Machine learning models predict total charges and drivers of cost for transcatheter aortic valve replacement.

Background: Given the increasing healthcare costs, there is an interest in developing machine learning (ML) prediction models for estimating hospitalization charges. We use ML algorithms to predict hospitalization charges for patients undergoing transfemoral transcatheter aortic valve replacement (TF-TAVR) utilizing the National Inpatient Sample (NIS) database. Methods: Patients who underwent TF-TAVR from 2012 to 2016 were included in the study. The primary outcome was total hospitalization charges. Study dataset was divided into 80% training and 20% testing sets. We used following ML regression algorithms: random forest, gradient boosting, k-nearest neighbors (KNN), multi-layer perceptron and linear regression. ML algorithms were built for for 3 stages: Stage 1, including variables that were known pre-procedurally (prior to TF-TAVR); Stage 2, including variables that were known post-procedurally; Stage 3, including length of stay (LOS) in addition to the stage 2 variables. Results: A total of 18,793 hospitalization for TF-TAVR were analyzed. The mean and median adjusted hospitalization charges were $220,725.2 ($137,675.1) and $187,212.0 ($137,971.0-264,824.8) respectively. Random forest regression algorithm outperformed other ML algorithms at all stages with higher R2 score and lower mean absolute error (MAE), root mean squared area (RMSE) and root mean squared logarithmic error (RMSLE) (Stage 1: MAE 79,979.11, R2 0.157; Stage 2: MAE 76,200.09, R2 0.256; Stage 3: MAE 69,350.09, R2 0.453). LOS was the most important predictor of hospitalization charges. Conclusions: We built ML algorithms that predict hospitalization charges with good accuracy in patients undergoing TF-TAVR at different stages of hospitalization and that can be used by healthcare providers to better understand the drivers of charges.

Five years of a comprehensive ST-elevation myocardial infarction protocol and its association with sex disparities.

Aims: To determine whether a comprehensive ST-elevation myocardial infarction (STEMI) protocol is associated with reduced sex disparities over 5 years. Methods and results: This was an observational cohort study of 1833 consecutive STEMI patients treated with percutaneous coronary intervention (PCI) before (1 January 2011-14 July 2014, control group) and after (15 July 2014-15 July 2019, protocol group) implementation of a protocol for early guideline-directed medical therapy (GDMT), rapid door to balloon time (D2BT), and use of trans-radial PCI. In the control group, females had less GDMT (77.1% vs. 68.1%, P = 0.03), similarly low trans-radial PCI (19.0% vs. 17.6%, P = 0.73), and longer D2BT [104 min (79, 133) vs. 112 min (85, 147), P = 0.02] corresponding to higher in-hospital mortality [4.5% vs. 10.3%, odds ratio (OR) 2.44 (1.34-4.46), P = 0.004], major adverse cardiac and cerebrovascular events [MACCE, 9.8% vs. 16.3%, OR 1.79 (1.14-2.84), P = 0.01], and net adverse clinical events [NACE, 16.1% vs. 28.3%, OR 2.06 (1.42-2.99), P < 0.001]. In the protocol group, no significant sex differences were observed in GDMT (87.2% vs. 86.4%, P = 0.81) or D2BT [85 min (64-106) vs. 89 min (65-111), P = 0.06], but trans-radial PCI was used less in females (77.6% vs. 71.2%, P = 0.03). In-hospital mortality [2.5% vs. 4.4%, OR 1.78 (0.91-3.51), P = 0.09] and MACCE [9.0% vs. 11.1%, OR 1.27 (0.83-1.92), P = 0.26] were similar between sexes, but higher NACE in females approached significance [14.8% vs. 19.4%, OR 1.38 (0.99-1.92), P = 0.05] due to higher bleeding risk [7.2% vs. 11.1%, OR 1.60 (1.04-2.46), P = 0.03]. Conclusions: A comprehensive STEMI protocol was associated with sustained reductions for in-hospital ischaemic outcomes over 5 years, but higher bleeding rates in females persisted.

Trends in Outcomes of Transcatheter and Surgical Aortic Valve Replacement in the United States (2012-2017).

As the use of transcatheter aortic valve implantation (TAVI) expands to varying patient populations, impacting the landscape of surgical aortic valve replacement (SAVR), this study sought to assess volume and performance trends of aortic valve replacement (AVR) in the United States during 2012-2017. The Nationwide Readmissions Database was queried for patients who underwent endovascular/transapical TAVI, isolated SAVR, or complex aortic valve surgery between 2012 and 2017. Temporal trends in annual case volume, admission costs, in-hospital outcomes, and 30-day readmission were evaluated. Of 624,303 patients (median age 72 years) who received AVR, 387,011 (62%) were men. Among these patients, 170,521 (27%) underwent TAVI and 453,782 (73%) underwent SAVR with 299,398 isolated and 154,384 complex aortic valve surgery. TAVI patients were significantly older and higher risk compared with SAVR patients. From 2012 to 2017, the annual number of TAVI increased from 8,295 to 55,168 whereas SAVR volume remained remarkably stable. Patients who underwent AVR demonstrated significant improvements in mortality, stroke, duration of hospitalization, and 30-day readmission. In conclusion, this large contemporary analysis reports the considerable growth of AVR in the United States. It remains unequivocal that the treatment of aortic stenosis is improving overall with reduced mortality following AVR, highlighting the effectiveness of various process improvements such as newer valves, enhanced patient selection, and the interdisciplinary Heart Team approach.

Impact of Economic Status on Utilization and Outcomes of Transcatheter Aortic Valve Implantation and Mitraclip.

Data on the impact of economic status on Transcatheter aortic valve implantation (TAVI) and MitraClip (MC) is lacking. Patients who underwent TAVI and/or MC during 2012 to 2017 were identified in the Nationwide Readmission Database and divided by zip code estimated income quartile into 4 groups (Q1 to Q4). The utilization of TAVI and/or MC was defined as the number of TAVIs and/or MCs over all admissions with an aortic and/or mitral valve disease (AVD and/or MVD) and represented per 1,000 admissions. A total of 168,853 patients underwent TAVI; 20.6% in Q1, 26.3% in Q2, 27.3% in Q3, and 25.8% in Q4, while 15,387 patients underwent MC; 22% in Q1, 26.2% in Q2, 26.3% in Q3, and 25.5% in Q4. The annual utilization of TAVIs and/or MCs increased over the study period and was generally lower with lower income. In 2012, TAVI was performed for 8.2, 8.8, 10.8, and 11.3 per 1,000 AVD admissions in Q1, Q2, Q3, and Q4, respectively. In 2017, TAVI was performed for 54.1, 65.1, 68.6, and 71 per 1,000 AVD admissions in Q1, Q2, Q3, and Q4, respectively. In 2014, MC was performed for 1.6, 2.1, 1.8, and 1.9 per 1,000 MVD admissions in Q1, Q2, Q3, and Q4, respectively. In 2017, MC was performed for 5.6, 6.5, 8, and 8 per 1,000 MVD admissions in Q1, Q2, Q3, and Q4, respectively. In-hospital mortality, stroke, and 30-day readmissions were generally comparable across groups. Lower-income patients may be underrepresented among patients undergoing TAVI and MC despite comparable outcomes. Further studies are needed to examine the etiologies behind these disparities and identify targeted strategies for its mitigation.