing with the aging population in the US. In order to cope with this situation, decision makers in healthcare (managerial, administrative and/or clinical) need to be increasingly more effective and efficient at what they do. Along with expertise, information and knowledge are the other key sources for better decisions. Data mining techniques are becoming a popular tool for extracting information/knowledge hidden deep into large healthcare databases. In this study, using a large, feature-rich, nationwide inpatient databases along with four popular machine learning techniques, we developed predictive models; and using an information fusion based sensitivity analysis on these models, we explained the surgical outcome of a patient undergoing a coronary artery bypass grafting. In this study, support vector machines produced the best prediction results (87.74%) followed by decision trees and neural networks. Studies like this illustrate the fact that accurate prediction and better understanding of such complex medical interventions can potentially lead to more favorable outcomes and optimal use of limited healthcare resources.

Highlights

► Demand for high-quality, affordable healthcare services increasing with the aging population in the US. ► Decision makers in healthcare need to be increasingly more effective and efficient at what they do. ► Analytic techniques should be used to augment accurate and timely decision making in healthcare. ► Machine learning techniques can be used to accurately predict CABG surgery outcome. ► Sensitivity analysis on trained models can be used to effectively explain prognostic factors.

Introduction

In recent years, healthcare has become one of the most spoken issues that have a direct impact on quality of life in the US and abroad. While the demand for healthcare services is increasing with the aging population, the supply side is having serious problems to keep up with the needed level and quality of service. In order to close the gap, healthcare systems ought to significantly improve their operational efficacy (i.e., effectiveness and efficiency). Effectiveness (doing the right thing, e.g., diagnosing and treating accurately) and efficiency (doing it the right way, e.g., using least amount of resources and using the least amount of time) are the two fundamental pillars upon which the healthcare system can be revived [16]. One promising way to improve the healthcare efficacy is to take advantage of advanced modeling techniques and large and feature rich data sources (true reflections of medical and healthcare experiences) to support accurate and timely decision making [16].

Decisions in healthcare can roughly be classified as either managerial or clinical. Managerial decisions involve optimal allocation of resources to various demand centers (which at the more detail level involves forecasting, capacity planning, scheduling, etc.). Medical decisions involve making the right diagnosis and implementing the right treatment. Computer systems that help in making medical decisions are often called medical decision support systems or clinical decision support systems. Clinical decision support systems (CDSS) are interactive computer programs, which are designed to assist physicians and other healthcare professionals with decision making tasks. The main goal of a CDSS is to link health observations (i.e., data) with health knowledge (i.e., expertise) to guide health choices by clinicians and managers for an improved healthcare. These CDSS systems encompass knowledge to guide the decision maker in optimal analysis of patient's condition, and subsequently leading to accurate diagnosis and treatment. The knowledge component of these systems may come from the experts (via expert systems) and/or it can be extracted from historical data sources (via data mining).

According to the American Heart Association, cardiovascular disease (CVD) was the underlying cause for 20.67% of deaths in the US [14]. Since 1900, CVD has been the number one killer every year except 1918, which was the year of the great flu pandemic. CVD kills more people than the next four leading causes of deaths combined; cancer, chronic lower respiratory disease, accidents and diabetes mellitus. Out of all CVD deaths, more than half of them are attributed to coronary diseases. Not only does CVD take a huge toll on the personal health and well-being of the population, it is also a great drain on the healthcare resources in the US and elsewhere in the world. The direct and indirect costs associated with CVD for a year is estimated to in excess of $500 billion [14]. Even though the cost of a coronary artery bypass surgery depends on the patient and service provider related factors [6], the average rate is between $50,000 and $100,000 in the US [6].

As is the case in many complex problem situations, the outcome, if predicted accurately beforehand, could lead to better decision making. Especially the prediction of survival time is a clinically important and challenging problem [18], [32]. In the case of cardiac surgery, which in itself poses great risk to the patient, this becomes a matter of life and death. Today coronary artery bypass surgery (CABG) is a common surgical procedure thousands being performed every year in the US and more around the world. CABG surgery is advised for selected groups of patients with significant narrowing and blockage of the heart arteries. CABG surgery creates new routes around narrowed and blocked arteries, allowing sufficient blood flow to deliver oxygen and nutrients to the heart muscles [1], [15]. The coronary artery disease happens with hardening of the arteries or plaque buildup on the walls of the arteries. Smoking can increase this plaque, as well as high blood pressure, high cholesterol, diabetes, among many others. Age also increases the risk, as well as similar family history. When the arteries narrow, the blood supply to the heart is not enough to meet increased oxygen demand. The heart muscle becomes starved of oxygen, and can cause chest pain or death of heart tissue [15].

In this paper, we report on a study where we employed various data mining methods to predict the outcome of coronary artery bypass surgery (CABG), and applied an information fusion based sensitivity analysis on the trained models to better understand the importance of the prognostic factors. Our main goal was to illustrate that predictive and explanatory analysis of large and future rich data sets provides invaluable information to make more efficient and effective decisions in healthcare. Machine learning-based prediction models integrated into a CDSS that provides accurate prediction of the procedure and help practitioners better understand the causal effects of the patient prognostic factors can improve the timeliness of the procedure and optimal allocation of resource, and hence, can improve the chances of success/survival for a significantly more number of patients. The rest of the paper is organized as follows. The next section provides a review of the relevant literature of this medical prediction domain. Section 3 describes the methodology (i.e., data, prediction model types and the evaluation methods used in the study) followed by Section 4 which provides the prediction and sensitivity analysis results. Finally, Section 5 summarizes the study, discusses the findings, and identifies the limitations and future research directions.

Section snippets

Literature review

In the last couple of decades, there have been numerous studies on the risk of death with CABG, but only recently a number of large single-center and multi-center cardiac surgical databases were established for use in risk stratification models. These databases typically contain patient and disease characteristics, of which inclusion (or exclusion) in a statistical method or algorithm could potentially influence the predictive power of the model. The committee led by K.A. Eagle to revise the

Research method

At the highest level, the methodology that we followed in this study can be depicted as a 4-step process as illustrated in Fig. 1. Step 1 involves data identification, data acquisition, and data preparation related tasks. The outcome of this step is a pre-processed dataset which can be used in the rest of the steps. Step 2 involves the model building and model calibration tasks. Here we used four different types of models (artificial neural networks, support vector machines, and two types of

Prediction results

To compare the classification models, three performance criteria are adopted as follows:Accuracy=TP+TNTP+TN+FP+FNSensitivity=TPTP+FNSpecificity=TNTN+FPwhere TP, TN, FP, FN denote true positive, true negative, false positive, and false negative, respectively. Accuracy, shown by Eq. (1), measures the proportion of correctly classified test examples, therefore predicting the overall probability of the correct classification. Sensitivity and specificity, shown by Eqs. (2), (3) respectively, measure

Discussion and conclusions

Data mining, as a knowledge discovery tool, is becoming a popular enabler for improving the decision performance and hence the effectiveness and efficiency in healthcare. Better information and knowledge provided to the right person at the right time in a healthcare setting will undoubtedly lead to better decisions and more favorable outcomes. Such data/fact driven information/knowledge resources packaged in a user-friendly decision support system can be an invaluable aid to both managerial as

Dr. Dursun Delen is the William S. Spears Chair in Business Administration and Associate Professor of Management Science and Information Systems in the Spears School of Business at Oklahoma State University (OSU). He received his Ph.D. in Industrial Engineering and Management from OSU in 1997. Prior to his appointment as an Assistant Professor at OSU in 2001, he worked for a privately-owned research and consultancy company, Knowledge Based Systems Inc., in College Station, Texas, as a research

References (34)

  • M. Rowan

    The use of artificial neural networks to stratify the length of stay of cardiac patients based on preoperative and initial postoperative factors

    Artificial Intelligence in Medicine

    (2007)
  • A. Saltelli

    Making best use of model evaluations to compute sensitivity indices

    Computer Physics Communications

    (2002)
  • D. Sheppard et al.

    Predicting cytomegalovirus disease after renal transplantation: an artificial neural network approach

    International Journal of Medical Informatics

    (1999)
  • R. Wilson et al.

    Bankruptcy prediction using neural networks

    Decision Support Systems

    (1994)
  • BCBS

    MedBrief: Coronary Artery Bypass Graft Surgery (CABG) for the Treatment of Coronary Artery Disease (CAD)

  • L. Breiman et al.

    Classification and Regression Trees

    (1984)
  • B. Bridgewater

    Predicting operative risk for coronary artery surgery in the United Kingdom: a comparison of various risk prediction algorithms

    Heart

    (1998)
  • Cited by (54)

    • The influence of hemodynamics on graft patency prediction model based on support vector machine

      2020, Journal of Biomechanics
      Citation Excerpt :

      In the field of cardiovascular diseases, most researchers use the image data obtained from coronary angiography, CTA and IVUS as well as the collected ECG data as characteristics, and use SVM to carry out diagnosis and postoperative effect prediction (Ismail et al., 2010a,b; Araki et al., 2015; Azam et al., 2017; Matheny et al., 2007; Jawaid et al., 2017). Among them, Dursun et al. used a variety of machine learning methods to predict the surgical results of CABG, and the results showed that SVM worked best (Delen et al., 2012). However, hemodynamic factors were not included in the selection of characteristics.

    View all citing articles on Scopus

    Dr. Dursun Delen is the William S. Spears Chair in Business Administration and Associate Professor of Management Science and Information Systems in the Spears School of Business at Oklahoma State University (OSU). He received his Ph.D. in Industrial Engineering and Management from OSU in 1997. Prior to his appointment as an Assistant Professor at OSU in 2001, he worked for a privately-owned research and consultancy company, Knowledge Based Systems Inc., in College Station, Texas, as a research scientist for five years, during which he led a number of decision support and other information systems related research projects funded by federal agencies, including DoD, NASA, NIST and DOE. His research has appeared in major journals including Decision Support Systems, Communications of the ACM, Computers and Operations Research, Computers in Industry, Journal of Production Operations Management, Artificial Intelligence in Medicine, Expert Systems with Applications, among others. He recently published three books: Advanced Data Mining Techniques with Springer, 2008; Decision Support and Business Intelligence Systems with Prentice Hall, 2010; and Business Intelligence: A Managerial Approach, with Prentice Hall, 2010. He is often invited to national and international conferences for keynote addresses on topics related to Data/Text Mining, Business Intelligence, Decision Support Systems, and Knowledge Management. He served as the general co-chair for the 4th International Conference on Network Computing and Advanced Information Management (September 2–4, 2008 in Seoul, South Korea), and regularly chairs tracks and mini-tracks at various information systems conferences. He is the associate editor-in-chief for International Journal of Experimental Algorithms, associate editor for International Journal of RF Technologies, and is on editorial boards of five other technical journals. His research and teaching interests are in data and text mining, decision support systems, knowledge management, business intelligence and enterprise modeling.

    Dr. Asil Oztekin received the B.S. degree from Yildiz Technical University, Istanbul, Turkey, in 2004, and the M.S. degree from Fatih University, Istanbul, Turkey, in 2006, both in industrial engineering. He also completed the Ph.D. degree in the School of Industrial Engineering and Management at Oklahoma State University (OSU), Stillwater, in 2010. He joined the Robert J. Manning School of Business at University of Massachusetts Lowell in July 2011. Prior to joining UMass Lowell, he worked as a Visiting Assistant Professor in the Department of Statistics at Oklahoma State University. His research interests include medical informatics, healthcare engineering, decision analysis, multivariate data analysis, data mining, quality engineering, and human-computer interaction. His work has been published in leading journals such as Decision Support Systems, International Journal of Production Research, Production Planning & Control, International Journal of Medical Informatics, Artificial Intelligence in Medicine, International Journal of Industrial Ergonomics, and etc. He edited a special issue “Intelligent Computational Techniques in Science, Engineering, and Business” in Expert Systems with Applications journal. Dr. Oztekin has reviewed manuscripts for IIE Transactions, Decision Support Systems, International Journal of Production Research, Production Planning & Control, Computers in Industry, Journal of Systems and Software, Intelligent Information Management, and Applied Clinical Informatics. Dr. Oztekin is a member of ASQ, IIE, and INFORMS and was the recipient of the Alpha Pi Mu Outstanding Industrial Engineering and Management Research Assistant Award from OSU in 2009.

    Dr. Leman Tomak is an Associate Professor of Biostatistics in the College of Medicine at Ondokuz Mayis University, Samsun, Turkey. She received her M.D. degree from the College of Medicine in Ondokuz Mayis University in 1995. After obtaining her degree, she worked at several hospitals and medical centers for three years on Public Healthcare and Medicine under the supervision of the College of Medicine at Ondokuz Mayis University until 2003. She then obtained a Master of Biostatistics degree from the College of Medicine at Ondokuz Mayis University in 2008. As part of her graduate study, she has completed two theses. The subject of the first thesis is the nutrition in adolescents and the relationship between nutrition and development. The subject of the second study was quality control and quality control techniques in clinic chemistry laboratory. She is currently working in Biostatistics Department as a full-time faculty member in the College of Medicine at Ondokuz Mayis University. She has published in leading medical informatics and biostatistics journals. Her research interests include analytical and statistical techniques in medicine and healthcare and quality control techniques in clinical chemistry.

    View full text