Validation Of A Computable Algorithm For Medication Optimization In Heart Failure With Reduced Ejection Fraction

Abstract

Background Guideline-directed medical therapy (GDMT) optimization can improve outcomes in heart failure with reduced ejection fraction (HFrEF) but often does not occur in clinical practice. The objective of this study was to assess the prescribing of GDMT in clinical trial data using a novel algorithm to identify opportunities for medication optimization. Methods Clinical trial data from GUIDE-IT (Guiding Evidence-Based Therapy Using Biomarker Intensified Treatment in Heart Failure), a study using natriuretic peptides to guide medical therapy in HFrEF, was passed through a computable medication optimization algorithm to output GDMT recommendations. Algorithm-based recommendations were compared to actual medication changes at GUIDE-IT visits. Results Data were analyzed from 5733 GUIDE-IT visits in 841 unique patients. In patients not on therapy, the algorithm recommended initiation of angiotensin-converting enzyme inhibitor/angiotensin receptor blocker (ACEi/ARB), beta-blockers, and mineralocorticoid receptor antagonists (MRA) in 480 (52.8%), 51 (34.9%), and 1421 (68.1%) visits, respectively. Initiation of the agents only occurred in 100 (20.8%), 29 (56.9%), and 224 (15.8%) of those visits. The algorithm also recommended increasing doses in an additional 1475 (48.8%) visits for ACEi/ARBs and 1158 (39.4%) visits for beta-blockers. The dose increases only occurred in 358 (24.3%) of these visits for ACEi/ARBs and 426 (36.8%) visits for beta-blockers leaving many opportunities for medication optimization (Figure 1). The medication optimization score (MOS) increased over time (Figure 2). As the MOS increased over time, the hazard ratio for time to first CV death or HF hospitalization decreased. Conclusions The algorithm accurately identified patients that met the criteria for GDMT initiation and titration. Even in a clinical trial with robust protocols, GDMT could have been further optimized in a meaningful number of visits. The MOS generated from the algorithm may correlate with clinical outcomes. Implementation of algorithms into electronic health records could help identify suboptimal GDMT and improve HFrEF care. Guideline-directed medical therapy (GDMT) optimization can improve outcomes in heart failure with reduced ejection fraction (HFrEF) but often does not occur in clinical practice. The objective of this study was to assess the prescribing of GDMT in clinical trial data using a novel algorithm to identify opportunities for medication optimization. Clinical trial data from GUIDE-IT (Guiding Evidence-Based Therapy Using Biomarker Intensified Treatment in Heart Failure), a study using natriuretic peptides to guide medical therapy in HFrEF, was passed through a computable medication optimization algorithm to output GDMT recommendations. Algorithm-based recommendations were compared to actual medication changes at GUIDE-IT visits. Data were analyzed from 5733 GUIDE-IT visits in 841 unique patients. In patients not on therapy, the algorithm recommended initiation of angiotensin-converting enzyme inhibitor/angiotensin receptor blocker (ACEi/ARB), beta-blockers, and mineralocorticoid receptor antagonists (MRA) in 480 (52.8%), 51 (34.9%), and 1421 (68.1%) visits, respectively. Initiation of the agents only occurred in 100 (20.8%), 29 (56.9%), and 224 (15.8%) of those visits. The algorithm also recommended increasing doses in an additional 1475 (48.8%) visits for ACEi/ARBs and 1158 (39.4%) visits for beta-blockers. The dose increases only occurred in 358 (24.3%) of these visits for ACEi/ARBs and 426 (36.8%) visits for beta-blockers leaving many opportunities for medication optimization (Figure 1). The medication optimization score (MOS) increased over time (Figure 2). As the MOS increased over time, the hazard ratio for time to first CV death or HF hospitalization decreased. The algorithm accurately identified patients that met the criteria for GDMT initiation and titration. Even in a clinical trial with robust protocols, GDMT could have been further optimized in a meaningful number of visits. The MOS generated from the algorithm may correlate with clinical outcomes. Implementation of algorithms into electronic health records could help identify suboptimal GDMT and improve HFrEF care.