Using a cohort study of diabetes and peripheral artery disease to compare logistic regression and machine learning via random forest modeling.

BACKGROUND: This study illustrates the use of logistic regression and machine learning methods, specifically random forest models, in health services research by analyzing outcomes for a cohort of patients with concomitant peripheral artery disease and diabetes mellitus. METHODS: Cohort study using fee-for-service Medicare beneficiaries in 2015 who were newly diagnosed with peripheral artery disease and diabetes mellitus. Exposure variables include whether patients received preventive measures in the 6 months following their index date: HbA1c test, foot exam, or vascular imaging study. Outcomes include any reintervention, lower extremity amputation, and death. We fit both logistic regression models as well as random forest models. RESULTS: There were 88,898 fee-for-service Medicare beneficiaries diagnosed with peripheral artery disease and diabetes mellitus in our cohort. The rate of preventative treatments in the first six months following diagnosis were 52% (n = 45,971) with foot exams, 43% (n = 38,393) had vascular imaging, and 50% (n = 44,181) had an HbA1c test. The directionality of the influence for all covariates considered matched those results found with the random forest and logistic regression models. The most predictive covariate in each approach differs as determined by the t-statistics from logistic regression and variable importance (VI) in the random forest model. For amputation we see age 85 + (t = 53.17) urban-residing (VI = 83.42), and for death (t = 65.84, VI = 88.76) and reintervention (t = 34.40, VI = 81.22) both models indicate age is most predictive. CONCLUSIONS: The use of random forest models to analyze data and provide predictions for patients holds great potential in identifying modifiable patient-level and health-system factors and cohorts for increased surveillance and intervention to improve outcomes for patients. Random forests are incredibly high performing models with difficult interpretation most ideally suited for times when accurate prediction is most desirable and can be used in tandem with more common approaches to provide a more thorough analysis of observational data.

The financial evolution of endovascular aneurysm repair delivery in contemporary practice.

OBJECTIVE: The fiscal impact of endovascular repair (EVR) of aortic aneurysms and the requisite device costs have previously highlighted the tenuous long-term financial sustainability among Medicare beneficiaries. The Centers for Medicare & Medicaid Services have since reclassified EVR remuneration paradigms with new Medicare Severity Diagnosis-Related Groups (MS-DRGs) intended to better address the procedure's cost profile. The impact of this change remains unknown. The purpose of this analysis was to compare EVR-specific costs and revenue among Medicare beneficiaries both before and after this change. METHODS: All infrarenal EVRs performed in fiscal years (FYs) 2014 and 2015, before the MS-DRG change, and those performed in FYs 2017 and 2018, after the MS-DRG change, were identified using the DRG codes 238 (n = 108) and 269 (n = 84), respectively. We then identified those who were treated according to the instructions for use guidelines with a single manufacturer's device and billed to Medicare (n = 23 in FY14-15; n = 22 in FY17-18). From these cohorts, we determined total procedure technical costs, technical revenue, and net technical margin in conjunction with the hospital finance department. Results were then compared between these two groups. RESULTS: The two cohorts demonstrated similar demographic profiles (FY14-15 vs FY17-18 cohort: age, 78 years vs 74 years; median length of stay, 1.0 day vs 1.0 day). Mean total technical costs were slightly higher in the FY17-18 group ($24,511 in FY14-15 vs $26,445 in FY17-18). Graft implants continued to account for a significant portion of the total cost, with the device cost accounting for 56% of the total procedure costs in both cohorts. Net revenue was greater in the FY17-18 group by $5800 ($30,698 in FY14-15 vs $36,498 in FY17-18), resulting in an increased overall margin in the FY17-18 group compared with the FY14-15 group ($6188 in FY14-15 vs $10,053 in FY17-18). CONCLUSIONS: Device costs remain the single greatest cost driver associated with EVR delivery. DRG reclassification of EVR to address total procedure and implant costs appears to better address the requisite associated procedure costs and may thereby better support long-term fiscal sustainability of this procedure for hospitals and health systems alike.