Dataset Information

A Sneak-Peek into the Physician's Brain: A Retrospective Machine Learning-Driven Investigation of Decision-Making in TAVR versus SAVR for Young High-Risk Patients with Severe Symptomatic Aortic Stenosis.

ABSTRACT: Transcatheter aortic valve replacement (TAVR) has rapidly become a viable alternative to the conventional isolated surgical aortic valve replacement (iSAVR) for treating severe symptomatic aortic stenosis. However, data on younger patients is scarce and a gap exists between data-based recommendations and the clinical use of TAVR. In our study, we utilized a machine learning (ML) driven approach to model the complex decision-making process of Heart Teams when treating young patients with severe symptomatic aortic stenosis with either TAVR or iSAVR and to identify the relevant considerations. Out of the considered factors, the variables most prominently featured in our ML model were congestive heart failure, established risk assessment scores, previous cardiac surgeries, a reduced left ventricular ejection fraction and peripheral vascular disease. Our study demonstrates a viable application of ML-based approaches for studying and understanding complex clinical decision-making processes.

SUBMITTER: Hasimbegovic E

PROVIDER: S-EPMC8622882 | biostudies-literature | 2021 Oct

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A Sneak-Peek into the Physician's Brain: A Retrospective Machine Learning-Driven Investigation of Decision-Making in TAVR versus SAVR for Young High-Risk Patients with Severe Symptomatic Aortic Stenosis.

Hasimbegovic Ena E Papp Laszlo L Grahovac Marko M Krajnc Denis D Poschner Thomas T Hasan Waseem W Andreas Martin M Gross Christoph C Strouhal Andreas A Delle-Karth Georg G Grabenwöger Martin M Adlbrecht Christopher C Mach Markus M

Journal of personalized medicine 20211022 11

Transcatheter aortic valve replacement (TAVR) has rapidly become a viable alternative to the conventional isolated surgical aortic valve replacement (iSAVR) for treating severe symptomatic aortic stenosis. However, data on younger patients is scarce and a gap exists between data-based recommendations and the clinical use of TAVR. In our study, we utilized a machine learning (ML) driven approach to model the complex decision-making process of Heart Teams when treating young patients with severe s ...[more]

PMID: 34834414

Similar Datasets

Project description:ImportanceShared decision-making (SDM) is widely advocated for patients with valvular heart disease yet is not integrated into the heart team model for patients with symptomatic aortic stenosis. Decision aids (DAs) have been shown to improve patient-centered outcomes and may facilitate SDM.ObjectiveTo determine whether the repeated use of a DA by heart teams is associated with greater SDM, along with improved patient-centered outcomes and clinician attitudes about DAs.Design, setting, and participantsThis mixed-methods study included a nonrandomized pre-post intervention and clinician interviews. It was conducted between April 30, 2015, and December 7, 2017, with quantitative analysis performed between January 12, 2017, and May 26, 2017, within 2 academic medical centers in northern New England among 35 patients with symptomatic aortic stenosis who were at high to prohibitive risk for surgery. The qualitative analysis was performed between August 6, 2018, and May 7, 2019. The Severe Aortic Stenosis Decision Aid was delivered by 6 clinicians, with patients choosing between transcatheter aortic valve replacement and medical management.Main outcomes and measuresClinician SDM performance was measured using the Observer OPTION5 scale with dual-independent coding of audiotaped clinic visits. Previsit and postvisit surveys measured the patient's knowledge, satisfaction, and decisional conflict. Audiotaped clinician interviews were coded, and qualitative thematic analysis was performed.ResultsSix male clinicians and 35 patients (19 of 34 women [55.9%; 1 survey was missing]; mean [SD] age, 85.8 [7.8] years) participated in the study. Shared decision-making increased stepwise with repeated use of the DA (mean [SD] Observer OPTION5 scores: usual care, 17.9 [7.6]; first use of a DA, 60.5 [30.9]; fifth use of a DA, 79.0 [8.4]; P < .001 for comparison between usual care and fifth use of DA). Multiple uses of the DA were associated with increased patient knowledge (mean difference, 18.0%; 95% CI, 1.2%-34.8%; P = .04) and satisfaction (mean difference, 6.7%; 95% CI, 2.5%-10.8%; P = .01) but not decisional conflict (mean [SD]: usual care, 96.0% [9.4%]; first use of DA, 93.8% [12.5%]; fifth use of DA, 95.0% [11.2%]; P = .60). Qualitative analysis of clinicians' interviews revealed that clinicians perceived that they used an SDM approach without DAs and that the DA was not well understood by elderly patients. There was infrequent values clarification or discussion of stroke risk.Conclusion and relevanceIn a mixed-methods pilot study, use of a DA for severe aortic stenosis by heart team clinicians was associated with improved SDM and patient-centered outcomes. However, in qualitative interviews, heart team clinicians did not perceive a significant benefit of the DA, and therefore sustained implementation is unlikely. This pilot study of SDM clarifies new research directions for heart teams.

Project description:Background: Despite clear indications for intervention, therapeutic decision-making for elderly patients with severe symptomatic aortic stenosis (AS) remains a complex issue due to the wide variation in individual risk profiles and the involvement of patients' subjective preferences. We aimed to investigate the reasons leading to the decisions against intervention and the consequences thereof on survival. Methods: Data were derived from the China Elderly Valve Disease (China-DVD) Cohort Study on patients aged ≥60-year-old with severe symptomatic AS consecutively enrolled between September to December 2016. Patients were analyzed according to the initial therapeutic decisions made by consensus between patients and physicians at the time of the index evaluation: intervention group (patients who were evaluated as suitable for intervention and accepted the treatment proposal); patient-refusal group (patients who were evaluated as suitable for intervention but refused due to subjective preferences); physician-deny group (patients who were denied intervention by physicians after evaluation). The least absolute shrinkage and selection operator (LASSO)-penalized logistic regression model was used to identify the factors associated with physicians' decisions against intervention. Twelve-month survival was analyzed using Cox proportional hazards models, with multivariate adjustment using inverse probability weighting (IPW). Results: Among the enrolled 456 elderly patients with severe symptomatic AS, 52 (11.4%) patients refused intervention and 49 (10.7%) patients were denied intervention by their physicians. LASSO-penalized logistic regression model identified that reduced left ventricular ejection fraction and increased EuroSCORE-II were strongly associated with physicians' decisions against intervention. At 12-month follow-up, only 8 (15.4%) patients who initially refused the intervention proposal underwent the subsequent intervention, with an average delay of 195 days. Patients' initial decisions against intervention were significantly associated with 12-month mortality, even after IPW adjustment (Hazard ratio: 2.61; 95% confidence interval: 1.09-6.20; P = 0.031). Conclusions: The decision against intervention was taken in about one-fifth of elderly patients with symptomatic severe AS, half of which were due to patients' subjective preferences. Surgical risk remains the primary concern for physicians when making therapeutic decisions. Elderly patients' initial decisions against intervention have a profound impact on subsequent intervention rates and prognosis, and therefore should be treated as a "risk factor" at the subjective level. Clinical Trial Registration: clinicaltrials.gov/ct2/show/NCT02865798, China elDerly Valve Disease (China-DVD) cohort study (NCT02865798).

Dataset Information

A Sneak-Peek into the Physician's Brain: A Retrospective Machine Learning-Driven Investigation of Decision-Making in TAVR versus SAVR for Young High-Risk Patients with Severe Symptomatic Aortic Stenosis.

Publications

A Sneak-Peek into the Physician's Brain: A Retrospective Machine Learning-Driven Investigation of Decision-Making in TAVR versus SAVR for Young High-Risk Patients with Severe Symptomatic Aortic Stenosis.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets