Implementation of Patient-Centered Shared Decision Making for LVAD Candidates: Year 1 Results of a Multi-Site Study

Abstract

A decision aid (DA) prior to left ventricular assist device (LVAD) placement significantly increased patient knowledge and satisfaction with life after LVAD implant in a multi-site RCT (VADDA). Little evidence exists regarding best practices for implementation of a DA in real-world cardiovascular care. This project aims to evaluate DA implementation at nine U.S. hospitals with a focus on optimizing shared decision-making (SDM). Sites received an implementation plan and participated in a training webinar about ways to use the DA during LVAD evaluation to promote SDM. LVAD coordinators completed a 10-item SDM Implementation Fidelity Checklist (score 0-10) describing optimal DA use and integration of SDM as part of the DA deployment for each patient. DA "reach" is calculated by dividing the number of checklists received by the total number of patients receiving pre-LVAD education during evaluation. DA "fidelity" is calculated using LVAD coordinator self-report of the patient encounter on the Fidelity Checklist. Sites received ongoing monitoring and support from a coordinating center to discuss barriers and facilitators of DA use. Physicians and coordinators (n=30) were surveyed about their attitudes towards use of the LVAD decision aid during the first year of implementation. 454 patients received a DA from September 2018-September 2019 with an average Fidelity Checklist score of 8.5 (range 6.2-10.0). Reach ranged from 14.7%-90.2% of patients across sites with an overall reach of 53.2%. At baseline, a majority (60%, SD=1.00) of cardiologist and coordinator stakeholders believed DAs are relevant to clinical practice, and agree they improve patients' knowledge (63%, SD=0.87), satisfaction (57%, SD=0.95) and decrease anxiety (57%, SD=0.97). The most frequently cited barrier to using DAs was "not enough time to use the tool in its entirety" (50%, SD=1.34) and questioning that "the effectiveness of the aid was not well established" (40%, SD=1.42). These findings support the feasibility, high fidelity and usefulness of a LVAD DA into busy real-world cardiovascular care. The majority of clinical providers believe the aid is relevant to clinical practice.