A multi-site trial to test benefits of adding a personalized risk calculator to an online decision aid for left ventricular assist device

PROJECT SUMMARY
Continuous flow left ventricular assist devices (LVADs) now rival heart transplant as the gold standard for
advanced heart failure. However, uncertainty persists over which patients are most likely to benefit from
treatment. New technologies for a more personalized approach to risk prediction and supported decision-
making are crucially needed to improve current clinical risk calculators and decision aids which lack the ability
to (a) calculate patients’ individualized and highly variable personalized risks, (b) communicate those risks to
patients in the context of their values and goals and discuss these factors with their health care team, and (c)
capitalize upon digitalized health systems and platforms to provide most expeditious and efficient updates to
rapidly-changing risk. This represents a significant gap in informed and high-quality decision making that also
potentially negatively affects health and clinical outcomes. The objective of this proposal is to improve patients’
informed and shared decision making about LVAD by providing a holistic and personalized clinical decision
support system for LVAD candidates that interactively guides their understanding of how available treatment
options align with their expressed values, given their personalized risk estimates, and provides them with a tool
to communicate these values to their clinical team. We will do this by updating and integrating a validated
online risk prediction and communication tool, the Cardiac Outcomes Risk Assessment (CORA)
developed by our colleagues at Cornell University, with our efficacy-tested decision aid (Deciding
Together) for LVAD. To accomplish this, in Aim 1 we will conduct in-depth qualitative interviews to identify
patients’ and physicians’ practical, ethical, and contextual considerations while using CORA, while our
colleagues at Cornell simultaneously update and (re-)validate the updated model against existing LVAD risk
prediction models, as well as integrate CORA with an adapted and digitalized version of our decision aid. In Aim
2, we will then translate patients' and physicians’ practical, ethical and contextual considerations into
concrete improvements to CORA and conduct acceptability/feasibility testing with patients and physicians
using the updated, validated system. These steps will culminate in Aim 3, a multi-site randomized controlled trial
to evaluate the impact of a personalized approach to clinical decision making on informed and values-
concordant choice and shared decision-making outcomes. This contribution will be significant because it
will address the urgent need to better identify and respond to the specific and dynamic nature of patient needs
in seeking advanced heart failure treatment. Our approach is innovative in that it harnesses state of the art
technology prediction models and LVAD decision support, synthesizing diverse expertise from a team of
bioengineers, computer scientists, leading heart failure cardiologists, decision scientists, and medical ethicists.
This five-year project is feasible in that it builds on 6 years of research on the development, implementation and
dissemination of LVAD decision support and a decade of research into accurate risk prediction models for LVAD.

Grant Number: 5R01HS027784-05
NIH Institute/Center: AHRQ
Principal Investigator: Jennifer Blumenthal-Barby