Barriers to Optimization of Medical Therapy and the Role of Checklist‐Based Decision Support in Heart Failure Authors: Brett M. Montelaro, Edward Woods, Candace D. Speight, Eisha P. Udeshi, Andrea R. Mitchell, Michael A. Burke, Jamie Diamond, Alanna A. Morris, Engels N. Obi, Andra S. Stevenson, Molly Tiedeken, Neal W. Dickert Published: 2026 ABSTRACT BackgroundGuideline‐directed medical therapy (GDMT) reduces morbidity and mortality in heart failure with reduced ejection fraction, yet it remains underused. Patient‐activation interventions, including checklist‐based decision support, have shown improvement in GDMT optimization, and iterative refinement of such tools may enhance their impact. We analyzed recorded clinician–patient encounters, in which a checklist‐based activation tool was used, to identify barriers to GDMT optimization and opportunities to enhance decision support.MethodsThis was a secondary analysis of transcript data from the POCKET‐COST‐HF (Integrating Cost Into Shared Decision‐Making for Heart Failure With Reduced Ejection Fraction) trial, which implemented a checklist‐based tool focused on prescription price transparency. Patients and clinicians at 2 academic health centers received adapted versions of the EPIC‐HF (Electronically Delivered Patient‐Activation Tool for Intensification of Medications for Chronic Heart Failure With Reduced Ejection Fraction) checklist outlining approved medications and target doses for HFrEF. Encounters were audio recorded and transcribed and underwent content analysis to identify barriers and opportunities for checklist improvement.ResultsEncounters between 247 patients with heart failure with reduced ejection fraction (mean age 62.9, 29.5% female, 26.3% Black) and 39 clinicians were included. Baseline GDMT use was high (95% beta blockers, 81% angiotensin‐converting enzyme inhibitors/angiotensin receptor blockers/angiotensin receptor‐neprilysin inhibitors, 63% mineralocorticoid receptor antagonists, and 43% SGLT2 [sodium‐glucose cotransporter 2] inhibitors). The checklist was referenced in 48.6% of encounters, and GDMT optimization discussed in 75.7%. Barriers to optimization were identified in 61.5% of encounters, the most common (74.3%) being medical (eg, hypotension). Nonmedical barriers included medication cost (29.6%) and resistance from patients or clinicians (13.8%), often reflecting clinical inertia or desire to minimize medications.ConclusionsNonmedical barriers to GDMT optimization may be addressed by decision‐support tools such as checklists. These data suggest that attention to cost, clinical inertia, and desire to minimize medications should be prioritized in future iterations. Nonstandard Abbreviations and Acronyms guideline‐directed medical therapy heart failure with reduced ejection fraction Research PerspectiveWhat Is New? The impact of decision support tools on patient–clinician conversations regarding guideline‐ directed medical therapy optimization has not been assessed and may help refine such tools for large‐scale implementation.A significant number of outpatient encounters do not address medication optimization, suggesting continued value in refining decision‐support and patient‐activation tools. Medical comorbidities remain the most common barrier to medication optimization; however, clinical inertia and cost remain significant barriers to guideline‐ directed medical therapy optimization and could be addressed using decision support tools. What Question Should Be Addressed Next? Refining decision support tools to more clearly delineate therapeutic goals, highlight efficacy of treatment, and approximate financial impact may improve their uptake and efficacy as well as improve outpatient guideline‐ directed medical therapy optimization. Future studies should assess the impact of these refined decision support tools and track optimization rates, and studies should continue to address optimal approaches to implementation of decision support tools for both patients and physicians. Research Perspective What Is New? The impact of decision support tools on patient–clinician conversations regarding guideline‐ directed medical therapy optimization has not been assessed and may help refine such tools for large‐scale implementation.A significant number of outpatient encounters do not address medication optimization, suggesting continued value in refining decision‐support and patient‐activation tools. Medical comorbidities remain the most common barrier to medication optimization; however, clinical inertia and cost remain significant barriers to guideline‐ directed medical therapy optimization and could be addressed using decision support tools. The impact of decision support tools on patient–clinician conversations regarding guideline‐ directed medical therapy optimization has not been assessed and may help refine such tools for large‐scale implementation. The impact of decision support tools on patient–clinician conversations regarding guideline‐ directed medical therapy optimization has not been assessed and may help refine such tools for large‐scale implementation. A significant number of outpatient encounters do not address medication optimization, suggesting continued value in refining decision‐support and patient‐activation tools. Medical comorbidities remain the most common barrier to medication optimization; however, clinical inertia and cost remain significant barriers to guideline‐ directed medical therapy optimization and could be addressed using decision support tools. A significant number of outpatient encounters do not address medication optimization, suggesting continued value in refining decision‐support and patient‐activation tools. Medical comorbidities remain the most common barrier to medication optimization; however, clinical inertia and cost remain significant barriers to guideline‐ directed medical therapy optimization and could be addressed using decision support tools. What Question Should Be Addressed Next? Refining decision support tools to more clearly delineate therapeutic goals, highlight efficacy of treatment, and approximate financial impact may improve their uptake and efficacy as well as improve outpatient guideline‐ directed medical therapy optimization. Future studies should assess the impact of these refined decision support tools and track optimization rates, and studies should continue to address optimal approaches to implementation of decision support tools for both patients and physicians. Refining decision support tools to more clearly delineate therapeutic goals, highlight efficacy of treatment, and approximate financial impact may improve their uptake and efficacy as well as improve outpatient guideline‐ directed medical therapy optimization. Future studies should assess the impact of these refined decision support tools and track optimization rates, and studies should continue to address optimal approaches to implementation of decision support tools for both patients and physicians. Refining decision support tools to more clearly delineate therapeutic goals, highlight efficacy of treatment, and approximate financial impact may improve their uptake and efficacy as well as improve outpatient guideline‐ directed medical therapy optimization. Future studies should assess the impact of these refined decision support tools and track optimization rates, and studies should continue to address optimal approaches to implementation of decision support tools for both patients and physicians. Over 6.7 million Americans currently have heart failure, and this number is expected to increase to >8 million by 2030. 1 Heart failure drives substantial morbidity and mortality, both of which continue to rise despite advances in care. 1 In patients diagnosed with heart failure with reduced ejection fraction (HFrEF), the use of guideline‐directed medical therapy (GDMT) helps to combat these trends and is associated with substantial improvements in mortality, cardiac function, and quality of life. These benefits increase incrementally as medications are added and titrated to maximally tolerated doses. 2 Despite this, GDMT continues to be underused and underdosed, especially in the outpatient setting where the initiation or up‐titration of medication remains poor. 3 , 4 Although medical barriers such as hypotension or hyperkalemia are common, other more modifiable barriers such as clinical inertia, patient hesitancy, and cost also likely play important roles. 4 , 5 , 6 , 7 Without efforts to improve prescription rates, patients will not attain optimal outcomes from available medications. Given the poor clinical outcomes associated with HFrEF, it is imperative to develop interventions that can be readily integrated into practice to effectively address this gap. Clinician education, outpatient audits, and electronic medical record alerts have demonstrated mixed results, with some studies demonstrating modest benefits and others having been limited by inconsistences across medication classes or inclusion of higher risk patients. 8 , 9 , 10 , 11 , 12 , 13 , 14 Patient activation tools offer an attractive option to improve optimization efforts by prompting conversations regarding medication changes and allowing for shared decision making. In the EPIC‐HF trial (Electronically Delivered Patient‐Activation Tool for Intensification of Medications for Chronic Heart Failure With Reduced Ejection Fraction), a multimodality checklist‐based patient‐activation intervention resulted in a 19% absolute increase in GDMT intensification at a single health center; however, this tool has yet to be implemented at a larger scale. 15 Additionally, patterns in patient–clinician conversations regarding the optimization of GDMT, and the impact that patient activation tools have on these discussions, remain unclear. In this study, we sought to address the present knowledge gap by retrospectively analyzing discussions between clinicians and ambulatory patients with HFrEF from the POCKET‐COST‐HF (Integrating Cost Into Shared Decision‐Making for Heart Failure With Reduced Ejection Fraction) trial (NCT04793880), a cluster‐randomized trial in which participants were presented with a checklist‐based patient activation tool based on EPIC‐HF. Recorded encounters offer unique insights into patterns of clinician‐patient communication patterns. The primary goal of the POCKET‐COST‐HF trial was to assess the impact of cost disclosure on medication discussions. 16 The goal of this secondary analysis was to analyze these encounters to characterize patterns of discussion more specifically around barriers to GDMT optimization in order to identify opportunities to refine decision support tools to better address barriers and improve optimization rates. METHODS Data supporting the findings of this study are available from the corresponding author upon reasonable request for purposes of reproduction of results. Study Design This was a secondary analysis of transcripts of patient encounters from the POCKET‐COST‐HF trial, a stepped‐wedge cluster randomized trial implementing a checklist‐based decision‐support tool focusing on price transparency. 16 In this trial, patients and clinicians in both the control and intervention groups were provided with an adapted version of the EPIC‐HF checklist that outlined approved medications and target doses for HFrEF (Figure S1). Patients in the intervention group received a checklist that also contained patient‐specific cost information. These checklists were distributed at the time of outpatient clinic visits, which were audio recorded and transcribed. Setting and Participants Patient encounters occurred at HF and cardiology clinics within 2 large tertiary care centers—Emory Healthcare and University of Colorado—between June 2021 and August 2023. English‐speaking adult patients (≥18 years) with a diagnosis of HFrEF were included in the analysis. All patients provided informed consent, and the study was reviewed and approved by the Emory University Institutional Review Board. Baseline patient demographics are included in Table 1. Table 1Patient Characteristics (n=247)CharacteristicsOverall (n=247)Age, y, mean±SD62.9 (13.9)Sex, No. (%)Male174 (70.5)Race, No. (%)White158 (64.0)Black65 (26.3)>1 race11 (4.5)Asian3 (1.2)American Indian or Alaska Native2 (0.8)Native Hawaiian or Pacific Islander1 (0.4)Ethnicity, No. (%)Hispanic8 (3.2)Non‐Hispanic236 (95.6)Not reported3 (1.2)Baseline medications, No. (%)Medications where cost was disclosedSacubitril/valsartan114 (46.2)Eplerenone12 (4.9)Canagliflozin1 (0.4)Dapagliflozin64 (25.9)Empagliflozin41 (16.6)Ivabradine3 (1.2)Vericiguat0Other components of guideline‐directed medical therapyBeta blockers234 (94.7)Angiotensin‐converting enzyme inhibitors/angiotensin receptor blockers88 (35.6)Spironolactone144 (58.3)Insurance, No. (%)Private82 (33.2)Medicare Part D89 (36.0)Medicare Advantage29 (11.7)Medicaid18 (7.3)Uninsured20 (8.1)Government (TriCare and Department of Defense)8 (3.2)Veterans Affairs1 (0.4)Other* 7 (2.8) * None of the listed categories. Patient Characteristics (n=247) None of the listed categories. Statistical Analysis Interviews were audio recorded, transcribed verbatim, deidentified, and verified/corrected for errors. Transcripts were imported into MAXQDA 12, a qualitative analysis software package, for coding. Transcripts from all clinical encounters in the parent trial were included in this analysis. The primary analytic aim was qualitative description, focusing on understanding the range of reasons within the clinician–patient encounter that GDMT optimization was not pursued—either from the patient or clinician perspective—and ways the checklist was or was not integrated into the clinician–patient interaction. The preliminary codebook for this analysis was developed by the research team a priori based on known potential barriers to GDMT optimization in clinical practice such as medical barriers, cost, and clinical inertia (Table S1). 17 , 18 Other domains included assessments of worsening HFrEF, discussions about medication intensification and optimization, and ways in which the checklist was used by patients and clinicians during encounters. This codebook was expanded and refined inductively as themes emerged during transcript review using constant comparison, and then finalized after review (see Tables 2 and 3; Table S2). Consistent with the descriptive aim of this study, simple descriptive statistics were tabulated to indicate the frequency of particular patterns of discussion. Table 2Barriers to Intensification or Optimization (n=152)SubcodesQuotes Medical barriers/side effects as barrier to intensification or optimization—74.3% (n=113) Adverse effects, intolerances, or other reported medical issues that prevent adding GDMT or increasing dosages Clinician: In theory, the idea is [to] try and get to the maximum dose of all these different classes that we can. I think we're pretty close to being there. We're a little limited by kidneys. Then we're a little limited by … some blood pressure issues in the past … I don't think I want to change things here Clinician: Yeah. We've never tried you on Entresto, have we? Patient: No. Clinician: The issue is the blood pressure effect. It would replace Losartan, but it's a far more potent hypotensive drug. Patient: Oh. Clinician: If there's been issues, then … Patient: I'm sensitive. Clinician: let's live with that. Deescalation—6.6% (n=10) A reduction in dose or discontinuation of a GDMT agent Clinician: I don't love doing this, but I think we should perhaps lower that dose of Entresto to the low dose. I think this is part blood pressure related. I'd also say you could probably—you could increase your water intake by—at least you could have another 2 cups per day. You look dry. Cost as a barrier—29.6% (n=45) GDMT intensification or optimization not pursued because of actual or anticipated cost. Clinician: Oh. No. Okay. You're able to get the Entresto? Patient: I've been out of [Entresto] for a month … I got the Eliquis. I don't have the money to get all of’ em, just this one. Clinician: There are a number of potential options for medicines that we're limited by … what you'd have to pay. Some of these medicines can be $4 or $5000 a year. Inertia/resistance—13.8% (n=21) Hesitancy on behalf of either the patient or clinician to optimize GDMT as a result of perceived clinical stability, desire to limit polypharmacy, etc. Patient: I don't like adding more pills without taking some away. At this point I'm takin’ so many. Clinician: Most folks, when they take it, they don't feel anything. They're, like, “Well, why am I taking it?” I say, “Because we know that taking it helps keep you out of the hospital and live longer,” and do those kind of situations. You know, I think it's a good medicine. I also understand if you want to say, well, let's just, you know, don't rock the boat, 'cause you're doing well, that's also a wisdom in and of itself. Patient: Okay, well, yeah. Unless I've got something that kind of stands out, you know, I I guess part of that is, I'm really happy with the medicines I'm taking. Other barriers—18.4% (n=28) When intensification is discussed but not pursued due to timing, confusion regarding use, or factors otherwise not accounted for by other subcodes. Clinician: Jardiance. It starts with the letter “J.” Either “J” or empagliflozin … You take 10 milligrams. You're still taking it? Or you're off of it? You're off of it? Is there a reason why? Patient: Yeah. Because I read the side effects. Clinician: You're right. Initially there is a little impairment in the kidney function. However, the studies are done on medicines like … empagliflozin and dapagliflozin have shown that over years actually the kidney function not only is preserved, but it improves in folks who have diabetes and heart conditions. Clinician: I just don't wanna add anything, especially with you just starting back under treatments for the prostate cancer, so anything that might lower your blood pressure more than where we are.GDMT indicates guideline‐directed medical therapy. Barriers to Intensification or Optimization (n=152) Medical barriers/side effects as barrier to intensification or optimization—74.3% (n=113) Adverse effects, intolerances, or other reported medical issues that prevent adding GDMT or increasing dosages Clinician: Yeah. We've never tried you on Entresto, have we? Patient: No. Clinician: The issue is the blood pressure effect. It would replace Losartan, but it's a far more potent hypotensive drug. Patient: Oh. Clinician: If there's been issues, then … Patient: I'm sensitive. Clinician: let's live with that. Deescalation—6.6% (n=10) A reduction in dose or discontinuation of a GDMT agent Cost as a barrier—29.6% (n=45) GDMT intensification or optimization not pursued because of actual or anticipated cost. Clinician: Oh. No. Okay. You're able to get the Entresto? Patient: I've been out of [Entresto] for a month … I got the Eliquis. I don't have the money to get all of’ em, just this one. Inertia/resistance—13.8% (n=21) Hesitancy on behalf of either the patient or clinician to optimize GDMT as a result of perceived clinical stability, desire to limit polypharmacy, etc. Clinician: Most folks, when they take it, they don't feel anything. They're, like, “Well, why am I taking it?” I say, “Because we know that taking it helps keep you out of the hospital and live longer,” and do those kind of situations. You know, I think it's a good medicine. I also understand if you want to say, well, let's just, you know, don't rock the boat, 'cause you're doing well, that's also a wisdom in and of itself. Patient: Okay, well, yeah. Unless I've got something that kind of stands out, you know, I I guess part of that is, I'm really happy with the medicines I'm taking. Other barriers—18.4% (n=28) When intensification is discussed but not pursued due to timing, confusion regarding use, or factors otherwise not accounted for by other subcodes. Clinician: Jardiance. It starts with the letter “J.” Either “J” or empagliflozin … You take 10 milligrams. You're still taking it? Or you're off of it? You're off of it? Is there a reason why? Patient: Yeah. Because I read the side effects. Clinician: You're right. Initially there is a little impairment in the kidney function. However, the studies are done on medicines like … empagliflozin and dapagliflozin have shown that over years actually the kidney function not only is preserved, but it improves in folks who have diabetes and heart conditions. GDMT indicates guideline‐directed medical therapy. Table 3Ways the Checklist Is Used or Discussed by Clinicians and Patients (n=120)SubcodesQuotes Brief reference—15% (n=18) Checklist is mentioned briefly during an encounter but is not used to review or change medication regimen. Clinician: This is a great just reference sheet to think—to see all the common medicines that we use. Uses to review−19.2% (n=23) Checklist is used to review current medication regimen but no attempt to intensify or optimize GDMT is made. Clinician: Let me go over this … We can just go over it together real quick, 'cause you're already on the good stuff … you're taking the torsemide … you're on carvedilol … and then you're on Entresto … you take this spironolactone … Then I'm so happy they started you on the Jardiance. Uses but doesn't make changes—26.7% (n=32) Checklist is reviewed in an attempt to optimize GDMT, but barriers are identified. Clinician: These are just all the classes of medications for heart failure. We've tried ACE [angiotensin‐converting enzyme] inhibitors, and then you were taken off the lisinopril because your blood pressure got too low and also your kidneys. We were on spironolactone at one point—I'm gonna make little dashes here so you understand what that is—and we were taken off of that both because of kidneys and high potassium, okay? Uses and makes changes—39.2% (n=47) Checklist is reviewed and leveraged to optimize GDMT. Clinician: There's another medicine that's out there. It's called Farxiga. That's a medicine we would recommend we start at this point. It's on the list, there. It is one of these right here.Clinician: The purpose of increasing the medications is you get a better benefit for heart failure as we go up higher. We like to try to get to target dosages, and these are the target dosages listed out here … Today, what I wanted to do is increase the spironolactone to a whole pill, go to 25 milligramsOther checklist subcodes (n=120) Reference to cost—42.5% (n=51) Patient or clinician references the medication costs on the checklist Clinician: Okay. Let's go through this list and keep this list because this is helpful. I like how they put the cost up here.Clinician: Here we got this list. The other options are—one of the better medicines for this kind of heart disease is this newer medicine which is called Entresto. It's a twice a day medicine. One of the main issues is it can be very expensive, but at least our estimate is that it will cost you only $8 a month, which is excellent. Your insurance is gonna cover the great majority of this medicine, Checklist confusion or misunderstandings – 21.7% (n=26) Confusion about checklist use, labeling, or about items in the checklist (ex. confusion whether it is 30‐d vs 90‐d cost; questions about checklist headings) Patient: I was lookin’ at this list right here the lady gave me … What it is potassium raising? I take Inspra, but I'm takin’ potassium pill, too. Clinician: All that is to keep the potassium level up, and in the past, your potassium level has been … Patient: …and then I don't have a sodium problem, in fact, I'm kind of at the low end on the sodium, so I don't see why I would need anything like this. Clinician: Yeah, this was a medicine—it doesn't really affect your sodium levels, per se, but it does treat … your heart failure syndrome. Clinician: Okay, I wanted to point out here—I didn't know how much your Jardiance was costing you. Patient: Oh, it's $120 a month for a 90‐day supply. Clinician: For a 90‐day supply? Okay. All right, then that's fine. Patient: Yeah. Clinician: That is 'cause I think this is a monthly. Patient education—23.3% (n=28) Clinician uses the checklist for further patient education on medications or heart failure with reduced ejection fraction Patient: On this list under adrenaline beta blockers, it says metoprolol, but the word after it is not tartrate which is the one he takes, so I didn't know if I should check that one or not. His milligrams is much lower than 200. Clinician: Right, so really one good option, and this really should be a lateral switch. I don't know that it'll make a huge difference in the blood pressure, but it's possible. Metoprolol tartrate is a shorter‐acting beta blocker. It can help with lowering blood pressure and improving the heart rate control when somebody's in AFib [atrial fibrillation], but it's not as well research‐proven compared with metoprolol succinate which is Toprol XL or generic Toprol XL when we talk about heart failure. GDMT indicates guideline‐directed medical therapy. Ways the Checklist Is Used or Discussed by Clinicians and Patients (n=120) Brief reference—15% (n=18) Checklist is mentioned briefly during an encounter but is not used to review or change medication regimen. Uses to review−19.2% (n=23) Checklist is used to review current medication regimen but no attempt to intensify or optimize GDMT is made. Uses but doesn't make changes—26.7% (n=32) Checklist is reviewed in an attempt to optimize GDMT, but barriers are identified. Uses and makes changes—39.2% (n=47) Checklist is reviewed and leveraged to optimize GDMT. Reference to cost—42.5% (n=51) Patient or clinician references the medication costs on the checklist Checklist confusion or misunderstandings – 21.7% (n=26) Confusion about checklist use, labeling, or about items in the checklist (ex. confusion whether it is 30‐d vs 90‐d cost; questions about checklist headings) Patient: I was lookin’ at this list right here the lady gave me … What it is potassium raising? I take Inspra, but I'm takin’ potassium pill, too. Clinician: All that is to keep the potassium level up, and in the past, your potassium level has been … Patient: …and then I don't have a sodium problem, in fact, I'm kind of at the low end on the sodium, so I don't see why I would need anything like this. Clinician: Yeah, this was a medicine—it doesn't really affect your sodium levels, per se, but it does treat … your heart failure syndrome. Clinician: Okay, I wanted to point out here—I didn't know how much your Jardiance was costing you. Patient: Oh, it's $120 a month for a 90‐day supply. Clinician: For a 90‐day supply? Okay. All right, then that's fine. Patient: Yeah. Clinician: That is 'cause I think this is a monthly. Patient education—23.3% (n=28) Clinician uses the checklist for further patient education on medications or heart failure with reduced ejection fraction Patient: On this list under adrenaline beta blockers, it says metoprolol, but the word after it is not tartrate which is the one he takes, so I didn't know if I should check that one or not. His milligrams is much lower than 200. Clinician: Right, so really one good option, and this really should be a lateral switch. I don't know that it'll make a huge difference in the blood pressure, but it's possible. Metoprolol tartrate is a shorter‐acting beta blocker. It can help with lowering blood pressure and improving the heart rate control when somebody's in AFib [atrial fibrillation], but it's not as well research‐proven compared with metoprolol succinate which is Toprol XL or generic Toprol XL when we talk about heart failure. GDMT indicates guideline‐directed medical therapy. Transcripts were double‐coded by C.S. and one other coder (A.M., B.M., E.U., or E.W.), using the final codebook, and discrepancies were reviewed with the coding team and principal investigator (N.D.) and resolved by consensus. RESULTS Patient Characteristics From June 2021 to August 2023 247 patients were seen by 39 clinicians in the advanced HF and general cardiology clinics affiliated with the Emory Healthcare and University of Colorado Health systems. Clinicians were a mix of advanced HF and transplant cardiologists (18), general cardiologists (12), and advanced practice providers (9). The mean age of the patients was 62.9 years old, 29.5% of patients were female, 26.3% were Black, and 3.2% were Hispanic/LatinX (Table 1). Baseline GDMT use was high with 95% of patients taking a beta blocker, 81.8% taking an angiotensin‐converting enzyme inhibitor, angiotensin receptor blocker, or angiotensin receptor neprilysin inhibitor (46.2% taking angiotensin receptor‐neprilysin inhibitors and 35.6% angiotensin converting enzyme inhibitor/angiotensin receptor blocker), 63% taking a mineralocorticoid receptor antagonist, and 43% taking a (SGLT2) sodium glucose cotransporter 2 inhibitor. In this cohort, 27.1% of patients were taking all 4 classes of GDMT, 37.7% were taking 3 classes, 24.7% were taking 2 classes, 9.3% were taking 1 class, and 1.2% were not taking any component of GDMT. Primary insurance coverage was as follows: 33.2% private, 36% Medicare Part D, 11.7% Medicare Advantage, 7.3% Medicaid, and 3.6% government or Veterans Affairs. 8.1% were uninsured. Key Domains Major Themes Across All Encounters Clinicians asked patients questions to determine the status and trajectory of their HF in nearly every encounter (96%) (Table S2). These ranged from generalized assessments to more categorical assessments of a patient's relative functional status. A majority (75%) of encounters included discussions regarding intensification or optimization of GDMT and typically highlighted the established benefits of optimization. Barriers to GDMT optimization were identified in 61.5% of encounters, and the checklist tool was explicitly discussed in 48.6%. Barriers to Intensification or Optimization Medical Barriers Medical barriers were the most frequently identified reason for not optimizing GDMT, representing 74.3% of identified barriers (Table 2). Common medical barriers included renal dysfunction, electrolyte abnormalities, and hypotension. However, there were several instances where either the true underlying medical barrier was not clearly identified, medical optimization was not pursued due to perceived historic barriers, or previously identified barriers were likely no longer present. One patient described themselves as “sensitive” precluding optimization of GDMT, and 1 clinician noted, “Your potassium levels had been high, so I think it's okay to just hold everything as is.” Other times, reported barriers were not clearly related to the drug in question, such as “Metoprolol … made me sick … throwing up and passing out.” Rarely (6.6%), GDMT was deescalated at the time of the visit because of intolerance to therapy (eg, hypotension) or decline in patients' cardiac or renal function. Cost as a Barrier Medication cost was a commonly identified barrier (29.6%), with both patients and clinicians sharing concerns regarding affordability of the prescribed HF medications. In several instances, the financial burden of the medication directly affected patient adherence to GDMT. In others, clinicians referenced a potential financial burden, which was enough to affect medication choice, especially if the medications in question had frequently been found to be prohibitively expensive for other patients. One clinician noted, “There are a number of potential options for medicines that we're limited by … what you'd have to pay. Some of these medicines can be 4 or 5 thousand dollars a year.” Some of the cost barrier conversations were preceded by conversations about general health care costs (eg, other meds, HF medications, visit copays, etc.) that may have prompted these concerns. Among encounters in which cost barriers were identified, clinicians and patients initiated discussions about cost within the encounter at similar frequencies (51.1% and 48.9% respectively). Inertia/Resistance Inertia related to medication was another commonly encountered barrier, comprising 13.8% of all encountered barriers. In 66% of these encounters, clinicians exhibited inertia against GDMT intensification. Common themes that contributed to inertia among clinicians included minimal symptom burden and perceived clinical stability. For example, 1 physician said, “I also understand if you want to say … don't rock the boat, [because] you're doing well.” Inertia was often grounded in resistance among patients to perceived polypharmacy. One patient noted, “I don't like adding more pills without taking some away.” Other times, there was a hesitancy to make changes when the patient perceived themselves as doing well. One patient remarked, “Unless I've got something that kind of stands out … I'm really happy with the medicines that I'm taking.” Other Identified Barriers A final 18.4% of barriers identified included “other” categories such as clinicians wanting to wait until after certain procedures or additional appointments, wishing to allow another clinician to make medication changes, or due to confusion regarding the medications that the patient is currently taking. Checklist Use Either the clinician or the patient explicitly referenced the checklist‐based patient activation tool in 48.6% of encounters (Table 3), with the nature of use varying among encounters. In some instances, the checklist was briefly mentioned (15%). In other encounters (19.2%), the checklist was used to reconcile medications that the patient was taking but was not clearly leveraged into a discussion about optimization. In still other encounters, the checklist was used to discuss medications and identify barriers to further optimization (26.7%). Finally, some clinicians made changes and optimized GDMT based on discussions that involved the checklist (39.2%). When using the checklist, 42.5% of encounters involved a discussion regarding medication cost, and 23.3% used the checklist to conduct patient education. In 21.7% of encounters, there was evidence of confusion or misunderstanding related to use of the checklist; often this was regarding descriptions or labeling of medications on the checklist. DISCUSSION Despite clear benefits of GDMT for patients with HFrEF, rates of full implementation at target doses remain suboptimal. Attempts at decision support have demonstrated mixed results. 8 Importantly, the EPIC‐HF trial demonstrated an increase in intensification, primarily driven by an increase in beta‐blocker therapy, with a multimodality patient activation tool incorporating a checklist. 15 However, this tool was not truly implemented at a system level in that trial, and there may be alterations that could increase its impact. This secondary analysis of recorded encounters within the POCKETCOST‐HF trial—in which the EPIC‐HF checklist was used—offers important insights into opportunities for refinement of decision support strategies to facilitate implementation of this and other tools at scale within clinical practice. This study demonstrated high rates of discussion regarding HF therapy optimization (75% of encounters) in a context where all patients and clinicians received the checklist. This is encouraging, as only 27% of patients were on 4‐drug GDMT. However, these data also demonstrate room for improvement. Specifically, about 25% of encounters did not include a discussion about optimization of therapy, which suggests that there is value in continuing to explore and refine strategies for delivering decision support and patient activation tools to stimulate conversations about optimization. In POCKET‐COST HF, for example, a hard copy of the checklist was provided to patients and clinicians at the time of the visit. Use of patient portals for delivery to patients in advance and electronic medical record displays for clinicians are important, scalable tools that may be used and studied to potentially enhance impact of these interventions. 19 , 20 Additionally, these encounters revealed a number of barriers to optimization, some of which may be more amenable than others to refinement of decision support materials. For example, medical barriers are well established and were the most common barriers identified. Medical barriers are unlikely to be directly affected by a patient activation tool, though emphasizing the importance of GDMT could help to stimulate more in‐depth optimization conversations and especially evaluation—and importantly reevaluation—of reported historical adverse effects or barriers to GDMT. Some symptoms thought to be barriers, for example, have been reported at similar rates in both intervention and placebo groups in landmark trials for GDMT, with many being attributable to HFrEF itself and not to the medication in question. 21 Similarly, barriers such as renal function and blood pressure are not stagnant. In order to stimulate discussion about whether medical barriers are still present and significant, we have explicitly incorporated more direct messaging about the benefits of GDMT into a subsequent iteration of this checklist. Perhaps the barrier that a checklist‐based tool is most well positioned to affect relates to clinical inertia, a phenomenon we observed in both patients and clinicians. There were examples in which the checklist was used to prompt discussions about the importance of both adding classes of medications and increasing doses, but there also remained instances in which patients expressed the desire to minimize medications, and clinicians were hesitant to “rock the boat.” This phenomenon has been observed in other studies as well, and it may have been less frequent in our study because all patients and clinicians received the checklist or potentially because many of these encounters occurred in HF specialty clinics. 4 , 6 Inertia, however, was still a common occurrence and is important to address, because the treatment goal in HFrEF is to have the patient taking the highest tolerated dose of all possible components of GDMT. Prior studies have often focused on patient‐level factors that affect adherence and patient decision‐making and this continues to be of importance when optimizing future decision‐support tools. 19 , 20 However, it is also important to note that the present study found that the majority of instances of clinical inertia occurred on behalf of clinicians, a trend highlighted in a recent study by Greene et al. 5 In this regard, targeting refinement toward patients alone may be insufficient, and future iterations may benefit from more clinician‐directed refinement. Helping to shift the attitude of patients and clinicians toward this goal of medication maximization (rather than minimization) remains an important role for checklist‐based support, and we have incorporated more explicit messaging about the importance of this shift in orientation within a subsequent version of the checklist. There were a significant number of encounters where the checklist was used, but optimization of a patient's medication regimen was deferred at the time of the visit. However, we observed that during many of these encounters patients and clinicians discussed the potential for making changes in the future. This suggests the potential for decision‐support tools to aid in optimization efforts over time by helping to identify a plan to intensify therapy at a future visit or between visits. Refinement of the checklist to better delineate therapeutic goals may help to encourage optimization longitudinally by facilitating the development of a clear plan for continued intensification, which may also allow for some of the encountered barriers to be anticipated and proactively addressed. This may be achieved through greater integration into patient navigation tools or other electronic medical record mechanisms. Another major barrier to optimization identified in these encounters was cost. Out‐of‐pocket costs have become a major concern with the approval of newer therapies for HFrEF that remain unavailable as generic medications. This has frequently been cited as a driver of suboptimal GDMT use and the persistently poor outcomes in observational studies of patients with HFrEF despite the additive benefit of these agents. 7 , 22 Importantly, the parent trial within which these data were gathered specifically addressed provision of out‐of‐pocket costs and demonstrated an increase in rates of conversations about cost when those costs were included in the checklist. 16 There are significant complexities in integrating patient‐specific costs into decision‐support tools, because patients may face highly variable out‐of‐pocket costs, but highlighting potential ranges of cost, for example, may help to bring cost to the forefront. In a subsequent version of this tool, we have indicated the medications for which a range of costs may be encountered in order to proactively identify when cost may be relevant and to stimulate appropriate discussion and cost mitigation strategies. When combined with the growing availability of electronic medical record‐based cost estimation tools, these considerations may be better integrated into decision‐making. These data provide important insights regarding ways to optimize the checklist approach for effective integration into clinical encounters, and we have indicated some of the changes we have made in subsequent versions. Portal messaging, for example, which is highly scalable, can deliver messaging to patients in advance and can be leveraged to track uptake and integration. In former iterations of the checklist, labels like “potassium raising,” and “sodium glucose lowering” led to confusion such as 1 patient remarking “I don't have a sodium problem, in fact, I'm kind of at the low end on the sodium, so I don't see why I would need anything like this.” We have simplified labels within the checklist and have removed vague wording surrounding mechanism of action for critical categories of medications that we hope will enhance patients' understanding and prioritize the 4 main pillars of GDMT in order to help optimize therapy. Enhancement of clear messaging regarding the benefits associated with intensification of optimal GDMT may increase buy‐in regarding the importance of intensification of therapy for patients and potentially help to combat clinical inertia, especially for patients who are clinically doing well. This could be further supplemented with the provision of online resources reviewing the efficacy and benefits of GDMT or dedicated pharmacy and nurse‐driven education reviewing the patient activation tool and GDMT benefits in some settings. Importantly, clinician‐facing materials may also be harnessed to enhance clinician uptake and to combat trends in clinical inertia on the part of clinicians, which was observed here among clinicians and was not isolated to patients. 5 Finally, continuing to develop strategies to address cost is important, as cost was observed to be a significant consideration in many encounters. In addition to the “flagging” of cost that we described, there was at times confusion regarding issues such as whether the costs displayed reflected 30 or 90‐day supply. Future iterations could consider the addition of supplemental information, including links to relevant online resources, that provide further information regarding specific cost considerations and relevant patient assistance programs to allow for more informed cost‐based decision‐making. Additionally, with changes to Medicare Part D out‐of‐pocket maximums, future iterations of the tool may be able to help patients understand their cost burden over time more effectively. This could be additionally supplemented by integration of increasingly available cost predictor tools into the checklist format so that comparative cost information is available. These additions and tools would allow for a more specific understanding of a patient's true cost burden and allow this information to be an asset to physicians' efforts to aggressively uptitrate GDMT medications. The most important step will be testing these refinements to demonstrate whether we “move the needle” on optimization, especially given that efforts to provide decision support for GDMT have had mixed results. The TAILORD‐HF trial (Tailored Electronic Intervention to Improve Therapy in a Diverse Cohort of Patients With Heart Failure; NCT06847438) and IICAPTAIN trial (Implementation and Interaction of Clinician and Patient‐facing Tools Aiming to Intensify Neurohormonal Medicines for Heart Failure; NCT06526988) are both now ongoing and build on further refined versions of the EPIC‐HF checklist; together these studies will provide important data on the impact of different implementation strategies. As indicated in data from this analysis, there are clearly residual areas for improvement in communication about GDMT. These opportunities for improvement, importantly, were observed in a context where many patients were under the care of HF specialists; gaps are likely even more significant and prevalent in the context of primary care or general cardiology. Limitations Our study has several limitations. First, there was no control group, as the study was embedded within a trial in which all patients and clinicians were provided a checklist‐based decision‐support tool, with the intervention group having more cost‐specific information. Second, the encounters occurred at 2 large academic health systems, and many of the encounters were in HF specialty clinics, which may limit external validity. We expect the relative effect of the checklist to be larger in a general cardiology or primary care setting, where the management of HFrEF is not always the primary focus and clinicians may have a lower comfort level with prescribing or uptitrating these medications. Third, by recording encounters there is the potential for confounding as a result of the Hawthorne effect. This impact typically diminishes over time with repeated clinic encounters as clinicians become comfortable with encounters being recorded; however, it could have affected the rates at which the checklist was discussed in this study. Finally, the checklist was provided to patients and clinicians at the time of the visit, and it is unclear what the relative uptake and impact would be if it was provided before the visit when implemented at scale and how this would affect its efficacy. The latter question, as mentioned, is a focus of an ongoing study. CONCLUSIONS Checklist‐based decision‐support tools have demonstrated significant promise in improving uptake of GDMT. However, their use has not been demonstrated at scale, and there may be further opportunities for optimizing clinician‐patient communication to remove barriers to optimal care. These data suggest that, although medical barriers are common and unavoidable, decision‐support tools have the potential to address considerations such as cost, clinical inertia, and patient education. Refining future tools to address these barriers could enhance their impact and improve efforts to optimize GDMT in HFrEF. Sources of Funding This study was funded by Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA. The original POCKET‐COST‐HF Trial was funded by the Agency for Healthcare Research and Quality (R01HS026081). Disclosures Engels N. Obi, Andra S. Stevenson, and Molly Tiedeken are employees of Merck Sharp & Dohme LLC, a subsidiary of Merck & Co., Inc., Rahway, NJ, USA, and may own restricted stock units or stock options in Merck & Co., Inc., Rahway, NJ, USA. Alanna A. Morris is now an employee of Bayer US LLC and may own restricted stock units or stock options in Bayer AG. Jamie Diamond reports research funding from Novo Nordisk MB and receives research funding from the National Institutes of Health and the National Heart, Lung, and Blood Institute. Neal W. Dickert reports research funding from National Institutes of Health, Patient Centered Outcomes Research Institute, Agency for Healthcare Research and Quality, American Heart Association, and Abiomed and consulting activity for Abiomed. The remaining authors have no disclosures to report. Supporting information Figure S1 Tables S1–S2