Early Prediction of Major Adverse Cardiovascular Events Using Remote Monitoring

Accurate assessment of cardiovascular risk is essential for clinical decision making in that the benefits, risks, and costs of alternative strategies must be weighed ahead of choosing the best treatment for individuals. Existing multivariable risk prediction models are vital components of current practice, and remain the logical standard to which new risk markers must be added and compared.7 The study described herein applies a practical framework for assessing the value of novel risk markers identified through patient reported outcomes (PROs), patient reported informatics (PRIs),8 and biomarkers in the forms of proteins and lipids. Though the purpose of the study is largely exploratory, it does take preliminary steps toward answering the question: "Do new PRO-, PRI-, and/or bio-markers add significant predictive information beyond that provided by established cardiac risk factors?" STUDY AIMS Aim 1: To measure cross-sectional correlations between PRIs, PROs, MACE biomarker candidates, and established MACE biomarker surrogates known to closely predict MACE itself (e.g. ultra-high sensitive troponin I \[u-hsTnI\], brain natriuretic peptide \[BNP\], and high sensitivity C-reactive protein \[hsCRP\], assay 1). Hypothesis 1: PRI metrics, PRO measure scores, and Candidate Biomarkers will correlate with MACE biomarker surrogates. Justification: Usual care may not identify subtle clinical changes that precede MACE. In order to justify future efforts to employ remote monitoring at scale to predict MACE, we will first evaluate for evidence of basic, cross-sectional correlations between PRIs, PROs, and known MACE surrogate biomarkers. Aim 2: To measure the longitudinal relationship between PRI metrics, PRO measure scores, Candidate Biomarkers, and changes in MACE surrogates. Hypothesis: Changes in PRI metrics, PRO measure scores, and candidate biomarkers will predict changes in MACE biomarker surrogates. Justification: If changes in PRI metrics, PRO measure scores, and candidate biomarkers can predict longitudinal changes in MACE biomarker surrogates, then it will provide biological plausibility that remote surveillance may predict MACE itself; this would justify a larger trial of remote digital monitoring vs. usual care and suggest the concept has merit. Exploratory Aim 2b: To assess improvement in risk prediction provided by risk markers identified in the above aims. Hypothesis: Using PRI-, PRO-, and Bio- marker predictors in combination with established risk factors will provide incremental prognostic information compared to models using established risk factors alone. Additionally, we will perform in-depth proteomic and bioinformatics analysis using baseline samples to explore potential molecular mechanisms driving MACE. Specific Aim 3: To estimate the cost-effectiveness and budget impact of remote monitoring for MACE. Hypothesis: The incremental cost of remote monitoring will be offset by downstream savings engendered by early and precise prediction of unexpected and costly MACE in stable moderate-risk IHD. Justification: Precision Medicine innovations must be cost-effective in order to be scaled across health systems and receive payer support. Using summary results from this study, we will create hypothesis-generating cost-effectiveness, cost-utility, and budget impact models to estimate the projected return on investment of remote monitoring. Importantly, these models are evaluative in nature and do not involve patient-level data - let alone identifiable information - of any sort.