A Machine Learning Approach for Predicting Real-time Risk of Intraoperative Hypotension in Traumatic Brain Injury.

BACKGROUND: Traumatic brain injury (TBI) is a major cause of death and disability. Episodes of hypotension are associated with worse TBI outcomes. Our aim was to model the real-time risk of intraoperative hypotension in TBI patients, compare machine learning and traditional modeling techniques, and identify key contributory features from the patient monitor and medical record for the prediction of intraoperative hypotension. METHODS: The data included neurosurgical procedures in 1005 TBI patients at an academic level 1 trauma center. The clinical event was intraoperative hypotension, defined as mean arterial pressure <65 mm Hg for 5 or more consecutive minutes. Two types of models were developed: one based on preoperative patient-level predictors and one based on intraoperative predictors measured per minute. For each of these models, we took 2 approaches to predict the occurrence of a hypotensive event: a logistic regression model and a gradient boosting tree model. RESULTS: The area under the receiver operating characteristic curve for the intraoperative logistic regression model was 0.80 (95% confidence interval [CI]: 0.78-0.83), and for the gradient boosting model was 0.83 (95% CI: 0.81-0.85). The area under the precision-recall curve for the intraoperative logistic regression model was 0.16 (95% CI: 0.12-0.20), and for the gradient boosting model was 0.19 (95% CI: 0.14-0.24). Model performance based on preoperative predictors was poor. Features derived from the recent trend of mean arterial pressure emerged as dominantly predictive in both intraoperative models. CONCLUSIONS: This study developed a model for real-time prediction of intraoperative hypotension in TBI patients, which can use computationally efficient machine learning techniques and a streamlined feature-set derived from patient monitor data.

Utility of Genetic Testing in Addition to Mammography for Determining Risk of Breast Cancer Depends on Patient Age.

While screening and treatment have sharply reduced breast cancer mortality in the past 50 years, more targeted diagnostic testing may improve the accuracy and efficiency of care. Our retrospective, age-matched, case-control study evaluated the differential value of mammography and genetic variants to predict breast cancer depending on patient age. We developed predictive models using logistic regression with group lasso comparing (1) diagnostic mammography findings, (2) selected genetic variants, and (3) a combination of both. For women older than 60, mammography features were most predictive of breast cancer risk (imaging AUC = 0.74, genetic variants AUC = 0.54, combined AUC = 0.71). For women younger than 60 there is additional benefit to obtaining genetic testing (imaging AUC = 0.69, genetic variants AUC = 0.70, combined AUC = 0.72). In summary, genetic testing supplements mammography in younger women while mammography appears sufficient in older women for breast cancer risk prediction.

Improving breast cancer risk prediction by using demographic risk factors, abnormality features on mammograms and genetic variants.

The predictive capability of combining demographic risk factors, germline genetic variants, and mammogram abnormality features for breast cancer risk prediction is poorly understood. We evaluated the predictive performance of combinations of demographic risk factors, high risk single nucleotide polymorphisms (SNPs), and mammography features for women recommended for breast biopsy in a retrospective case-control study (n = 768) with four logistic regression models. The AUC of the baseline demographic features model was 0.580. Both genetic variants and mammography abnormality features augmented the performance of the baseline model: demographics + SNP (AUC =0.668), demographics + mammography (AUC =0.702). Finally, we found that the demographics + SNP + mammography model (AUC = 0.753) had the greatest predictive power, with a significant performance improvement over the other models. The combination of demographic risk factors, genetic variants and imaging features improves breast cancer risk prediction over prior methods utilizing only a subset of these features.

Multiple postoperative complications: Making sense of the trajectories.

BACKGROUND: Many studies have evaluated predictors of postoperative complications, yet little is known about the development of multiple complications. The goal of this study was to assess complication timing in cascades of multiple complications and the risk of future complications given a patient's first complication. METHODS: This study includes 30-day, postoperative complications from the American College of Surgeons National Surgical Quality Improvement Program for all patients who underwent major inpatient and outpatient operative procedures from 2005-2013. The timing and sequencing of complications were evaluated using χ2 analysis and pairwise comparisons. RESULTS: More severe postoperative complications (cardiac arrest or myocardial infarction, renal insufficiency or failure, stroke, intubation, septic shock, coma) had the greatest impact on the risk for developing further complications, increasing the relative risk of developing future, specific, severe complications by more than 40-fold. These more severe complications occur within a few days of other complications (whether as a preceding factor or an outcome), while less severe complications, such as surgical site infection and urinary tract infection, are linked less tightly to complication cascades. CONCLUSION: This analysis highlights both the risk for secondary complications after an initial complication and when those future complications are likely to occur. Physicians can use this information to target interventions to prevent high-risk complications.

Effects of a Cognitive Behavioral Therapy Intervention Trial to Improve Disease Outcomes in Children with Inflammatory Bowel Disease.

BACKGROUND: Studies testing the efficacy of behavioral interventions to modify psychosocial sequelae of inflammatory bowel disease in children are limited. This report presents outcomes through a 6-month follow-up from a large randomized controlled trial testing the efficacy of a cognitive behavioral intervention for children with inflammatory bowel disease and their parents. METHODS: One hundred eighty-five children aged 8 to 17 years with a diagnosis of Crohn's disease or ulcerative colitis and their parents were randomized to one of two 3-session conditions: (1) a social learning and cognitive behavioral therapy condition or (2) an education support condition designed to control for time and attention. RESULTS: There was a significant overall treatment effect for school absences due to Crohn's disease or ulcerative colitis (P < 0.05) at 6 months after treatment. There was also a significant overall effect after treatment for child-reported quality of life (P < 0.05), parent-reported increases in adaptive child coping (P < 0.001), and reductions in parents' maladaptive responses to children's symptoms (P < 0.05). Finally, exploratory analyses indicated that for children with a higher level of flares (2 or more) prebaseline, those in social learning and cognitive behavioral therapy condition experienced a greater reduction in flares after treatment. CONCLUSIONS: This trial suggests that a brief cognitive behavioral intervention for children with inflammatory bowel disease and their parents can result in improved child functioning and quality of life, and for some children may decrease disease activity.

Modeling the Temporal Evolution of Postoperative Complications.

Post-operative complications have a significant impact on patient morbidity and mortality; these impacts are exacerbated when patients experience multiple complications. However, the task of modeling the temporal sequencing of complications has not been previously addressed. We present an approach based on Markov chain models for characterizing the temporal evolution of post-operative complications represented in the American College of Surgeons National Surgery Quality Improvement Program database. Our work demonstrates that the models have significant predictive value. In particular, an inhomogenous Markov chain model effectively predicts the development of serious complications (coma longer than a day, cardiac arrest, myocardial infarction, septic shock, renal failure, pneumonia) and interventional complications (unplanned re-intubation, longer than 2 days on a ventilator and bleeding transfusion).