An interpretable machine learning model for the prevention of contrast-induced nephropathy in patients undergoing lower extremity endovascular interventions for peripheral arterial disease.

BACKGROUND: Contrast-induced nephropathy (CIN) is a postprocedural complication associated with increased morbidity and mortality. An important risk factor for development of CIN is renal impairment. Identification of patients at risk for acute renal failure will allow physicians to make appropriate decisions to minimize the incidence of CIN. We developed a machine learning model to stratify risk of acute renal failure that may assist in mitigating risk for CIN in patients with peripheral artery disease (PAD) undergoing endovascular interventions. METHODS: We utilized the American College of Surgeons National Surgical Quality Improvement Program database to extract clinical and laboratory information associated with 14,444 patients who underwent lower extremity endovascular procedures between 2011 and 2018. Using 11,604 cases from 2011 to 2017 for training and 2840 cases from 2018 for testing, we developed a random forest model to predict risk of 30-day acute renal failure following infra-inguinal endovascular procedures. RESULTS: Eight variables were identified as contributing optimally to model predictions, the most important being diabetes, preoperative BUN, and claudication. Using these variables, the model achieved an area under the receiver-operating characteristic (AU-ROC) curve of 0.81, accuracy of 0.83, sensitivity of 0.67, and specificity of 0.74. The model performed equally well on white and nonwhite patients (Delong p-value = 0.955) and patients age < 65 and patients age ≥ 65 (Delong p-value = 0.659). CONCLUSIONS: We develop a model that fairly and accurately stratifies 30-day acute renal failure risk in patients undergoing lower extremity endovascular procedures for PAD. This model may assist in identifying patients who may benefit from strategies to prevent CIN.

Risk stratification with explainable machine learning for 30-day procedure-related mortality and 30-day unplanned readmission in patients with peripheral arterial disease.

Predicting 30-day procedure-related mortality risk and 30-day unplanned readmission in patients undergoing lower extremity endovascular interventions for peripheral artery disease (PAD) may assist in improving patient outcomes. Risk prediction of 30-day mortality can help clinicians identify treatment plans to reduce the risk of death, and prediction of 30-day unplanned readmission may improve outcomes by identifying patients who may benefit from readmission prevention strategies. The goal of this study is to develop machine learning models to stratify risk of 30-day procedure-related mortality and 30-day unplanned readmission in patients undergoing lower extremity infra-inguinal endovascular interventions. We used a cohort of 14,444 cases from the American College of Surgeons National Surgical Quality Improvement Program database. For each outcome, we developed and evaluated multiple machine learning models, including Support Vector Machines, Multilayer Perceptrons, and Gradient Boosting Machines, and selected a random forest as the best-performing model for both outcomes. Our 30-day procedure-related mortality model achieved an AUC of 0.75 (95% CI: 0.71-0.79) and our 30-day unplanned readmission model achieved an AUC of 0.68 (95% CI: 0.67-0.71). Stratification of the test set by race (white and non-white), sex (male and female), and age (≥65 years and <65 years) and subsequent evaluation of demographic parity by AUC shows that both models perform equally well across race, sex, and age groups. We interpret the model globally and locally using Gini impurity and SHapley Additive exPlanations (SHAP). Using the top five predictors for death and mortality, we demonstrate differences in survival for subgroups stratified by these predictors, which underscores the utility of our model.

Interpretable Machine Learning for the Prediction of Amputation Risk Following Lower Extremity Infrainguinal Endovascular Interventions for Peripheral Arterial Disease.

PURPOSE: Severe peripheral artery disease (PAD) may result in lower extremity amputation or require multiple procedures to achieve limb salvage. Current prediction models for major amputation risk have had limited performance at the individual level. We developed an interpretable machine learning model that will allow clinicians to identify patients at risk of amputation and optimize treatment decisions for PAD patients. METHODS: We utilized the American College of Surgeons National Surgical Quality Improvement Program database to collect preoperative clinical and laboratory information on 14,444 patients who underwent lower extremity endovascular procedures for PAD from 2011 to 2018. Using data from 2011 to 2017 for training and data from 2018 for testing, we developed a machine learning model to predict 30 day amputation in this patient population. We present performance metrics overall and stratified by race, sex, and age. We also demonstrate model interpretability using Gini importance and SHapley Additive exPlanations. RESULTS: A random forest machine learning model achieved an area under the receiver-operator curve (AU-ROC) of 0.81. The most important features of the model were elective surgery designation, claudication, open wound/wound infection, white blood cell count, and albumin. The model performed equally well on white and non-white patients (Delong p-value = 0.189), males and females (Delong p-value = 0.572), and patients under age 65 and patients age 65 and older (Delong p-value = 0.704). CONCLUSION: We present a machine learning model that predicts 30 day major amputation events in PAD patients undergoing lower extremity endovascular procedures. This model can optimize clinical decision-making for patients with PAD.

Utilization of medical services for the treatment of acute low back pain: conformance with clinical guidelines.

OBJECTIVE: To assess the utilization of diagnostic and therapeutic medical services for the management of acute low back pain in a primary care setting, and to determine whether such utilization conforms to suggested guidelines for the management of this condition. STUDY DESIGN: A retrospective chart audit of consecutive cases of acute low back pain. Specific elements of the diagnostic and therapeutic approach were judged appropriate or inappropriate based on comparison with published recommendations supported by the medical literature. SETTING: The primary care adult practice of a university-affiliated health maintenance organization. PATIENTS: One hundred eighty-three patients presenting with acute low back pain of musculoskeletal origin. MEASUREMENTS AND MAIN RESULTS: According to suggested guidelines for the care of acute low back pain, 26% of plain lumbar x-rays (10/38), 66% of computed tomography (CT) and magnetic resonance imaging (MRI) scans (12/18), and 82% (23/28) of subspecialty referrals were categorized as inappropriate. Among patients without indications for these services, 12% (10/85) had received lumbar x-rays, 7% (12/168) had received lumbar MRI or CT scans, and 14% (23/168) had received subspecialty referrals. Underutilization of these services had occurred in 71% (70/98) of patients with an indication for plain lumbar radiography, and 47% (7/15) of patients with potential indications for surgical referral or CT/MRI scanning. Neither overutilization nor underutilization had led to adverse outcomes or delays in diagnosis in this small sample. CONCLUSIONS: According to guidelines from the medical literature, the primary care physicians in this study both overutilized and underutilized diagnostic and referral services in cases of acute low back pain. It is necessary to determine whether underutilization of plain lumbar radiography adversely affects diagnostic accuracy and whether overutilization of other services improves important clinical outcomes, given the generally benign natural history of this condition.