TAVR and cancer: machine learning-augmented propensity score mortality and cost analysis in over 30 million patients

Monitoring adverse drug events or pharmacovigilance has been promoted by the World Health Organization to assure the safety of medicines through a timely and reliable information exchange regarding drug safety issues. We aim to discuss the application of machine learning methods as well as causal inference paradigms in pharmacovigilance. We first reviewed data sources for pharmacovigilance. Then, we examined traditional causal inference paradigms, their applications in pharmacovigilance, and how machine learning methods and causal inference paradigms were integrated to enhance the performance of traditional causal inference paradigms. Finally, we summarized issues with currently mainstream correlation-based machine learning models and how the machine learning community has tried to address these issues by incorporating causal inference paradigms. Our literature search revealed that most existing data sources and tasks for pharmacovigilance were not designed for causal inference. Additionally, pharmacovigilance was lagging in adopting machine learning-causal inference integrated models. We highlight several currently trending directions or gaps to integrate causal inference with machine learning in pharmacovigilance research. Finally, our literature search revealed that the adoption of causal paradigms can mitigate known issues with machine learning models. We foresee that the pharmacovigilance domain can benefit from the progress in the machine learning field.

Interventional cardiology in cancer patients: A position paper from the Portuguese Cardiovascular Intervention Association and the Portuguese Cardio-Oncology Study Group of the Portuguese Society of Cardiology

Background and objectives: Little is known about outcome improvements and disparities in cardiac arrest and active cancer. We performed the first known AI and propensity score (PS)-augmented clinical, cost-effectiveness, and computational ethical analysis of cardio-oncology cardiac arrests including left heart catheterization (LHC)-related mortality reduction and related disparities. Materials and methods: A nationally representative cohort analysis was performed for mortality and cost by active cancer using the largest United States all-payer inpatient dataset, the National Inpatient Sample, from 2016 to 2018, using deep learning and machine learning augmented propensity score-adjusted (ML-PS) multivariable regression which informed cost-effectiveness and ethical analyses. The Cardiac Arrest Cardio-Oncology Score (CACOS) was then created for the above population and validated. The results informed the computational ethical analysis to determine ethical and related policy recommendations. Results: Of the 101,521,656 hospitalizations, 6,656,883 (6.56%) suffered cardiac arrest of whom 61,300 (0.92%) had active cancer. Patients with versus without active cancer were significantly less likely to receive an inpatient LHC (7.42% versus 20.79%, p < 0.001). In ML-PS regression in active cancer, post-arrest LHC significantly reduced mortality (OR 0.18, 95%CI 0.14–0.24, p < 0.001) which PS matching confirmed by up to 42.87% (95%CI 35.56–50.18, p < 0.001). The CACOS model included the predictors of no inpatient LHC, PEA initial rhythm, metastatic malignancy, and high-risk malignancy (leukemia, pancreas, liver, biliary, and lung). Cost-benefit analysis indicated 292 racial minorities and $2.16 billion could be saved annually by reducing racial disparities in LHC. Ethical analysis indicated the convergent consensus across diverse belief systems that such disparities should be eliminated to optimize just and equitable outcomes. Conclusions: This AI-guided empirical and ethical analysis provides a novel demonstration of LHC mortality reductions in cardio-oncology cardiac arrest and related disparities, along with an innovative predictive model that can be integrated within the digital ecosystem of modern healthcare systems to improve equitable clinical and public health outcomes.

Carcinoid heart disease is increasingly recognized and challenging to manage due to limited outcomes data. This is the largest known cohort study of valvular pathology, treatment (including pulmonary and tricuspid valve replacements [PVR and TVR]), dispairties, mortality, and cost in patients with malignant carcinoid tumor (MCT). Machine learning-augmented propensity score-adjusted multivariable regression was conducted for clincal outcomes in the 2016-2018 U.S. National Inpatient Sample (NIS). Regression models were weighted by the complex survey design and adjusted for known confounders and the likelihood of undergoing valvular procedures. Among 101,521,656 hospitalizations, 55,910 (0.06%) had MCT. Patients with MCT vs. those without had significantly higher inpatient mortality (2.93 vs. 2.04%, p = 0.002), longer mean length of stay (12.20 vs. 4.62, p < 0.001), and increased mean total cost of stay ($70,252.18 vs. 51,092.01, p < 0.001). There was a step-wise increased rate of TVR and PVR with each subsequent year, with significantly more TV (0.16% vs. 0.01, p < 0.001) and PV (0.03 vs. 0.00, p = 0.040) diagnosed with vs. without MCT for 2016, with comparable trends in 2017 and 2018. There were no significant procedural disparities among patients with MCT for sex, race, income, urban density, or geographic region, except in 2017, when the highest prevalence of PV procedures were performed in the Western North at 50.00% (p = 0.034). In machine learning and propensity score augmented multivariable regression, MCT did not significantly increase the likelihood of TVR or PVR. In sub-group analysis restricted to MCT, neither TVR nor PVR significantly increased mortality, though it did increase cost (respectively, $141,082.30, p = 0.015; $355,356.40, p = 0.012). This analysis reflects a favorable trend in recognizing the need for TVR and PVR in patients with MCT, with associated increased cost but not mortality. Our study also suggests that pulmonic valve pathology is increasingly recognized in MCT as reflected by the upward trend in PVRs. Further research and updated societal guidelines may need to focus on the "forgotten pulmonic valve" to improve outcomes and disparities in this understudied patient population.

Background and Objectives: There are no nationally representative studies of mortality and cost effectiveness for fractional flow reserve (FFR) guided percutaneous coronary interventions (PCI) in patients with cancer. Our study aims to show how this patient population may benefit from FFR-guided PCI. Materials and Methods: Propensity score matched analysis and backward propagation neural network machine learning supported multivariable regression was performed for inpatient mortality in this case-control study of the 2016 National Inpatient Sample (NIS). Regression results were adjusted for age, race, income, geographic region, metastases, mortality risk, and the likelihood of undergoing FFR versus non-FFR PCI. All analyses were adjusted for the complex survey design to produce nationally representative estimates. Results: Of the 30,195,722 hospitalized patients meeting criteria, 3.37% of the PCIs performed included FFR. In propensity score adjusted multivariable regression, FFR versus non-FFR PCI significantly reduced inpatient mortality (OR 0.47, 95%CI 0.35–0.63; p < 0.001) and length of stay (LOS) (in days; beta −0.23, 95%CI −0.37–−0.09; p = 0.001) while increasing cost (in USD; beta $5708.63, 95%CI, 3042.70–8374.57; p < 0.001), without significantly increasing complications overall. FFR versus non-FFR PCI did not specifically change cancer patients’ inpatient mortality, LOS, or cost. However, FFR versus non-FFR PCI significantly increased inpatient mortality for Hodgkin’s lymphoma (OR 52.48, 95%CI 7.16–384.53; p < 0.001) and rectal cancer (OR 24.38, 95%CI 2.24–265.73; p = 0.009). Conclusions: FFR-guided PCI may be safely utilized in patients with cancer as it does not significantly increase inpatient mortality, complications, and LOS. These findings support the need for an increased utilization of FFR-guided PCI and further studies to evaluate its long-term impact.

2021. The year in review. Structural heart interventions

Post-cardiac arrest PCI is underutilized among cancer patients: Machine learning augmented nationally representative case-control study of 30 million hospitalizations

This review provides an in-depth analysis of the effect of length of stay (LOS), comorbidities, and procedural complications on the cost-effectiveness of transcatheter aortic valve replacement (TAVR) in comparison to surgical aortic valve replacement (SAVR). We found that the average LOS was shorter for patients undergoing TAVR, contributing to lower average costs associated with the procedure, although the LOS varied between patients due to the severity of illness and comorbidities present. TAVR has also been found to improve the quality of life for patients receiving aortic valve replacement compared to SAVR. Although TAVR has a lower rate of most post-operative complications caused by SAVR, such as bleeding and cardiac complications, TAVR shows an increased rate of permanent pacemaker (PPM) implantation due to mechanical trauma on the heart's conduction system. In addition, our findings suggest that the cost-effectiveness of each procedure varies based on the types of valve, the patient history of other medical conditions, and the procedural methods. Our findings show that TAVR is preferred over SAVR in terms of cost-effectiveness across a variety of patients with other coexisting medical conditions, including cancer, advanced kidney disease, cirrhosis, diabetes mellitus, and bundle branch block. TAVR also appears to be superior to SAVR with fewer post-operative complications. However, TAVR appears to have a higher rate of PPM implantation rates as compared to SAVR. The comorbidities of the valve recipient must be considered when deciding whether to use TAVR or SAVR as cost-effectiveness varies with the patient background.