Evaluation of machine learning methodology for the prediction of healthcare resource utilization and healthcare costs in patients with critical limb ischemia-is preventive and personalized approach on the horizon?

BACKGROUND: Critical limb ischemia (CLI) is a severe stage of peripheral arterial disease and has a substantial disease and economic burden not only to patients and families, but also to the society and healthcare systems. We aim to develop a personalized prediction model that utilizes baseline patient characteristics prior to CLI diagnosis to predict subsequent 1-year all-cause hospitalizations and total annual healthcare cost, using a novel Bayesian machine learning platform, Reverse Engineering Forward Simulation™ (REFS™), to support a paradigm shift from reactive healthcare to Predictive Preventive and Personalized Medicine (PPPM)-driven healthcare. METHODS: Patients ≥ 50 years with CLI plus clinical activity for a 6-month pre-index and a 12-month post-index period or death during the post-index period were included in this retrospective cohort of the linked Optum-Humedica databases. REFS™ built an ensemble of 256 predictive models to identify predictors of all-cause hospitalizations and total annual all-cause healthcare costs during the 12-month post-index interval. RESULTS: The mean age of 3189 eligible patients was 71.9 years. The most common CLI-related comorbidities were hypertension (79.5%), dyslipidemia (61.4%), coronary atherosclerosis and other heart disease (42.3%), and type 2 diabetes (39.2%). Post-index CLI-related healthcare utilization included inpatient services (14.6%) and ≥ 1 outpatient visits (32.1%). Median annual all-cause and CLI-related costs per patient were $30,514 and $2196, respectively. REFS™ identified diagnosis of skin and subcutaneous tissue infections, cellulitis and abscess, use of nonselective beta-blockers, other aftercare, and osteoarthritis as high confidence predictors of all-cause hospitalizations. The leading predictors for total all-cause costs included region of residence and comorbid health conditions including other diseases of kidney and ureters, blindness of vision defects, chronic ulcer of skin, and chronic ulcer of leg or foot. CONCLUSIONS: REFS™ identified baseline predictors of subsequent healthcare resource utilization and costs in CLI patients. Machine learning and model-based, data-driven medicine may complement physicians' evidence-based medical services. These findings also support the PPPM framework that a paradigm shift from post-diagnosis disease care to early management of comorbidities and targeted prevention is warranted to deliver a cost-effective medical services and desirable healthcare economy.

Proposal for a national coverage determination for the treatment of varicose veins and venous disease due to disparate Centers for Medicare and Medicaid Services local coverage determination policies.

Varicose veins and chronic venous disease are common problems in the United States. Persons with these conditions often have an adversely affected quality of life. There are a number of proven interventions to treat varicose veins and to improve patients' life quality, but these interventions are often restricted by the Centers for Medicare and Medicaid Services and private third-party payers. The Centers for Medicare and Medicaid Services have private contractors that administer Medicare policies in 10 jurisdictions across the United States. There is no national policy or coverage for the treatment of varicose veins; rather, there are multiple, disparate regional policies written by the contractors that cover the same Medicare beneficiary population. These disparate policies are not evidence based and provide unfair coverage of the same disease to the Medicare population, depending on where they live. Our proposal is for a national coverage determination policy for the treatment of varicose veins.

Program requirements for fellowship education in venous and lymphatic medicine.

Background In every field of medicine, comprehensive education should be delivered at the graduate level. Currently, no single specialty routinely provides a standardized comprehensive curriculum in venous and lymphatic disease. Method The American Board of Venous & Lymphatic Medicine formed a task force, made up of experts from the specialties of dermatology, family practice, interventional radiology, interventional cardiology, phlebology, vascular medicine, and vascular surgery, to develop a consensus document describing the program requirements for fellowship medical education in venous and lymphatic medicine. Result The Program Requirements for Fellowship Education in Venous and Lymphatic Medicine identify the knowledge and skills that physicians must master through the course of fellowship training in venous and lymphatic medicine. They also specify the requirements for venous and lymphatic training programs. The document is based on the Core Content for Training in Venous and Lymphatic Medicine and follows the ACGME format that all subspecialties in the United States use to specify the requirements for training program accreditation. The American Board of Venous & Lymphatic Medicine Board of Directors approved this document in May 2016. Conclusion The pathway to a vein practice is diverse, and there is no standardized format available for physician education and training. The Program Requirements for Fellowship Education in Venous and Lymphatic Medicine establishes educational standards for teaching programs in venous and lymphatic medicine and will facilitate graduation of physicians who have had comprehensive training in the field.

Cost-Effectiveness Analysis of Initial Treatment Strategies for Nonembolic Acute Limb Ischemia Using Real-Word Data.

BACKGROUND: Nonembolic acute limb ischemia (ALI) is a condition characterized by a sudden decrease in limb perfusion and requires immediate interventions. There are multiple treatment options available including surgery, catheter-directed thrombolysis (CDT), endovascular procedures, and hybrid treatment (a combination of open and endovascular techniques). Randomized trials provide information only on clinical efficacy, but not on economic outcomes. The objective of the study was to perform the cost-effective analysis comparing different treatment alternatives of ALI. METHODS: The data were collected from 4r ProMedica community hospitals in the Northwest Ohio from January 2009 to December 2012. Patients were included if they were treated within 14 days of onset of symptoms for nonembolic ALI and were divided into groups of receiving CDT, surgery, endovascular, or hybrid treatments. Demographics, comorbidities, medications taken before admission, and smoking status were collected at baseline for all patients and were compared among the treatment groups. A cost-effectiveness decision tree was developed to calculate expected costs and life years gained associated with available treatment options. A probabilistic sensitivity analysis was also performed to check the robustness of the model. RESULTS: A population of 205 patients with the diagnosis of ALI was included and divided into different treatment groups. There was no major significant difference in baseline characteristics among the studied groups (P > 0.05). The total costs were $17,163.47 for surgery, $20,620.39 for endovascular, $21,277.61 for hybrid, and $30,675.42 for CDT. The life years gained were 17.25 for surgery, 18 for endovascular, 18 for hybrid, and 17 for CDT. CDT was dominated because of the high cost and the low effectiveness, while hybrid treatment was dominated when compared with endovascular treatment because these 2 treatments have similar outcomes. The incremental cost-effectiveness ratio of the endovascular group over the surgery group was found to be $4,609.23 per life year gained. The sensitivity analysis showed that the endovascular treatment was found to be cost-effective under willingness to pay $50,000. CONCLUSIONS: This study provides economic evaluation of ALI treatments for a defined clinical population in the real-world setting. Compared with other available alternatives, the endovascular treatment showed to be a cost-effective use of healthcare resources.

Practice patterns of endovenous ablation therapy for the treatment of venous reflux disease.

OBJECTIVE: The purpose of this study was to assess practice patterns of endovenous ablation therapy for the treatment of venous reflux disease among the vein specialist members of the American Venous Forum (AVF). METHODS: An online survey was conducted of AVF members designed to identify demographics, treatment practices, and clinical variables in the selection of vein ablation devices. RESULTS: The survey was distributed to 798 practicing physicians, of whom 129 (16%) responded. The specialty distribution of respondents was as follows: vascular surgeons, 54%; phlebologists, 14%; general surgeons, 11%; interventional radiologists, 9%; and other specialties, 6%. The majority (81%) were from the United States, and 65% were self-employed. Almost half (47%) were in practice for >20 years, with 33% of all respondents performing three to five saphenous vein ablations per week. Three-quarters (79%) of respondents preferred radiofrequency ablation (RFA), with 47% believing that it was more cost-effective and more than half (57%) reporting improved patient satisfaction with this technique. Most of them (63%) responded that previous capital investment played a significant role in their choice of vein ablation device along with the associated cost of disposable equipment. A large majority (77%) of physicians responded that they had a significant role in choosing the treatment device, whereas only 17% thought that patients' choice played a major role in device choice. The capital investment affected choice of modality more significantly in newer practices (P < .0.5). CONCLUSIONS: The majority of AVF vein specialists prefer an RFA technique to laser, believing that RFA is associated with improved patient outcomes and is more cost-effective. Advances in technology, device costs, and reimbursement levels may have an impact on such preferences in the future.

Clinical outcomes and cost-effectiveness of initial treatment strategies for nonembolic acute limb ischemia in real-life clinical settings.

OBJECTIVE: The optimal initial treatment for patients with acute limb ischemia (ALI) remains undefined. Although clinical outcome data are inconsistent, catheter-directed thrombolysis (CDT) with tissue plasminogen activator is increasingly used. Patient-level analysis combining clinical and economic data in a real-life setting is lacking. This study compared clinical outcomes and cost-effectiveness of initial treatment strategies for nonembolic ALI using real-life patient-level data. METHODS: Medical records and data for hospital costs were analyzed for nonembolic ALI patients treated in four hospitals over 3 years. A cost-effectiveness analysis was performed using a decision tree analytic model. All costs were valued based on cost-to-charge ratios. RESULTS: In 205 patients, initial treatments were CDT alone in 68 or with angioplasty in 16, open surgery in 60, endovascular in 33, and hybrid in 28. Although clinical outcomes did not differ significantly among the groups, reintervention rates during hospital stay, readmission rates, and costs were highest in the CDT group. Reintervention was required in 62% of patients after CDT compared with 7% after open surgery, and 16% of the CDT patients needed more than one reintervention. The mean total hospital cost was $34,800 per patient in CDT group compared with $10,677 in open surgery group. CONCLUSIONS: In this real-life study, initial treatment of nonembolic ALI with currently available CDT options was associated with greater health care resource consumption and cost compared with other initial treatment options.

Multicenter assessment of the repeatability and reproducibility of the revised Venous Clinical Severity Score (rVCSS).

BACKGROUND: The Venous Clinical Severity Score (VCSS) was designed and validated as an objective measure of disease severity in patients with chronic venous disease (CVD). Recently, a revision of the VCSS (rVCSS) was performed to resolve ambiguity in the clinical descriptors and improve clarity and ease of use. This new revised VCSS requires validation to determine its repeatability and reproducibility in clinical evaluation of patients with varying levels of CVD. METHODS: A prospective multicenter protocol was designed to enroll patients undergoing evaluation for CVD at venous practices with experience using the original VCSS. At the time of initial evaluation, two clinicians independently assessed both lower extremities to determine the rVCSS and the CEAP clinical score. Between 1 and 6 weeks, patients returned and received repeat assessment of the rVCSS by the same two clinicians independently. Patients were excluded if any venous intervention occurred between the two separate evaluation visits. Scores were compared to determine inter- and intra-observer variability overall and within each CEAP clinical class. RESULTS: Seven centers enrolled a total of 136 limbs yielding 248 paired evaluations for interobserver variability and 258 paired evaluations for intraobserver variability. The mean interobserver rVCSS difference was 1.4 ± 1.7 and the mean intraobserver variability was 1.3 ± 1.6. Statistical assessment with weighted kappa yielded good repeatability (κ = 0.68; P < .0001) and good reproducibility (κ = 0.72; P < .000001) for the rVCSS. The rVCSS correlated well with the CEAP clinical class with significant differences between rVCSS in increasing classes. (P < .0001). CONCLUSIONS: In this multicenter evaluation, the rVCSS was demonstrated to be a reliable and reproducible instrument for documentation of the severity of symptoms in patients with lower extremity venous insufficiency.

Multicenter assessment of venous reflux by duplex ultrasound.

OBJECTIVE: This prospective multicenter investigation was conducted to define the repeatability of duplex-based identification of venous reflux and the relative effect of key parameters on the reproducibility of the test. METHODS: Repeatability was studied by having the same technologist perform duplicate tests, at the same time of the day, using the same reflux-provoking maneuver and with the patient in the same position. Reproducibility was examined by having two different technologists perform the test at the same time of the day, using the same reflux-provoking maneuver and with the patient in the same position. Facilitated reproducibility was studied by having two different technologists examine the same patients immediately after an educational intervention. Limits of agreement between two duplex scans were studied by changing three elements of the test: time of the day (morning vs afternoon), patient's position (standing vs supine), and reflux initiation (manual vs automatic compression-decompression). RESULTS: The study enrolled 17 healthy volunteers and 57 patients with primary chronic venous disease. Repeatability of reflux time measurements in deep veins did not significantly differ with the time of day, the patient's position, or the reflux-provoking maneuver. Reflux measurements in the superficial veins were more repeatable (P < .05) when performed in the morning with the patient standing. The agreement between the clinical interpretations significantly depended on a selected cut point (Spearman's ρ, -0.4; P < .01). Interpretations agreed in 93.4% of the replicated measurements when a 0.5-second cut point was selected. The training intervention improved the frequency of agreement to 94.4% (κ = 0.9). Alternations of the time of the duplex scan, the patient's position, and the reflux-provoking maneuver significantly decreased reliability. CONCLUSIONS: This study provides evidence to develop a new standard for duplex ultrasound detection of venous reflux. Reports should include information on the time of the test, the patient's position, and the provoking maneuver used. Adopting a uniform cut point of 0.5 second for pathologic reflux can significantly improve the reliability of reflux detection. Implementation of a standard protocol should elevate the minimal standard for agreement between repeated tests from the current 70% to at least 80% and with more rigid standardization, to 90%.

Reproducibility of ultrasound scan in the assessment of volume flow in the veins of the lower extremities.

PURPOSE: The purpose of this study was the identification of the optimal settings of ultrasound scan flow measurement in the veins and the determination of whether the standardization of these settings can provide acceptable reproducibility of the venous flow measurements in individual segments of the lower extremity veins. METHODS: The venous cross-sectional area, the time average mean velocity, and the venous volume flow of 25 healthy volunteers were examined with duplex ultrasound scanning. Reproducibility was examined with different measurement settings. Doppler scan sample volume size, ultrasound scan beam incident angle, and time interval of measurement were varied across a spectrum for arrival at the setting for highest reproducibility of the flow volume measurements. Test-retest reproducibility of venous flow volume measurements then was investigated with optimized settings. RESULTS: The highest repeatability of volume flow measurements was achieved when the full lumen of the vein was insonated (coefficient of repeatability [CR] = 1.88 cm/s), the ultrasound scan beam incident angle was equal to 60 degrees (CR = 1.56 cm/s), and the measurement time was more than 40 seconds (CR = 1.64 cm/s). The mean values of volume flow were 360 mL/min in the common femoral vein, 147 mL/min in the superficial femoral vein, 86 mL/min in the profunda femoral vein, and 38 mL/min in the greater saphenous vein. Test-retest repeatability coefficients were 96.9 mL/min for the common femoral vein, 70.2 mL/min for the superficial femoral vein, 40.8 mL/min for the profunda femoral vein, and 16.8 mL/min for the greater saphenous vein. CONCLUSION: The reproducibility of ultrasound scan measurements of volume flow in veins is optimized with the use of sampling volumes that cover the entire venous lumen, with an incident angle of 60 degrees and measuring for 40-second intervals or longer. With these defined variables, volumetric measurements are sufficiently repeatable. the values of flow volume measured with duplex ultrasound scanning were comparable to those with thermodilution techniques that were reported previously.