A decision support system to facilitate management of patients with acute gastrointestinal bleeding

OBJECTIVE To develop a model to predict the bleeding source and identify the cohort amongst patients with acute gastrointestinal bleeding (GIB) who require urgent intervention, including endoscopy. Patients with acute GIB, an unpredictable event, are most commonly evaluated and managed by non-gastroenterologists. Rapid and consistently reliable risk stratification of patients with acute GIB for urgent endoscopy may potentially improve outcomes amongst such patients by targeting scarce healthcare resources to those who need it the most. DESIGN AND METHODS Using ICD-9 codes for acute GIB, 189 patients with acute GIB and all available data variables required to develop and test models were identified from a hospital medical records database. Data on 122 patients was utilized for development of the model and on 67 patients utilized to perform comparative analysis of the models. Clinical data such as presenting signs and symptoms, demographic data, presence of co-morbidities, laboratory data and corresponding endoscopic diagnosis and outcomes were collected. Clinical data and endoscopic diagnosis collected for each patient was utilized to retrospectively ascertain optimal management for each patient. Clinical presentations and corresponding treatment was utilized as training examples. Eight mathematical models including artificial neural network (ANN), support vector machine (SVM), k-nearest neighbor, linear discriminant analysis (LDA), shrunken centroid (SC), random forest (RF), logistic regression, and boosting were trained and tested. The performance of these models was compared using standard statistical analysis and ROC curves. RESULTS Overall the random forest model best predicted the source, need for resuscitation, and disposition with accuracies of approximately 80% or higher (accuracy for endoscopy was greater than 75%). The area under ROC curve for RF was greater than 0.85, indicating excellent performance by the random forest model. CONCLUSION While most mathematical models are effective as a decision support system for evaluation and management of patients with acute GIB, in our testing, the RF model consistently demonstrated the best performance. Amongst patients presenting with acute GIB, mathematical models may facilitate the identification of the source of GIB, need for intervention and allow optimization of care and healthcare resource allocation; these however require further validation.

[1]  D. Podolsky,et al.  Physician specialty and variations in the cost of treating patients with acute upper gastrointestinal bleeding. , 1997, Gastroenterology.

[2]  J M Tenías Burillo,et al.  Pre-endoscopic prognostic factors in non-varicose upper gastrointestinal bleeding. Development of a predictive algorithm. , 2002, Revista espanola de enfermedades digestivas : organo oficial de la Sociedad Espanola de Patologia Digestiva.

[3]  Yu-Chuan Li,et al.  Stratification of Adverse Outcomes by Preoperative Risk Factors in Coronary Artery Bypass Graft Patients: An Artificial Neural Network Prediction Model , 2003, AMIA.

[4]  S. Tenner,et al.  Early Intensive Resuscitation of Patients With Upper Gastrointestinal Bleeding Decreases Mortality , 2001, American Journal of Gastroenterology.

[5]  Helmut Messmann,et al.  Risk factors for mortality in severe upper gastrointestinal bleeding , 2004, International Journal of Colorectal Disease.

[6]  James J. Chen,et al.  Classification by ensembles from random partitions of high-dimensional data , 2007, Comput. Stat. Data Anal..

[7]  R. Schapire.  The Strength of Weak Learnability , 1990, Machine Learning.

[8]  J. Elashoff,et al.  Upper gastrointestinal hemorrhage clinical--guideline determining the optimal hospital length of stay. , 1996, The American journal of medicine.

[9]  Vladimir N. Vapnik,et al.  The Nature of Statistical Learning Theory , 2000, Statistics for Engineering and Information Science.

[10]  Amitabh Chak,et al.  Prediction of outcome in acute lower-gastrointestinal haemorrhage based on an artificial neural network: internal and external validation of a predictive model , 2003, The Lancet.

[11]  G. Swingler,et al.  Clinical predictors of outcome in acute upper gastrointestinal bleeding. , 2003, South African medical journal = Suid-Afrikaanse tydskrif vir geneeskunde.

[12]  Ann Williamson,et al.  Early predictors of severe lower gastrointestinal bleeding and adverse outcomes: a prospective study. , 2004, Clinical gastroenterology and hepatology : the official clinical practice journal of the American Gastroenterological Association.

[13]  O. Blatchford,et al.  A risk score to predict need for treatment for uppergastrointestinal haemorrhage , 2000, The Lancet.

[14]  G. Elta,et al.  Urgent colonoscopy for acute lower-GI bleeding. , 2004, Gastrointestinal endoscopy.

[15]  Andreas M. Stefan,et al.  Early Indicators of Prognosis in Upper Gastrointestinal Hemorrhage , 1998, American Journal of Gastroenterology.

[16]  Elizabeth Rajan,et al.  ASGE guideline: The role of endoscopy in acute non-variceal upper-GI hemorrhage. , 2004, Gastrointestinal endoscopy.

[17]  Paulo J. G. Lisboa,et al.  A review of evidence of health benefit from artificial neural networks in medical intervention , 2002, Neural Networks.

[18]  E John Orav,et al.  Early predictors of severity in acute lower intestinal tract bleeding. , 2003, Archives of internal medicine.

[19]  K. Palmer,et al.  Non-variceal upper gastrointestinal haemorrhage: guidelines , 2002, Gut.

[20]  Noam Harpaz,et al.  Artificial neural networks distinguish among subtypes of neoplastic colorectal lesions. , 2002, Gastroenterology.

[21]  Ananya Das,et al.  Prediction of outcome of acute GI hemorrhage: a review of risk scores and predictive models. , 2004, Gastrointestinal endoscopy.

[22]  T. C. Northfield,et al.  Selection of patients for early discharge or outpatient care after acute upper gastrointestinal haemorrhage , 1996, The Lancet.

[23]  M. D. Martínez-Miranda,et al.  Computational Statistics and Data Analysis , 2009 .

[24]  Richard C K Wong.  Acute GI bleeding: upper and lower, but is there a "middle kingdom"? , 2003, Gastrointestinal endoscopy.

[25]  A. Ellrodt,et al.  Prospective evaluation of a clinical guideline recommending hospital length of stay in upper gastrointestinal tract hemorrhage. , 1997, JAMA.

[26]  H B Devlin,et al.  Risk assessment after acute upper gastrointestinal haemorrhage. , 1996, Gut.

[27]  Lars H Lund,et al.  Computerized interpretation of the electrocardiogram. , 2004, Archives of internal medicine.

[28]  H B Devlin,et al.  Incidence of and mortality from acute upper gastrointestinal haemorrhage in the United Kingdom , 1995 .

[29]  J G Dolan,et al.  Early clinical signs identify low-risk patients with acute upper gastrointestinal hemorrhage. , 1985, JAMA.

[30]  Emanuel F Petricoin,et al.  Serum proteomics in cancer diagnosis and management. , 2004, Annual review of medicine.

[31]  P. Mortensen,et al.  The diagnostic value of serum urea/creatinine ratio in distinguishing between upper and lower gastrointestinal bleeding. A prospective study. , 1994, Danish medical bulletin.

[32]  Alan Barkun,et al.  Consensus Recommendations for Managing Patients with Nonvariceal Upper Gastrointestinal Bleeding , 2003, Annals of Internal Medicine.

[33]  R Beeri,et al.  Predictors of mortality in patients admitted to hospital for acute upper gastrointestinal hemorrhage. , 1995, Scandinavian journal of gastroenterology.

[34]  Annette M. Molinaro,et al.  Prediction error estimation: a comparison of resampling methods , 2005, Bioinform..

[35]  G. Zuckerman,et al.  Acute small bowel bleeding: a distinct entity with significantly different economic implications compared with GI bleeding from other locations. , 2003, Gastrointestinal endoscopy.

[36]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[37]  D. Bamber.  The area above the ordinal dominance graph and the area below the receiver operating characteristic graph , 1975 .

[38]  H S Fraser,et al.  An artificial neural network system for diagnosis of acute myocardial infarction (AMI) in the accident and emergency department: evaluation and comparison with serum myoglobin measurements. , 1997, Computer methods and programs in biomedicine.

[39]  Leo Breiman,et al.  Bagging Predictors , 1996, Machine Learning.

[40]  J Vandewalle,et al.  Artificial neural network models for the preoperative discrimination between malignant and benign adnexal masses , 1999, Ultrasound in obstetrics & gynecology : the official journal of the International Society of Ultrasound in Obstetrics and Gynecology.

[41]  R. Tibshirani,et al.  Diagnosis of multiple cancer types by shrunken centroids of gene expression , 2002, Proceedings of the National Academy of Sciences of the United States of America.

[42]  J. Terdiman,et al.  Risk of persistent or recurrent and intractable upper gastrointestinal bleeding in the era of therapeutic endoscopy. , 1997, The American journal of gastroenterology.

[43]  A ZaragozaMarcet,et al.  Pre-endoscopic prognostic factors in non-varicose upper gastrointestinal bleeding. Development of a predictive algorithm. , 2002 .

[44]  R Beeri,et al.  Predictors of Mortality in Patients Admitted to Hospital for Acute Upper Gastrointestinal Hemorrhage , 1995 .

[45]  H B Devlin,et al.  Incidence of and mortality from acute upper gastrointestinal haemorrhage in the United Kingdom. Steering Committee and members of the National Audit of Acute Upper Gastrointestinal Haemorrhage. , 1995, BMJ.