Abstract
BACKGROUND
Scalable and safe approaches for heart failure guideline-directed medical therapy (GDMT) optimization are needed.
OBJECTIVES
The authors assessed the safety and effectiveness of a virtual care team guided strategy on GDMT optimization in hospitalized patients with heart failure with reduced ejection fraction (HFrEF).
METHODS
In a multicenter implementation trial, we allocated 252 hospital encounters in patients with left ventricular ejection fraction ≤40% to a virtual care team guided strategy (107 encounters among 83 patients) or usual care (145 encounters among 115 patients) across 3 centers in an integrated health system. In the virtual care team group, clinicians received up to 1 daily GDMT optimization suggestion from a physician-pharmacist team. The primary effectiveness outcome was in-hospital change in GDMT optimization score (+2 initiations, +1 dose up-titrations, −1 dose down-titrations, −2 discontinuations summed across classes). In-hospital safety outcomes were adjudicated by an independent clinical events committee.
RESULTS
Among 252 encounters, the mean age was 69 ± 14 years, 85 (34%) were women, 35 (14%) were Black, and 43 (17%) were Hispanic. The virtual care team strategy significantly improved GDMT optimization scores vs usual care (adjusted difference: +1.2; 95% CI: 0.7-1.8; P < 0.001). New initiations (44% vs 23%; absolute difference: +21%; P = 0.001) and net intensifications (44% vs 24%; absolute difference: +20%; P = 0.002) during hospitalization were higher in the virtual care team group, translating to a number needed to intervene of 5 encounters. Overall, 23 (21%) in the virtual care team group and 40 (28%) in usual care experienced 1 or more adverse events (P = 0.30). Acute kidney injury, bradycardia, hypotension, hyperkalemia, and hospital length of stay were similar between groups.
CONCLUSIONS
Among patients hospitalized with HFrEF, a virtual care team guided strategy for GDMT optimization was safe and improved GDMT across multiple hospitals in an integrated health system. Virtual teams represent a centralized and scalable approach to optimize GDMT.
Keywords: heart failure, guideline-directed medical therapy, implementation, virtual care team
Multiple evidence-based disease-modifying therapies are well studied and available for the treatment of heart failure with reduced ejection fraction (HFrEF).1 Rapid implementation of multidrug comprehensive medical therapy for HFrEF is now guideline endorsed and is estimated to meaningfully extend survival free of worsening heart failure (HF) events.2 Despite this, contemporary global data suggest that implementation of guideline-directed medical therapy (GDMT) for HFrEF remains low and that optimal doses are rarely, slowly, and inequitably achieved.3–6 Thus, there is an imperative to develop and evaluate implementation strategies to improve GDMT use in this population.
Ideal implementation strategies should be effective, safe, and scalable. Electronic health record (EHR)-based clinical decision nudges may improve GDMT use,7 but improvements in HF therapeutic optimization have been relatively modest and alert fatigue might limit the potential to drive multidrug implementation. Recently, the STRONG-HF (Safety, Tolerability and Efficacy of Rapid Optimization, Helped by NT-proBNP and GDF-15, of Heart Failure Therapies; NCT03412201) trial demonstrated a structured protocol with regular in-person visits and biomarker-guided optimization safely improved GDMT utilization and clinical outcomes after an acute HF presentation.8,9 However, such an approach is resource and personnel intensive and may be challenging to implement in some care settings. Therefore, highly effective strategies that improve GDMT utilization but rely less on in person interactions, such that they can be efficiently scaled across health systems are needed.
Hospitalization, either for acute HF or for other reasons, represents an important opportunity for GDMT optimization, because it allows for frequent hemodynamic and laboratory monitoring, symptom assessment, and educational reinforcement.10 In an earlier nonrandomized pilot study at a single hospital, we found that a virtual care team strategy was safe and associated with improvements in GDMT use among patients with a history of HFrEF admitted for other nonprimary HF indications.11 We therefore designed the IMPLEMENT-HF (Implementation of Medical Therapy in Hospitalized Patients With Heart Failure With Reduced Ejection Fraction) trial to evaluate the effectiveness and safety of a virtual care team guided strategy vs usual care to improve GDMT utilization in hospitalized patients with new or established diagnoses of HFrEF across multiple hospitals caring for a diverse population in an integrated health system.
METHODS
IMPLEMENT-HF was a prospective implementation study among patients with HFrEF admitted to 3 hospitals within an integrated health care delivery system (Mass General Brigham, Boston, Massachusetts) from October 2021 through June 2022. Eligible patients were identified by an EHR-based query. To enhance pragmatism, allocation to a virtual care team guided strategy vs usual care was performed by birth month (6 months allocated to the control group; 6 months allocated to the intervention group). Study sites included an 812-bed quaternary care academic medical center (Brigham and Women’s Hospital, Boston, Massachusetts) and 2 community-based teaching hospitals (Brigham and Women’s Faulkner Hospital, Jamaica Plain, Massachusetts: 171 beds; Salem Hospital, Salem, Massachusetts: 395 beds). All hospitals had cardiovascular specialists; advanced heart failure services were centralized at the quaternary care center. Because of the health system quality improvement nature of the intervention supporting guideline-concordant care without any direct patient contact and with all data sourced from medical chart review, the Institutional Review Board determined that the study did not require patient-level informed consent.
STUDY POPULATION.
Hospitalized adult patients with previously or newly diagnosed HFrEF (left ventricular ejection fraction [LVEF] ≤40%) admitted for at least 1 midnight to a primary nonintensive care unit (ICU) medical or surgical service were eligible for inclusion. Eligible patients were identified daily from an EHR-generated report of admitted patients with LVEF ≤40% documented on a transthoracic echocardiogram performed in the health system within the previous 12 months. Patients with prior history of HFrEF were identified at the time of hospital admission or shortly thereafter and patients with de novo presentations of HFrEF were identified at the time of documentation of newly reduced LVEF based on available imaging. Daily screening for eligibility was performed by a centralized study physician via chart review across all 3 care entities. Patients were eligible across multiple hospitalization encounters. Patients admitted for acute HF or for other admission reasons were permitted as were patients with newly diagnosed (de-novo) HF during the index encounter. Exclusion criteria included ICU admission during the current hospitalization, inotropic or mechanical circulatory support use during the current admission, acute coronary syndrome, stroke, or major cardiovascular surgery within 30 days, systolic blood pressure <90 mm Hg in the preceding 24 hours, severe uncorrected valvular disease or moderate or greater right ventricular (RV) dysfunction on most recent transthoracic echocardiogram, admission for coronavirus disease 2019, and physician discretion (eg, ongoing acute kidney injury, possible upcoming surgery). Comprehensive trial eligibility criteria are provided in the Supplemental Appendix Exhibit 1.
INTERVENTION.
Eligible patients were allocated to receive a virtual care team guided strategy for GDMT optimization vs usual care. The virtual care team consisted of a centralized physician, study staff, and local pharmacist. Primary treating teams caring for patients allocated to the intervention group received up to once daily recommendations for GDMT optimization according to an evidence-based algorithm (Supplemental Appendix Exhibits 2 and 3). The protocol was based on American Heart Association/American College of Cardiology/Heart Failure Society of America guidelines,1,12 randomized clinical trial evidence, drug-specific U.S. Food and Drug Administration package inserts, and expert consensus documents.13,14 The primary goal of the protocol was to facilitate early treatment with 4 major therapeutic classes with known benefit in HFrEF in patients without contraindications: evidence-based β-blockers, angiotensin-converting enzyme (ACE) inhibitor/angiotensin receptor blocker (ARB)/angiotensin receptor–neprilysin inhibitor (ARNI), mineralocorticoid receptor antagonist (MRA), and sodium-glucose cotransporter-2 inhibitor (SGLT2i). Additional aims included switching from ACE inhibitor/ARB to ARNI in eligible patients followed by progressive up-titration of GDMT elements to target doses, with preference for evidence-based β-blockers.15 In patients not on ACE inhibitor or ARB, preference was to initiate ARNI de novo in the absence of contraindications. Class-wide indications, contraindications, and monitoring recommendations were also provided. If relative contraindications arose, the virtual care team could elect to not provide care suggestions on a particular day (eg, day before planned procedure) or to cease providing suggestions if there was a change in care plan (eg, change in goals of care). Class-specific considerations were highlighted to facilitate inpatient drug initiation, titration, and consolidation to chronic therapy (Supplemental Appendix Exhibit 4).
Algorithm-based recommendations were developed by study pharmacists; virtual care team members then reviewed recommendations in the context of patient-specific factors after further EHR review, including vital signs, laboratory tests, and progress notes (Supplemental Appendix Exhibit 5). Final care optimization recommendations were communicated by progress notes in the EHR. The note included detailed instructions regarding medication dosing, follow-up monitoring, possible adverse effects, information regarding efficacy of the recommended drug in HFrEF, and a virtual care team contact for any follow-up queries and communications. Text page communication was sent to treating teams to inform them of care suggestions.
Throughout this process, there was no contact between the virtual care team and patients. The team did not perform patient interviews, physical examinations, or volume status assessments. Virtual care team recommendations were communicated as suggestions (Supplemental Appendix Exhibit 6), and all decisions to adopt any recommendations was left to the discretion of each patient’s treating team.
OUTCOMES.
The primary effectiveness outcome was a composite in-hospital GDMT optimization score, defined as the sum of care optimization changes (+2 for new GDMT initiations, +1 for dose up-titrations) and care deoptimization changes (−2 for GDMT discontinuation, −1 for dose down-titrations). Scores were developed by comparing discharge medication regimens to those prior to admission. Transition from ACE inhibitor/ARB to ARNI was considered an initiation of ARNI but not a discontinuation of ACE inhibitor/ARB and was allotted a score of +2. Switching from ARNI to ACE inhibitor/ARB was considered a discontinuation of ARNI but not an initiation of ACE inhibitor/ARB and was allotted a score of −2 (Supplemental Appendix Exhibit 2). Dosing of each GDMT was indexed according to American College of Cardiology/American Heart Association target dosing (Supplemental Appendix Exhibit 4) at hospital admission and at hospital discharge. Changes made within the same drug class were categorized as up-titrations or down-titrations based on increases or decreases in percentage target dose, respectively. The composite GDMT score could range from −8 (indicating 4 discontinuations) to +8 (indicating 4 initiations).
Secondary effectiveness outcomes included the proportion of clinical encounters with GDMT initiations (when these therapies were not used at admission), the proportion of clinical encounters with either GDMT initiations and/or dose uptitrations, net GDMT intensification (optimization score >0), and changes in the prescription of individual GDMT elements from time of hospital admission to discharge.
Potential safety outcomes included the incidence of hypotension (systolic blood pressure <90 mm Hg on 3 consecutive measurements, vasopressor initiation, or ICU transfer because of hypotension), hyperkalemia (serum potassium >5.5 mmol/L or treatment with acute potassium-lowering therapies), AKI (doubling of admission serum creatinine or new kidney replacement therapy), bradycardia (heart rate <40 beats/min on 3 consecutive measurements, new temporary or permanent cardiac pacing, or treatment with acute heart rate–raising therapies [atropine, dopamine, isoproterenol, epinephrine]), and all-cause in-hospital death (Supplemental Appendix Exhibit 2). Safety outcomes were assessed and adjudicated by a blinded clinical events committee (CEC) consisting of 2 independent reviewers. In cases of discrepancy, a third reviewer (CEC chair) made final adjudicated determinations.
STATISTICAL ANALYSIS.
Effectiveness analyses on use and dosing of medical therapies were conducted according to intention-to-treat principles. Three encounters during which patients died during hospitalization were excluded from effectiveness analyses because discharge GDMT could not be determined in those encounters. Safety analyses were conducted among all 252 encounters. The primary endpoint was analyzed continuously by linear regression adjusted for baseline number of GDMT therapies. Beta-coefficients (95% CIs) derived from linear regression models for the primary endpoint reflect mean between-group differences in GDMT scores. Subgroup analyses for the primary endpoint were conducted by age, sex, race, ethnicity, primary language, chronicity of HF, admission indication, and site of enrollment. All effectiveness and safety outcomes were compared between strategy arms using logistic or linear regression, as appropriate, using clustered-robust standard errors to account for within-patient clustering. A 2-tailed P value of <0.05 was considered to be statistically significant without adjustment for multiplicity. Analyses were conducted with the use of Stata version 16 (Stata Corp).
RESULTS
STUDY COHORT.
Of 808 screened clinical encounters, 252 (31%) from 198 unique patients met inclusion criteria and were allocated to virtual care team guided strategy (107 encounters among 83 patients) or usual care (145 encounters among 115 patients) (Figure 1). Mean age was 69 ± 14 years, 85 (34%) were women, 183 (73%) were White, 35 (14%) were Black, 43 (17%) were Hispanic, and 30 (12%) were Spanish-speaking (Table 1). Baseline characteristics were generally balanced between study arms. Geographic distribution of patients’ residential communities is shown in Supplemental Appendix Exhibit 7. Median hospital length of stay was 6 days (IQR: 3-11 days) in the intervention group and 6 days (IQR: 3-10 days) in the usual care group (Supplemental Appendix Exhibit 8).
FIGURE 1. Study Design of IMPLEMENT-HF.
Flow chart summarizes screening, allocation, and follow-up in the IMPLEMENT-HF study, conducted across 3 hospitals in an integrated health system.
TABLE 1.
Patient Characteristics
| Virtual Care Team Strategy (n = 107) | Usual Care (n = 145) | |
|---|---|---|
| Demographic characteristics | ||
| Age, y | 70.3 ± 12.1 | 68.6 ± 14.8 |
| Women | 37 (34.6) | 48 (33.1) |
| Race | ||
| White | 83 (77.6) | 100 (70.9) |
| Black | 14 (13.1) | 21 (14.9) |
| Other | 10 (9.3) | 20 (14.2) |
| Hispanic ethnicity | 18 (17.0) | 25 (17.7) |
| Primary language | ||
| English | 90 (86.5) | 123 (85.4) |
| Spanish | 14 (13.5) | 16 (11.1) |
| Other | 0 (0.0) | 5 (3.5) |
|
| ||
| Initial presentation | ||
| Site of enrollment | ||
| Brigham and Women’s Hospital | 25 (23.4) | 49 (33.8) |
| Brigham Faulkner Hospital | 30 (28.0) | 37 (25.5) |
| Salem Hospital | 52 (48.6) | 59 (40.7) |
| Primary admission diagnosis of HF (vs other diagnosis) | 27 (25.2) | 35 (24.1) |
| De novo presentation of HF | 24 (22.4) | 26 (17.9) |
| Medicine admitting service (vs other service line) | 89 (83.2) | 111 (76.6) |
| Primary responding clinician | ||
| Attending | 36 (33.6) | 42 (29.0) |
| House staff | 53 (49.5) | 79 (54.5) |
| Advanced practice provider | 18 (16.8) | 24 (16.6) |
|
| ||
| Medical history | ||
| Left ventricular ejection fraction, % | 33.1 ± 9.4 | 32.4 ± 8.7 |
| Atrial fibrillation or flutter | 34 (31.8) | 53 (36.6) |
| Coronary artery disease | 51 (47.7) | 71 (49.0) |
| Cancera | 18 (16.8) | 24 (16.6) |
| Diabetes mellitus | 50 (46.7) | 56 (38.6) |
| End-stage kidney disease | 11 (10.3) | 18 (12.4) |
| Hyperlipidemia | 71 (66.4) | 79 (54.5) |
| Hypertension | 93 (86.9) | 112 (77.2) |
| Implantable cardioverter-defibrillator | 12 (11.2) | 35 (24.1) |
| Cardiac resynchronization therapy | 2 (1.9) | 8 (5.5) |
|
| ||
| Admission vital signs and laboratory measures | ||
| Systolic blood pressure, mm Hg | 133.6 ± 28.9 | 131.9 ± 25.2 |
| Heart rate, beats/min | 88.3 ± 21.1 | 88.9 ± 22.8 |
| Sodium, mEq/L | 138.0 ± 3.6 | 137.3 ± 4.4 |
| Potassium, mEq/L | 4.2 ± 0.6 | 4.3 ± 0.7 |
| eGFR, mL/min/1.73 m2 | 61.0 ± 31.3 | 61.6 ± 31.8 |
|
| ||
| Discharge vital signs and laboratory measures | ||
| Systolic blood pressure, mm Hg | 125.6 ± 19.5 | 123.3 ± 19.4 |
| Heart rate, beats/min | 79.1 ± 15.0 | 76.6 ± 14.8 |
| Sodium, mEq/L | 137.6 ± 3.1 | 138.0 ± 4.2 |
| Potassium, mEq/L | 4.1 ± 0.5 | 4.2 ± 0.5 |
| eGFR, mL/min/1.73 m2 | 63.5 ± 30.2 | 65.8 ± 32.6 |
Values are mean ± SD or n (%).
Active or diagnosed within the last 5 years, excluding superficial skin cancers.
eGFR = estimated glomerular filtration rate; HF = heart failure.
EFFECTIVENESS OUTCOMES.
A total of 187 unique recommendations were made by the virtual care team to clinicians caring for patients in the intervention arm, ranging from 0 to 8 per encounter. In encounters allocated to the intervention arm, 15 (14%), 46 (43%), and 46 (43%) had 0, 1, or 2 or more recommendations for optimization made during hospitalization by the virtual care team, respectively. In those surviving hospitalization, 182 (73%) were receiving evidence-based β-blocker, 108 (43%) ACE inhibitor or ARB, 55 (22%) MRA, 39 (16%) SGLT2i, and 27 (11%) ARNI at admission. Few patients were treated with multidrug regimens (26 [10%] on triple therapy and 23 [9%] on quadruple therapy) at admission. More patients were treated with MRA, SGLT2i, and ARNI during encounters allocated to the intervention group than to usual care at admission (Table 2).
TABLE 2.
Net Change in Medication Use From Admission to Dischargea
| Virtual Care Team Guided Strategy (n = 106 Encounters) | Usual Care (n = 143 Encounters) | % Change Admission to Discharge | P Value for Between-Arm Difference | ||||
|---|---|---|---|---|---|---|---|
| Admission | Discharge | Admission | Discharge | Intervention | Usual Care | ||
| Beta-blocker | 75 (70.8) | 99 (93.4) | 107 (74.8) | 118 (82.5) | +22.6 | +7.7 | 0.008 |
| ACE inhibitor/ARB | 48 (45.3) | 50 (47.2) | 60 (42.0) | 51 (35.7) | +1.9 | −6.3 | 0.14 |
| ARNI | 22 (20.8) | 26 (24.5) | 5 (3.5) | 12 (8.4) | +3.8 | +4.9 | 0.75 |
| MRA | 33 (31.1) | 52 (49.1) | 22 (15.4) | 30 (21.0) | +17.9 | +5.6 | 0.024 |
| SGLT2i | 26 (24.5) | 34 (32.1) | 13 (9.1) | 22 (15.4) | +7.5 | +6.3 | 0.78 |
Values are n (%) unless otherwise indicated.
Excluded 3 encounters of in-hospital deaths (because discharge guideline-directed medical therapy could not be determined in those individuals).
ACE = angiotensin-converting enzyme; ARB = angiotensin receptor blocker; ARNI = angiotensin receptor–neprilysin inhibitor; MRA = mineralocorticoid receptor antagonist; SGLT2i = sodium-glucose cotransporter-2 inhibitor.
Among encounters allocated to receive the virtual care team guided intervention, the composite GDMT score was 1.1 ± 2.6 compared with 0.4 ± 2.3 among encounters receiving usual care (unadjusted β-coefficient +0.7; 95% CI: +0.1 to +1.4; P = 0.02). After adjusting for the baseline imbalance in number of GDMT between arms, the virtual care team guided intervention remained significant in improving in-hospital GDMT optimization (adjusted β-coefficient: +1.2; 95% CI: +0.7 to +1.8; P < 0.001) (Figure 2).
FIGURE 2. Primary Effectiveness Outcome (GDMT Optimization Score) by Treatment Assignment.
Guideline-directed medical therapy (GDMT) score defined as the sum of care optimization changes (+2 for new GDMT initiations, +1 for dose uptitrations) and care deoptimization changes (−2 for GDMT discontinuation, −1 for dose downtitrations) at hospital discharge. Mean between-group difference in GDMT score was estimated with the use of linear regression models, adjusted for number of GDMT therapies at admission and accounting for patient-level clustering.
The proportion of patients newly initiated on therapy during hospitalization among those not on treatment at admission was numerically higher among those allocated to the virtual care team intervention vs usual care across all therapeutic classes (Figure 3). New initiations among patients not on therapy on admission was significantly higher for evidence-based β-blockers (81% vs 44%; P = 0.005) and MRAs (32% vs 11%; P = 0.001) (Figure 3). The between-group differences in proportion of new initiations for ACE inhibitor/ARB, ARNI, and SGLT2i were not statistically significant. Distribution of GDMT doses as a function of target dosing at admission and discharge is presented in Figure 4.
FIGURE 3. Guideline-Directed Medical Therapy Initiations by Class.
Guideline-directed medical therapy (GDMT) use assessed from hospital admission to discharge. Each bar represents the proportion of encounters with an initiation of therapy among patients without prior prescription of therapy at the time of hospital admission. Any new GDMT initiation during hospitalization was calculated among encounters during which patients were not already treated with quadruple therapy inclusive of an ARNI at admission. ACE = angiotensin-converting enzyme; ARB = angiotensin receptor blocker; ARNI = angiotensin receptor–neprilysin inhibitor; MRA = mineralocorticoid receptor antagonist; SGLT2i = sodium-glucose cotransporter-2 inhibitor.
FIGURE 4. Guideline-Directed Medical Therapy Dosing.
Guideline-directed medical therapy (GDMT) dosing from hospital admission to discharge in (top) the usual care group and (bottom) the virtual care team guided group. Stacked bars represent the distribution of GDMT dosing as a percentage of American College of Cardiology/American Heart Association target dosing before hospital admission and at hospital discharge. ACE = angiotensin-converting enzyme; ARB = angiotensin receptor blocker; ARNI = angiotensin receptor–neprilysin inhibitor; MRA = mineralocorticoid receptor antagonist; SGLT2i = sodium-glucose cotransporter 2-inhibitor.
A higher proportion of clinical encounters had overall net improvement in GDMT during hospitalizations (GDMT score >0) in the virtual care team intervention than in usual care (44% vs 24%; absolute difference: +20%; P = 0.002). This corresponded to a number needed to intervene for a net optimization of GDMT during hospitalization of 5 encounters (Central Illustration). Among encounters not already on 4-drug therapy (inclusive of ARNI), the proportion with at least 1 new initiation was significantly higher in the virtual care team intervention than in usual care (44% vs 23%; absolute difference: +21%; P = 0.001; number needed to intervene of 5 encounters). Similarly, any intensification (new initiation or dose up-titration) of at least 1 GDMT occurred more frequently in the virtual care team intervention compared with usual care (50% vs 28%; absolute difference: +22%; P = 0.001; number needed to intervene of 5 encounters).
CENTRAL ILLUSTRATION. Intensification of Guideline-Directed Medical Therapy by Treatment Arm.
Intensifications of guideline-directed medical therapy (GDMT) from hospital admission to discharge. Bars indicate proportions of encounters with at least 1 new initiation or dose uptitration, with at least 1 new initiation, and in which there was a net intensification of GDMT during hospitalization (GDMT optimization score >0) among encounters allocated to usual care vs the virtual care team guided strategy.
SUBGROUP ANALYSES.
Benefits on the primary endpoint were consistent by age, sex, and race subgroups, and were similar across the 3 sites. The virtual care team strategy was similarly beneficial in improving a GDMT optimization score among encounters with de novo vs established presentations of HFrEF and those encounters primarily for acute HF vs other admission reasons (Figure 5). Significant interactions were observed for ethnicity and language such that the virtual care team guided strategy was less effective among encounters involving predominantly Spanish-speaking patients and those of Hispanic ethnicity.
FIGURE 5. Effects of Virtual Care Team Guided Strategy on Primary Outcome Across Subgroups.
Forest plots summarize the effect of intervention vs usual care on guideline-directed medical therapy (GDMT) optimization scores across key subgroups, analyzed with the use of linear regression adjusted for number of guideline directed medical therapies at admission and accounting for patient-level clustering. Effect sizes are summarized as β-coefficient (95% CI).
SAFETY OUTCOMES.
Overall, among these hospital encounters, 40 (28%) in the usual care group and 23 (21%) in the virtual care team group experienced at least 1 safety event (P = 0.30). Hypotension was the most common safety event, occurring in 24 (17%) encounters allocated to usual care and 12 (11%) allocated to the virtual care team guided intervention (P = 0.28). Most hypotension events did not require ICU transfer or vasopressor use (Table 3). Rates of acute kidney injury, bradycardia, and hyperkalemia were similar in those allocated to usual care and to the virtual care team guided intervention. There were 3 in-hospital deaths during the course of the study, 2 (1%) in those allocated to usual care and 1 (1%) in those allocated to the virtual care team. Site-reported cause of death among usual care encounters included 1 patient with pneumonia and 1 patient with infected lower extremity ulcer complicated by aspiration pneumonia. Cause of death in the patient allocated to the intervention group was metastatic cholangiocarcinoma. The virtual care team made 2 recommendations for this patient, neither of which were adopted by the treating team.
TABLE 3.
Adjudicated Safety Outcomes
| Overall (N = 252) | Virtual Care Team Strategy (n = 107) | Usual Care (n = 145) | P Valuea | |
|---|---|---|---|---|
| At Least 1 safety event | 63 (25.0) | 23 (21.5) | 40 (27.6) | 0.30 |
|
| ||||
| Hypotension | 36 (14.3) | 12 (11.2) | 24 (16.6) | 0.28 |
| 3 consecutive systolic blood pressures <90 mm Hg | 35 (13.9) | 12 (11.2) | 23 (15.9) | 0.36 |
| Vasopressor/intensive care unit use due to hypotension | 9 (3.6) | 2 (1.9) | 7 (4.8) | 0.31 |
|
| ||||
| Hyperkalemia | 26 (10.3) | 8 (7.5) | 18 (12.4) | 0.22 |
| Serum K+ >5.5 mmol/L | 24 (9.5) | 6 (5.6) | 18 (12.4) | 0.08 |
| Serum K+ >6.0 mmol/L | 6 (2.4) | 0 (0.0) | 6 (4.1) | 0.04 |
| Acute potassium-Lowering therapy use | 12 (4.8) | 6 (5.6) | 6 (4.1) | 0.77 |
|
| ||||
| Acute kidney injury | 8 (3.2) | 5 (4.7) | 3 (2.1) | 0.29 |
| Doubling of admission serum creatinine | 6 (2.4) | 5 (4.7) | 1 (0.7) | 0.09 |
| New kidney replacement therapy | 2 (0.8) | 0 (0.0) | 2 (1.4) | 0.51 |
|
| ||||
| Bradycardia | 0 (0.0) | 0 (0.0) | 0 (0.0) | - |
| 3 consecutive heart rates ≤40 beats/min | 0 (0.0) | 0 (0.0) | 0 (0.0) | - |
| New temporary or permanent cardiac pacing use | 0 (0.0) | 0 (0.0) | 0 (0.0) | - |
| Acute heart rate increasing therapy use | 0 (0.0) | 0 (0.0) | 0 (0.0) | - |
|
| ||||
| In-hospital death | 3 (1.2) | 1 (0.9) | 2 (1.4) | 1.00 |
Values are n (%).
Fisher exact test.
DISCUSSION
In this prospective implementation trial, we found that a virtual care team guided strategy improved GDMT during hospitalization for patients with HFrEF across 3 hospitals in an integrated health care delivery system. A virtual care team guided strategy nearly doubled β-blocker prescriptions and nearly tripled MRA prescriptions in patients not previously on these treatments at hospital admission. In addition, the virtual care team guided strategy led to a 20% absolute improvement in net in-hospital optimization, with a number needed to intervene of only 5 clinical encounters for 1 net optimization. Results were consistent across those admitted for acute HF vs other indications and those with established and de novo presentations of HFrEF. A virtual care team guided strategy was safe, with no significant excess in adjudicated serious adverse events and no increase in hospital length of stay.
Our findings identify another important implementation strategy to improve guideline-concordant use in patients with HFrEF. Despite compelling randomized clinical trial evidence supporting use16–20 and codification in guidelines,1,12 well documented gaps in the prescription and target dose achievement of GDMT for HFrEF exist globally.3–6 Reasons for such gaps likely involve patient, provider, and systems-related factors, each contributing to clinical inertia. As such, several interventions have been recently tested aimed at improving GDMT in this population.
Two large multicenter randomized clinical trials demonstrated that neither intensive efforts around audit-and-feedback of quality-of-care measures nor an enhanced intensive suite of transitional care services improved clinical outcomes after HF hospitalization.21,22 Although patients with HFrEF face considerable risk,23 contextualizing 1-year risk estimates to treating clinicians also did not improve clinical outcomes or augment guideline-concordant HF medication use.24 PROMPT-HF (Pragmatic Trial of Messaging to Providers About Treatment of Heart Failure; NCT04514458) demonstrated that pairing risk estimates with tailored EHR-based clinical alerts led to an improvement in GDMT utilization over a 30-day period in outpatients with HFrEF.7 Though highly scalable and fully electronic, effects on GDMT optimization were relatively modest and there might be potential for diminishing returns or alert fatigue with several EHR-based best practice advisory notifications.25,26 EPIC-HF (Electronically Delivered Patient-Activation Tool for Intensification of Medications for Chronic Heart Failure With Reduced Ejection Fraction; NCT03334188) demonstrated effectiveness of an educational video paired with a 1-page medication checklist delivered to patients ahead of a cardiovascular visit.27 This trial demonstrated the feasibility of direct patient activation; given the outpatient clinic–based approach undertaken, this strategy may miss opportunities to optimize care during the vulnerable period surrounding hospitalization. The STRONG-HF (Safety, Tolerability, and Efficacy of Rapid Optimization, Helped by NT-proBNP and GDF-15, of Heart Failure Therapies; NCT03412201) trial demonstrated that a high-intensity care strategy led to GDMT improvement and lower clinical events among recently hospitalized patients with HF.8,9 Importantly, the protocol called for regular visits at 1, 2, 3, and 6 weeks and additional safety visits 1 week after dose up-titration; this intensity of care may not be feasible in settings outside of a clinical trial. In this context, use of a fully virtual model in IMPLEMENT-HF may compliment (and be paired with) the prior successful implementation approaches to improve the reach, scale, and effect on the ultimate aim of facilitating, improving, and sustaining optimal guideline concordant care for HFrEF.
There are several distinctive features of this design that should be noted. First, we demonstrated effectiveness across encounters for acute HF and admissions for other primary admission indications. Secondary hospitalizations in patients with HF are highly prevalent,28 and some of these hospitalizations may even meet consensus criteria for worsening HF.29,30 Second, this study included patients with newly diagnosed HFrEF during index hospitalization. While expert opinion has supported rapid initiation of multiple GDMT elements during de novo HF admissions,31 many efficacy and implementation trials have been conducted in outpatient settings or have excluded de novo presentations of HFrEF.7,24,27 In IMPLEMENT-HF, nearly 1 in 5 patients had de novo presentations of HF, and the virtual care team guided strategy had similar effectiveness across these encounters and those with established HFrEF. Third, the virtual care team guided strategy with a central physician team influenced care favorably across multiple hospitals in an integrated health care delivery system. An earlier pilot study demonstrated the effectiveness of a peer-to-peer consultation strategy for GDMT optimization in a single large academic medical center.32 Increasing consolidation in health care has contributed to the growth of large complex health care delivery systems that include hospitals of varying size, resources, and patient populations. These hospitals often have shared processes, technologies (eg, electronic medical records), and workflows. Developing a fully virtual platform or team may allow for increasing reach, scale, and cost-effectiveness. Strategies including the virtual team used in IMPLEMENT-HF hold promise for the efficient implementation of high-quality care strategies across other similarly organized integrated health systems.
The predominant impact of the virtual care team guided strategy was to improve the rates of foundational treatments for HF, namely β-blocker and MRA, mostly at low doses with lesser impact on optimization of newer agents including ARNI and SGLT2i. Reasons for this include the structure of the IMPLEMENT-HF algorithm, which prioritized establishing multicomponent GDMT even at low doses before dose up-titration, a strategy that is consistent with current guidelines. Furthermore, lack of familiarity with newer therapies in addition to perceived or actual cost or access barriers may have limited the adoption of recommendations for ARNI and SGLT2i. Finally, these therapies were suggested after foundational therapies generally suggested and implemented, possibly further accounting for the observed differences in optimization strategy by GDMT class. As such, this in-hospital intervention may serve as an initial step in establishing multicomponent GDMT regimens safely; pairing this intervention with efficient strategies in the transitional and early outpatient periods may further serve to accelerate incorporation of newer GDMT elements, encourage dose optimization, and promote treatment persistence.33
Importantly, we observed significant interaction in the effectiveness of the intervention, such that patients who were predominantly Spanish speaking and of Hispanic ethnicity derived less benefit from the virtual care team guided intervention. Notably, these findings of patient level interaction are intriguing given the provider-directed nature of the intervention. Reasons for these findings are unclear but may include the bias or inertial barriers associated with counseling predominantly non-English-speaking patients on multiple medication changes. While virtual care team notes included guidance for patients on necessary follow-up to be included in patient instructions in the discharge summary, alternative language translations of this text were not provided for non-English-speaking patients. Finally, additional socioeconomic factors may have contributed. These findings highlight the importance for prospective evaluation of inequities in implementation interventions, and the potential need for enhanced support among predominantly non-English-speaking patients in future studies.
STUDY LIMITATIONS.
Although this trial included hospital sites of varying size and patient populations, all sites were part of an integrated health system. Reproducibility in other hospital types, geographies, and populations requires additional study. In addition, larger trials with longer-term follow-ups are needed to evaluate the durability of GDMT optimization and association with clinical outcomes. Allocation in this study was based on hospital encounters, which might have led to contamination (because the intervention was clinician directed) as an individual clinician might have cared for individuals in both study arms. However, treating teams at the study hospitals included many clinicians, including house staff, advanced practice providers, and attending physicians, each with varying schedules, making clustering at the provider level challenging. Such contamination would have been expected to bias the result toward the null. Finally, although allocation to each strategy arm was intended to be random (as a virtue of patient birth month), the number of encounters in each group was not 1:1, and some baseline imbalances were identified such that GDMT use was higher at admission in the intervention group. We expect that these differences are most likely due to random chance; other key demographic and clinical characteristics appeared balanced between treatment groups. Assessed effectiveness outcomes included change from admission to discharge, which should account for baseline imbalances, and the primary endpoint analysis was further adjusted for baseline number of HF therapies. Despite fewer opportunities for care optimization owing to higher baseline HF therapy use in the intervention group, the virtual care team guided strategy boosted in-hospital use and dosing of GDMT compared with usual care.
CONCLUSIONS
A virtual care team guided strategy safely improved in-hospital HFrEF GDMT optimization across multiple hospitals in an integrated health care system without increasing hospital length of stay. This strategy represents a potential highly effective, scalable intervention that can lead to accelerated implementation of guideline concordant HFrEF care.
Supplementary Material
PERSPECTIVES.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS:
A virtual care team guided strategy improves in-hospital medical therapy for patients with heart failure in an integrated health care system without increasing length of stay.
TRANSLATIONAL OUTLOOK:
Enhanced, appropriate support for non-English speaking patients is an important area of focus for future implementation initiatives.
ACKNOWLEDGMENTS
The authors thank the Brigham Health Care Redesign Incubator and Startup Program (BCRISP) staff and mentors.
FUNDING SUPPORT AND AUTHOR DISCLOSURES
Funding was provided by the Brigham Health Care Redesign Incubator and Startup Program, Brigham and Women’s Hospital, Mass General Brigham, Boston, Massachusetts. Dr Varshney has received consulting fees from Broadview Ventures. Dr Claggett has received consulting fees from Cardurion, Corvia, Cytokinetics, Intellia, and Novartis. Dr Eaton is employed at Brigham and Women’s hospital, but is also employed by Janssen Pharmaceuticals. Dr Cunningham has received consulting fees from Roche Diagnostics and Occlutech. Dr Choudhry has received research grant support to Brigham and Women‘s Hospital from Merck, Sanofi, AstraZeneca, CVS Health, and Medisafe. Dr Solomon has received research grants from Actelion, Alnylam, Amgen, AstraZeneca, Bellerophon, Bayer, Bristol Myers Squibb, Celladon, Cytokinetics, Eidos, Gilead, GlaxoSmithKline, Ionis, Lilly, Mesoblast, MyoKardia, National Institutes of Health/National Heart, Lung, and Blood Institute, Neurotronik, Novartis, NovoNordisk, Respicardia, Sanofi Pasteur, Theracos, and US2.AI; and has consulted for Abbott, Action, Akros, Alnylam, Amgen, Arena, AstraZeneca, Bayer, Boehringer Ingelheim, Bristol Myers Squibb, Cardior, Cardurion, Corvia, Cytokinetics, Daiichi Sankyo, GlaxoSmithKline, Lilly, Merck, Myokardia, Novartis, Roche, Theracos, Quantum Genomics, Cardurion, Janssen, Cardiac Dimensions, Tenaya, Sanofi-Pasteur, Dinaqor, Tremeau, CellProThera, Moderna, American Regent, and Sarepta. Dr Vaduganathan has received research grant support from, served on advisory boards for, or had speaker engagements with American Regent, Amgen, AstraZeneca, Bayer, Baxter Healthcare, Boehringer Ingelheim, Chiesi, Cytokinetics, Lexicon Pharmaceuticals, Novartis, Novo Nordisk, Pharmacosmos, Relypsa, Roche Diagnostics, Sanofi, and Tricog Health; and participates on clinical trial committees for studies sponsored by AstraZeneca, Galmed, Novartis, Bayer, Occlutech, and Impulse Dynamics. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
ABBREVIATIONS AND ACRONYMS
- ACE
angiotensin-converting enzyme
- ARB
angiotensin receptor blocker
- ARNI
angiotensin receptor-neprilysin inhibitor
- EHR
electronic health record
- GDMT
guideline-directed medical therapy
- HFrEF
heart failure with reduced ejection fraction
- LVEF
left ventricular ejection fraction
- MRA
mineralocorticoid receptor antagonist
- SGLT2i
sodium-glucose cotransporter-2 inhibitor
Footnotes
APPENDIX For supplemental material, please see the online version of this paper.
REFERENCES
- 1.Heidenreich PA, Bozkurt B, Aguilar D, et al. 2022 AHA/ACC/HFSA guideline for the management of heart failure: a report of the ACC/AHA joint committee on clinical practice guidelines. J Am Coll Cardiol. 2022;79(17):e263–421. [DOI] [PubMed] [Google Scholar]
- 2.Vaduganathan M, Claggett BL, Jhund PS, et al. Estimating lifetime benefits of comprehensive disease-modifying pharmacological therapies in patients with heart failure with reduced ejection fraction: a comparative analysis of three randomised controlled trials. Lancet. 2020;396:121–128. [DOI] [PubMed] [Google Scholar]
- 3.Greene SJ, Butler J, Albert NM, et al. Medical therapy for heart failure with reduced ejection fraction: the CHAMP-HF registry. J Am Coll Cardiol. 2018;72:351–366. [DOI] [PubMed] [Google Scholar]
- 4.Greene SJ, Fonarow GC, DeVore AD, et al. Titration of medical therapy for heart failure with reduced ejection fraction. J Am Coll Cardiol. 2019;73:2365–2383. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Savarese G, Kishi T, Vardeny O, et al. Heart failure drug treatment-inertia, titration, and discontinuation: a multinational observational study (EVOLUTION HF). J Am Coll Cardiol HF. 2023;11:1–14. [DOI] [PubMed] [Google Scholar]
- 6.Tromp J, Ouwerkerk W, Teng T-HK, et al. Global disparities in prescription of guideline-recommended drugs for heart failure with reduced ejection fraction. Eur Heart J. 2022;43:2224–2234. [DOI] [PubMed] [Google Scholar]
- 7.Ghazi L, Yamamoto Y, Riello RJ, et al. Electronic alerts to improve heart failure therapy in outpatient practice: a cluster randomized trial. J Am Coll Cardiol. 2022;79:2203–2213. [DOI] [PubMed] [Google Scholar]
- 8.Kimmoun A, Cotter G, Davison B, et al. Safety, Tolerability and Efficacy of Rapid Optimization, Helped by NT-proBNP and GDF-15, of Heart Failure Therapies (STRONG-HF): rationale and design for a multicentre, randomized, parallel-group study. Eur J Heart Fail. 2019;21:1459–1467. [DOI] [PubMed] [Google Scholar]
- 9.Mebazaa A, Davison B, Chioncel O, et al. Safety, tolerability and efficacy of up-titration of guideline-directed medical therapies for acute heart failure (STRONG-HF): a multinational, open-label, randomised, trial. Lancet. 2022;400:1938–1952. [DOI] [PubMed] [Google Scholar]
- 10.Varshney AS, Bhatt AS, Vaduganathan M. Hospitalization of patients with (but not for) heart failure: an opportunity for accelerated guideline-directed medical therapy optimization? J Card Fail. 2021;27:910–912. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bhatt AS, Varshney AS, Nekoui M, et al. Virtual optimization of guideline-directed medical therapy in hospitalized patients with heart failure with reduced ejection fraction: the IMPLEMENT-HF pilot study. Eur J Heart Fail. 2021;23:1191–1201. [DOI] [PubMed] [Google Scholar]
- 12.Yancy CW, Jessup M, Bozkurt B, et al. 2017 ACC/AHA/HFSA focused update of the 2013 ACCF/AHA guideline for the management of heart failure: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Failure Society of America. J Am Coll Cardiol. 2017;70(6):776–803. [DOI] [PubMed] [Google Scholar]
- 13.Yancy CW, Januzzi JL, Allen LA, et al. 2017 ACC expert consensus decision pathway for optimization of heart failure treatment: answers to 10 pivotal issues about heart failure with reduced ejection fraction: a report of the American College of Cardiology Task Force on Expert Consensus Decision Pathways. J Am Coll Cardiol. 2018;71:201–230. [DOI] [PubMed] [Google Scholar]
- 14.Maddox TM, Januzzi JL, Allen LA, et al. 2021 update to the 2017 ACC expert consensus decision pathway for optimization of heart failure treatment: answers to 10 pivotal issues about heart failure with reduced ejection fraction: a report of the American College of Cardiology Solution Set Oversight Committee. J Am Coll Cardiol. 2021;77:772–810. [DOI] [PubMed] [Google Scholar]
- 15.Bristow MR, Gilbert EM, Abraham WT, et al. , MOCHA Investigators. Carvedilol produces dose-related improvements in left ventricular function and survival in subjects with chronic heart failure. Circulation. 1996;94:2807–2816. [DOI] [PubMed] [Google Scholar]
- 16.Effect of metoprolol CR/XL in chronic heart failure: Metoprolol CR/XL Randomised Intervention Trial in Congestive Heart Failure (MERIT-HF). Lancet. 1999;353:2001–2007. [PubMed] [Google Scholar]
- 17.Pitt B, Zannad F, Remme WJ, et al. , Randomized Aldactone Evaluation Study Investigators. The effect of spironolactone on morbidity and mortality in patients with severe heart failure. N Engl J Med. 1999;341:709–717. [DOI] [PubMed] [Google Scholar]
- 18.Zannad F, McMurray JJV, Krum H, et al. Eplerenone in patients with systolic heart failure and mild symptoms. N Engl J Med. 2011;364:11–21. [DOI] [PubMed] [Google Scholar]
- 19.McMurray JJV, Packer M, Desai AS, et al. Angiotensin-neprilysin inhibition versus enalapril in heart failure. N Engl J Med. 2014;371:993–1004. [DOI] [PubMed] [Google Scholar]
- 20.McMurray JJV, Solomon SD, Inzucchi SE, et al. Dapagliflozin in patients with heart failure and reduced ejection fraction. N Engl J Med. 2019;381:1995–2008. [DOI] [PubMed] [Google Scholar]
- 21.DeVore AD, Granger BB, Fonarow GC, et al. Care optimization through patient and hospital engagement clinical trial for heart failure: rationale and design of CONNECT-HF. Am Heart J. 2020;220:41–50. [DOI] [PubMed] [Google Scholar]
- 22.van Spall HGC, Lee SF, Xie F, et al. Effect of patient-centered transitional care services on clinical outcomes in patients hospitalized for heart failure: the PACT-HF randomized clinical trial. JAMA. 2019;321:753–761. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Greene SJ, Butler J, Fonarow GC. Contextualizing risk among patients with heart failure. JAMA. 2021;326:2261–2262. [DOI] [PubMed] [Google Scholar]
- 24.Ahmad T, Desai NR, Yamamoto Y, et al. Alerting clinicians to 1-year mortality risk in patients hospitalized with heart failure: the REVEALHF randomized clinical trial. JAMA Cardiol. 2022;7:905–912. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Chaparro JD, Hussain C, Lee JA, Hehmeyer J, Nguyen M, Hoffman J. Reducing interruptive alert burden using quality improvement methodology. Appl Clin Inform. 2020;11:46–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Kawamanto K, Flynn MC, Kukhareva P, et al. A pragmatic guide to establishing clinical decision support governance and addressing decision support fatigue: a case study. AMIA Annu Symp Proc. 2018;2018:624–633. [PMC free article] [PubMed] [Google Scholar]
- 27.Allen LA, Venechuk G, McIlvennan CK, et al. An electronically delivered patient-activation tool for intensification of medications for chronic heart failure with reduced ejection fraction: the EPIC-HF trial. Circulation. 2021;143:427–437. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Varshney AS, Minhas AMK, Bhatt AS, Ambrosy AP, Fudim M, Vaduganathan M. Contemporary burden of primary versus secondary heart failure hospitalizations in the united states. Am J Cardiol. 2021;156:140–142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Ambrosy AP, Parikh RV, Sung SH, et al. A natural Language processing-based approach for identifying hospitalizations for worsening heart faiLure within an integrated health care delivery system. JAMA Netw Open. 2021;4:e2135152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Ambrosy AP, Parikh RV, Sung SH, et al. Analysis of worsening heart failure events in an integrated health care system. J Am Coll Cardiol. 2022;80:111–122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Greene SJ, Butler J, Fonarow GC. Simultaneous or rapid sequence initiation of quadruple medical therapy for heart failure-optimizing therapy with the need for speed. JAMA Cardiol. 2021;6:743–744. [DOI] [PubMed] [Google Scholar]
- 32.Rao VN, Shah A, McDermott J, et al. In-hospital virtual peer-to-peer consultation to increase guideline-directed medical therapy for heart failure: a pilot randomized trial. Circ Heart Fail. 2022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Desai AS, Maclean T, Blood AJ, et al. Remote optimization of guideline-directed medical therapy in patients with heart failure with reduced ejection fraction. JAMA Cardiol. 2020;5:1430–1434. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.