Dataset Information

The use of machine learning for the identification of peripheral artery disease and future mortality risk.

ABSTRACT: A key aspect of the precision medicine effort is the development of informatics tools that can analyze and interpret "big data" sets in an automated and adaptive fashion while providing accurate and actionable clinical information. The aims of this study were to develop machine learning algorithms for the identification of disease and the prognostication of mortality risk and to determine whether such models perform better than classical statistical analyses.Focusing on peripheral artery disease (PAD), patient data were derived from a prospective, observational study of 1755 patients who presented for elective coronary angiography. We employed multiple supervised machine learning algorithms and used diverse clinical, demographic, imaging, and genomic information in a hypothesis-free manner to build models that could identify patients with PAD and predict future mortality. Comparison was made to standard stepwise linear regression models.Our machine-learned models outperformed stepwise logistic regression models both for the identification of patients with PAD (area under the curve, 0.87 vs 0.76, respectively; P = .03) and for the prediction of future mortality (area under the curve, 0.76 vs 0.65, respectively; P = .10). Both machine-learned models were markedly better calibrated than the stepwise logistic regression models, thus providing more accurate disease and mortality risk estimates.Machine learning approaches can produce more accurate disease classification and prediction models. These tools may prove clinically useful for the automated identification of patients with highly morbid diseases for which aggressive risk factor management can improve outcomes.

SUBMITTER: Ross EG

PROVIDER: S-EPMC5079774 | biostudies-literature | 2016 Nov

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

The use of machine learning for the identification of peripheral artery disease and future mortality risk.

Ross Elsie Gyang EG Shah Nigam H NH Dalman Ronald L RL Nead Kevin T KT Cooke John P JP Leeper Nicholas J NJ

Journal of vascular surgery 20160603 5

<h4>Objective</h4>A key aspect of the precision medicine effort is the development of informatics tools that can analyze and interpret "big data" sets in an automated and adaptive fashion while providing accurate and actionable clinical information. The aims of this study were to develop machine learning algorithms for the identification of disease and the prognostication of mortality risk and to determine whether such models perform better than classical statistical analyses.<h4>Methods</h4>Foc ...[more]

PMID: 27266594

Similar Datasets

Project description:BACKGROUND AND AIMS:Reduced kidney function is a risk factor for lower-extremity peripheral artery disease (PAD). However, the associations of novel filtration markers with PAD are yet to be quantified. Moreover, little is known on whether bone-mineral metabolism (BMM) markers are related to incident PAD beyond kidney function. METHODS:Using data from 12,472 participants at baseline (1990-1992) of the Atherosclerosis Risk in Communities (ARIC) study, we comprehensively quantified the associations of kidney related markers with incident PAD (defined as hospitalizations with diagnosis of lower-extremity atherosclerosis, revascularization, or amputation). Kidney related markers of interest included estimated glomerular filtration rate (eGFR) (based on creatinine, cystatin C, and both), cystatin C, beta-2 microglobulin (B2M), and BMM markers (serum fibroblast growth factor 23, parathyroid hormone, calcium, and phosphorus). RESULTS:During a median follow-up of 21 years, 471 participants developed incident PAD. Low eGFR was significantly associated with future PAD risk, with slightly stronger relationship when cystatin C was used (adjusted hazard ratio [HR] 6.3-8.3 for eGFR <30 and 2.4-3.5 for eGFR 30-59 vs. eGFR ?90 mL/min/1.73 m2). Among all filtration markers, B2M had the strongest association with incident PAD (HR for top vs. bottom quartile 2.60 [95% CI: 1.91-3.54] for B2M vs. 1.20 [0.91-1.58] for creatinine-based eGFR). Among BMM markers, only phosphorus remained significant for PAD risk beyond potential confounders, including kidney function (HR 1.47 [1.11-1.94] in top quartile). CONCLUSIONS:Kidney dysfunction was independently associated with future PAD risk, particularly when assessed with cystatin C and B2M. Among the BMM markers tested, phosphorus was most robustly associated with incident PAD beyond kidney function. Our results suggest the potential usefulness of novel filtration markers for PAD risk assessment and the possible role of phosphorus in the pathophysiology of PAD.

Project description:BackgroundPeripheral artery disease (PAD) is underrecognized, undertreated, and understudied: each of these endeavors requires efficient and accurate identification of patients with PAD. Currently, PAD patient identification relies on diagnosis/procedure codes or lists of patients diagnosed or treated by specific providers in specific locations and ways. The goal of this research was to leverage natural language processing to more accurately identify patients with PAD in an electronic health record system compared with a structured data-based approach.MethodsThe clinical notes from a cohort of 6861 patients in our health system whose PAD status had previously been adjudicated were used to train, test, and validate a natural language processing model using 10-fold cross-validation. The performance of this model was described using the area under the receiver operating characteristic and average precision curves; its performance was quantitatively compared with an administrative data-based least absolute shrinkage and selection operator (LASSO) approach using the DeLong test.ResultsThe median (SD) of the area under the receiver operating characteristic curve for the natural language processing model was 0.888 (0.009) versus 0.801 (0.017) for the LASSO-based approach alone (DeLong P<0.0001). The median (SD) of the area under the precision curve was 0.909 (0.008) versus 0.816 (0.012) for the structured data-based approach. When sensitivity was set at 90%, the precision for LASSO was 65% and the machine learning approach was 74%, while the specificity for LASSO was 41% and for the machine learning approach was 62%.ConclusionsUsing a natural language processing approach in addition to partial cohort preprocessing with a LASSO-based model, we were able to meaningfully improve our ability to identify patients with PAD compared with an approach using structured data alone. This model has potential applications to both interventions targeted at improving patient care as well as efficient, large-scale PAD research. Graphic Abstract: A graphic abstract is available for this article.

Project description:BACKGROUND:Prescription opioids account for 40% of all U.S. opioid overdose deaths, and national efforts have intensified to reduce opioid prescriptions. Little is known about the relationship between peripheral artery disease (PAD) and high-risk opioid use. The objectives of this study were to evaluate this relationship and to assess the impact of PAD treatment on opiate use. METHODS:In this retrospective cohort study, the Truven Health MarketScan database (Truven Health Analytics, Ann Arbor, Mich), a deidentified national private insurance claims database, was queried to identify patients with PAD (two or more International Classification of Diseases, Ninth Revision diagnosis codes of PAD ?2 months apart, with at least 2 years of continuous enrollment) from 2007 to 2015. Critical limb ischemia (CLI) was defined as the presence of rest pain, ulcers, or gangrene. The primary outcome was high opioid use, defined as two or more opioid prescriptions within a 1-year period. Multivariable analysis was used to determine risk factors for high opioid use. RESULTS:A total of 178,880 patients met the inclusion criteria, 35% of whom had CLI. Mean ± standard deviation follow-up time was 5.3 ± 2.1 years. An average of 24.7% of patients met the high opioid use criteria in any given calendar year, with a small but significant decline in high opioid use after 2010 (P < .01). During years of high opioid use, 5.9 ± 5.5 yearly prescriptions were filled. A new diagnosis of PAD increased high opioid use (21.7% before diagnosis vs 27.3% after diagnosis; P < .001). A diagnosis of CLI was also associated with increased high opioid use (25.4% before diagnosis vs 34.5% after diagnosis; P < .001). Multivariable analysis identified back pain (odds ratio [OR], 1.89; 95% confidence interval [CI], 1.84-1.93; P < .001) and illicit drug use (OR, 1.87; 95% CI, 1.72-2.03; P < .001) as the highest predictors of high opioid use. A diagnosis of CLI was also associated with higher risk (OR, 1.61; 95% CI, 1.57-1.64; P < .001). A total of 43,443 PAD patients (24.3%) underwent 80,816 PAD-related procedures. After exclusion of periprocedural opioid prescriptions (4.9% of all opioid prescriptions), the yearly percentage of high opioid users increased from 25.8% before treatment to 29.6% after treatment (P < .001). CONCLUSIONS:Patients with PAD are at increased risk for high opioid use, with nearly one-quarter meeting described criteria. CLI and treatment for PAD additionally increase high opioid use. In addition to heightened awareness and active opioid management, our findings warrant further investigation into underlying causes and deterrents of high-risk opioid use.

Project description:ObjectivesAmong changes in demographics, aging is the most relevant cardiovascular risk factor. The prevalence of peripheral artery disease (PAD) is high in elderly patients and is associated with a worse prognosis. Despite optimal treatments, mortality in the high-risk population of octo- and nonagenarians with PAD remains excessive, and predictive factors need to be identified. The objective of this study was to investigate predictors of mortality in octo- and nonagenarians with PAD.MethodsCases of treated octo- and nonagenarians, including the clinical characteristics and markers of myocardial injury and heart failure, were studied retrospectively with respect to all-cause mortality. Hazard ratios [HR] were calculated and survival was analyzed by Kaplan-Meyer curves and receiver operating characteristic curved were assessed for troponin-ultra and N-terminal pro-brain natriuretic peptide (NT-proBNP) levels and chronic limb-threatening ischemia (CLTI).ResultsA total of 123 octo- and nonagenarians admitted for PAD were eligible. The troponin level was the major predictor of all-cause mortality (HR: 4.6, 95% confidence interval [CI]: 1.4-15.3), followed by the NT-proBNP level (HR: 3.9, 95% CI 1.8-8.8) and CLTI (HR: 3.1, 95% CI 1.6-5.9). Multivariate regression revealed that each increment of 1 standard deviation in log troponin and log NT-proBNP was associated with a 2.7-fold (95% CI 1.8-4.1) and a 1.9-fold (95% CI 1.2-2.9) increased risk of all-cause death. Receiver operating characteristic curve analysis using a combination of all predictors yielded an improved area under the curve of 0.888. In a control group of an equal number of younger individuals, only NT-proBNP (HR: 4.2, 95% CI 1.2-14.1) and CLTI (HR: 6.1, 95% CI 1.6-23.4) were predictive of mortality.ConclusionOur study demonstrates that cardiovascular biomarkers and CLTI are the primary predictors of increased mortality in elderly PAD patients. Further risk stratification through biomarkers in this high-risk population of octo- and nonagenarians with PAD is necessary.

Project description:Machine learning (ML) and deep learning (DL) can successfully predict high prevalence events in very large databases (big data), but the value of this methodology for risk prediction in smaller cohorts with uncommon diseases and infrequent events is uncertain. The clinical course of spontaneous coronary artery dissection (SCAD) is variable, and no reliable methods are available to predict mortality. Based on the hypothesis that machine learning (ML) and deep learning (DL) techniques could enhance the identification of patients at risk, we applied a deep neural network to information available in electronic health records (EHR) to predict in-hospital mortality in patients with SCAD. We extracted patient data from the EHR of an extensive urban health system and applied several ML and DL models using candidate clinical variables potentially associated with mortality. We partitioned the data into training and evaluation sets with cross-validation. We estimated model performance based on the area under the receiver-operator characteristics curve (AUC) and balanced accuracy. As sensitivity analyses, we examined results limited to cases with complete clinical information available. We identified 375 SCAD patients of which mortality during the index hospitalization was 11.5%. The best-performing DL algorithm identified in-hospital mortality with AUC 0.98 (95% CI 0.97-0.99), compared to other ML models (P < 0.0001). For prediction of mortality using ML models in patients with SCAD, the AUC ranged from 0.50 with the random forest method (95% CI 0.41-0.58) to 0.95 with the AdaBoost model (95% CI 0.93-0.96), with intermediate performance using logistic regression, decision tree, support vector machine, K-nearest neighbors, and extreme gradient boosting methods. A deep neural network model was associated with higher predictive accuracy and discriminative power than logistic regression or ML models for identification of patients with ACS due to SCAD prone to early mortality.

Dataset Information

The use of machine learning for the identification of peripheral artery disease and future mortality risk.

Publications

The use of machine learning for the identification of peripheral artery disease and future mortality risk.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets