Evaluation of Image-Based Modelling on Clinical Decisions in Coarctation of the Aorta

Overview

To answer the research question: "Would image-based modelling result in different clinical decisions as compared to clinical practice guidelines?", we will conduct a randomized controlled experiment in which we will compare the hypothetical decisions made by interventional cardiologists who are presented with imaging parameters currently recommended by clinical practice guidelines vs. hypothetical decisions made by interventional cardiologists receiving an expanded list of parameters, including simulation modelling.

In collaboration with our three clinical partners, we will first generate two separate imaging datasets for a maximum of three patients recruited to participate in CARDIOPROOF. The first dataset will include the imaging parameters currently recommended by clinical practice guidelines (referred to as "limited dataset"). The second dataset will include an expanded list of parameters, inclusive of information that is available from traditional imaging parameters (as recommended by the guidelines) and simulation modeling (referred to as "image-based modelling dataset"). We will generate both limited and image-based modelling datasets from fully de-identified patients already enrolled in CARDIOPROOF (NCT02591940) who have consented to publication of data in anonymized form. Using a computerized random-sample function, we will randomly allocate interventional cardiologists into two separate groups and present them with one set of imaging data. The first group will receive a "limited" dataset including only information that is available from traditional diagnostics (as recommended by the clinical practice guidelines) for a pre-specified number of patients (maximum of 3). The second group will receive the full, detailed dataset inclusive of information that is available from traditional diagnostics (as recommended by the guidelines) and simulation modelling for the same set of patients. We will then ask the interventional cardiologists in the two groups to make (hypothetical) clinical decisions using the dataset of imaging parameters presented to them. The clinical decisions will be hypothetical because patients will have been treated according to clinical practice guidelines and this experiment will retrospectively involve interventional cardiologists who are not directly involved in the care of the patients participating in CARDIOPROOF. The analysis will focus on each hypothetical scenario and compare the proportions of cardiologists making different types of intervention decisions in the two randomly allocated groups.