HEART Camp Connect—Promoting adherence to exercise in adults with heart failure with preserved ejection fraction

Abstract

Aims

Most adults with stable heart failure are safe to exercise at a moderate intensity for 150 min/week. Regular participation in exercise may improve outcomes in adults with heart failure with preserved ejection fraction (HFpEF). Few adults with HFpEF initiate and sustain long‐term exercise. To promote exercise adherence in adults with HFpEF, we developed the Heart Failure Exercise and Resistance Training (HEART) Camp Connect intervention that is tested in this clinical trial. This trial tests our central hypothesis that theory‐informed coaching strategies delivered virtually will promote long‐term adherence to exercise in adults with HFpEF and drive clinically meaningful, and cost‐effective improvements in physiological and patient‐reported outcomes. Our aims are to (a) evaluate the effects of virtual and in‐person exercise and coaching on long‐term adherence, (b) determine a benchmark of minutes of moderate intensity exercise associated with health status as related to key biobehavioural outcomes, (c) examine behaviour change theory‐defined constructs as mediators of exercise adherence and (d) evaluate intervention costs.

Methods

This 18 month, three‐group, repeated measures randomized controlled trial is enrolling 300 adults with HFpEF. Participants are randomized to enhanced usual care (EUC), virtual coaching, or in‐person coaching. Our intervention applies coaching strategies, informed by behaviour change theories, in one‐on‐one and group settings weekly for 12 months. Our objective is to compare the effects of each delivery method to the other and EUC on exercise adherence (defined as ≥ 120 min of moderate intensity exercise/week) at 12 months (primary endpoint) and 18 months (sustainability endpoint). Secondary outcomes include minutes of moderate intensity exercise needed to drive minimal clinically important differences in health status, biomarkers, patient‐reported symptoms and cost. Behaviour change theory‐defined constructs (e.g., self‐efficacy and outcome expectations) will be tested as mediators of exercise adherence.

Results

We expect that virtual coaching is equally as efficacious and more cost effective at promoting exercise adherence as in‐person coaching. Effects on exercise adherence may be mediated by theory‐defined constructs. We also expect to identify a threshold for minutes of moderate intensity exercise to potentially serve as an adherence benchmark in adults with HFpEF, one that may differ from the 120 min of exercise in our current definition.

Conclusions

These findings could shift the paradigm of exercise coaching in HF towards virtual delivery and increase the generalizability and reach of exercise training. This is especially important for adults with HFpEF as they are excluded from Medicare reimbursement for traditional cardiopulmonary rehabilitation.

Introduction

Exercise training is recommended as a non‐pharmacological therapy that is safe and effective for individuals with heart failure (HF) (Class I—level of Evidence A). ¹ , ² Exercise trials have demonstrated positive outcomes in adults with HF, including improved exercise capacity (as measured by oxygen uptake, peak VO₂), physical function and quality of life (QOL). ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ However, as many as 91% of adults with HF do not participate in regular exercise and exercise adherence remains a major limitation. ¹⁰ , ¹¹ , ¹² Exercise trials report modest adherence, particularly long term, beyond 6 months. ⁹ , ¹³ , ¹⁴

Behavioural interventions improve long‐term adherence to moderate intensity exercise at 12 and 18 months compared with usual care in adults with stable HF. ¹⁵ Trials that included both adults with HF with preserved ejection fraction (HFpEF) and HF with reduced ejection fraction (HFrEF) demonstrated particularly favourable effects in those with HFpEF, suggesting that individuals with HFpEF may benefit more from exercise than those with HFrEF. ¹⁶ , ¹⁷ Despite a clear need, specific programmes to teach adults with HFpEF how to initiate and adhere to exercise behaviour over time do not exist. Centre‐based cardiac rehabilitation is the predominate model for initiating exercise in adults with HF, but current Medicare guidelines do not reimburse rehabilitation services for adults with HFpEF and exercise is not often sustained long‐term after supervised cardiac rehabilitation is completed. ¹⁴ , ¹⁸ , ¹⁹ , ²⁰ New and refined interventions to promote exercise adherence in adults with HFpEF are needed and are a high priority of the National Heart Lung and Blood Institute of the National Institutes of Health. ²¹

In response to this critical need, our team proposed the Heart Failure Exercise and Resistance Training (HEART) Camp Connect clinical trial. Our central hypothesis is that theory‐informed coaching strategies delivered virtually will promote long‐term adherence to exercise in adults with HFpEF and drive clinically meaningful and cost‐effective improvements in physiological and patient‐reported outcomes. We test this central hypothesis across specific aims focused on (a) evaluating the effects of virtual and in‐person exercise and coaching on long‐term adherence, (b) determining a benchmark of minutes of moderate intensity exercise associated with health status as related to key biobehavioural outcomes, (c) examining behaviour change theory‐defined constructs as mediators of exercise adherence and (d) evaluating intervention costs and benefits.

Study design

This prospective clinical trial uses a three‐group, 2–2–1 randomized repeated measures experimental design to compare the effect of two intervention arms to each other and versus enhanced usual care (EUC) in achieving long‐term [12 months (primary endpoint) and 18 months (sustainability endpoint)] adherence to moderate intensity exercise in adults with HFpEF. Moderate intensity is individually defined for each participant as exercise that reaches 40%–80% of heart rate reserve (HRR), using the peak‐rest method. The American College of Sports Medicine recommends adults with HF be started at 40%–50% HRR with progression to 60%–80% HRR as tolerated. ²² The adherence target for our sample of currently physically inactive adults with HFpEF is ≥120 min of moderate intensity exercise per week. This study is approved by a central Institutional Review Board and is registered on ClinicalTrials.gov (NCT05784753) and complies with the Declaration of Helsinki.

Sample, setting and recruitment

This study will enrol 300 adults with HFpEF from academic medical centres in Omaha, Nebraska and Detroit, Michigan. Inclusion/exclusion criteria are shown in Table 1. Participants are included if they have invasive haemodynamic evidence suggestive of HFpEF defined as resting pulmonary capillary wedge pressure (PCWP) ≥15 mmHg or exercise PCWP ≥ 25 mmHg measured during right heart catheterization, ²³ , ²⁴ a score ≥6 on the H₂FPEF algorithm, or a clinical diagnosis of HFpEF defined as the presence of common medical comorbidities or risk factors known to contribute to HFpEF along with classical signs/symptoms of HFpEF and echocardiographic abnormalities consistent with diastolic dysfunction as validated/assessed by a clinical expert. The H₂FPEF algorithm uses clinical characteristics, demographics and data derived from echocardiography collected from the electronic health record to compute a total score. Specifically, body mass index >30 kg/m ²⁵ earns 2 points; hypertension, defined as ≥2 antihypertensive medications, earns 1 point; atrial fibrillation, defined as presence of paroxysmal or persistent earns 3 points; age >60 years earns 1 point; and from echocardiography, pulmonary artery systolic pressure >35 mmHg earns 1 point; and filling pressure (E/e′) >9 earns 1 point. Individuals with scores ≥ 6 are demonstrated to have a >90% probability of having HFpEF. ²⁵ , ²⁶ , ²⁷

Table 1.

Inclusion/exclusion criteria.

Inclusion criteria

Exclusion criteria

1. Diagnosis of HF with an EF ≥ 50% 2. Invasive haemodynamic evidence of HFpEF, score > 6 on H2FPEF algorithm or confirmed clinical diagnosis of HFpEF 3. Age of majority in state of recruitment a 4. Echocardiogram in prior 24 months 5. Stable pharmacologic therapy in past 30 days

Life‐limiting illness precluding completion of 18 month study Diagnosis of Alzheimer's or other dementia Clinical evidence of decompensated HF Unstable angina or marked shortness of breath on exertion at <2 metabolic equivalents Myocardial infarction, coronary artery bypass surgery or biventricular pacemaker <6 weeks Orthopaedic or neuromuscular disorders preventing participation in aerobic exercise Cardiopulmonary exercise test results that preclude safe exercise Unwilling/unable to complete run‐in period Pregnancy

Abbreviations: EF, ejection fraction; HF, heart failure; HFpEF, HF with preserved EF.

^a Age of the majority in Nebraska, USA, is 19 years old.

Participants are recruited using a dual recruitment approach through relationships with our HF, gerontology and primary care clinical partners in Omaha and Detroit and an existing database of individuals willing to be contacted for participation in clinical research. Potential participants are approached for consent by research personnel via telephone or email if they have opted to receive research communication.

Run‐in period

As shown in the study overview (Figure 1), all participants complete a run‐in period prior to randomization. The purpose of the run‐in period is to ensure safety and adequate orientation to exercise at a moderate intensity. The run‐in period includes a cardiopulmonary exercise test (CPET), three sessions in early outpatient (i.e., Phase II) cardiopulmonary rehabilitation, investigator‐developed educational sessions for HFpEF and orientation to the study. CPET is used to determine suitability for exercise (no significant ischemia or arrhythmias) and the individualized exercise prescriptions. CPET results are transmitted to our CPET Research Core laboratory for determination of the exercise training prescription using the %HRR method. ²⁸ Percentage HRR is a practical measurement for determining exercise intensity, previously demonstrated to have a linear relationship to %VO₂ reserve. Additionally, %HRR is more readily available for community settings that are without access to the equipment and trained staff needed to determine gas exchange during exercise. ²⁹ , ³⁰ Applying %HRR to a participant's CPET results tailors the physiologic HR response for each participant as is recommended for anyone with HF, irrespective of taking medications that may attenuate HR response to exercise. ⁵ Exercise prescriptions are provided to cardiopulmonary rehabilitation staff to guide monitored sessions. The supervised sessions in cardiopulmonary rehabilitation and the use of moderate intensity aerobic (40%–80% HRR) ¹⁵ , ²⁸ and resistance (10–15 repetitions to volitional fatigue) ³¹ , ³² training exercises combine to help ensure the safety of all participants prior to transitioning to in‐person or virtual exercise. ²⁸ , ³³ , ³⁴

Figure 1.

HEART Camp Connect study overview. This figure summarizes the run‐in period and subsequent randomization to one of three groups.

All participants are given a wrist‐worn, Polar Pacer (PolarElectro, Kempele, Finland) exercise watch and forearm‐worn VeritySense (PolarElectro,Kempele, Finland) HR monitor that allows real‐time cloud‐based capture of data for our primary outcome and allows participants to self‐monitor HR during exercise. Polar watches and heart rate sensors are validated in adults with cardiovascular disease. ³⁵ We have successfully used Polar heart rate sensors and watches in our prior studies. ³³ Participants are also instructed to maintain a rating of perceived exertion of 12–14 on the Borg 6 to 20 rating of perceived exertion scale. ³⁶ Participants track their exercise sessions by initiating a session on their watch and heart rate monitor when they begin exercising. This approach familiarizes them with monitoring heart rate during exercise. For participants that cannot reliably use heart rate to self‐monitor exercise intensity due to beta‐blocker use, pacemakers or atrial fibrillation, the Borg rating of perceived exertion is used to self‐monitor. In addition, all participants complete daily exercise diaries that are used to reconcile heart rate data and record the rating of perceived exertion for each exercise session. These study procedures also reduce the risk of misattributing arrhythmia related tachycardia or artefact as exercise training. Orientation includes hands‐on training with all study procedures including completion of daily exercise diaries and videoconference demonstration for those randomized to virtual coaching.

Fitness centre memberships

Access to a fitness centre or fitness classes is reported as a barrier to exercise adherence, particularly in underserved populations. ³⁷ To reduce the potential of access to confound our adherence findings, we provide paid access to a fitness centre, either virtually or in‐person, for all participants for the 18 month study period.

Randomization and group assignments

Participants are randomized to one of two interventions or EUC using a 2–2–1 allocation schedule created by our statistician using block permutation with mixed block size. The sample will be stratified by biological sex and site to ensure equal female and male representation across conditions. Participants are randomized to (a) HEART Camp, an in‐person coaching intervention (n = 120), (b) HEART Camp Connect, a virtual coaching intervention (n = 120), or EUC (n = 60). Each of these groups is described below and summarized in Figure 2 above.

Figure 2.

HEART Camp Connect coaching workflow. This figure demonstrates a timeline and workflow of coaching for HEART Camp and HEART Camp Connect groups.

EUC

Participants randomized to EUC receive virtual fitness centre memberships. Virtual memberships provide access to a virtual exercise platform managed by the medical fitness centre at our primary site for 18 months. The platform includes exercise sessions that are viewed on demand from any internet‐capable device.

Intervention groups

Participants in the HEART Camp group are coached in person and have access to a brick‐and‐mortar medical fitness centre. Participants in the HEART Camp Connect group are coached via videoconference, have virtual fitness centre memberships only and do not access the brick‐and‐mortar fitness centre. Participants in HEART Camp Connect also receive weekly motivational text messages on Thursday afternoons if they have not reached the 120 min adherence target by that time. These messages are sent via an automated process incorporated into our study's web interface. The messaging is designed to incorporate individualized exercise data along with a message to encourage exercise. For example, a participant that has not achieved 120 min of exercise by Thursday afternoon will receive a message saying ‘You are [x] minutes away from your weekly exercise goal. Keep up the great work!’ Exercise coaches follow the same schedule for both intervention groups with structured activities that operationalize behaviour change theory‐defined constructs conceptualized from social cognitive theory and the theory of future‐oriented motivation and self‐regulation proposed by Miller and Brickman. ³⁸ , ³⁹ , ⁴⁰ Coaches emphasize strategies found to favourably mediate exercise adherence in our prior work including self‐efficacy, attitudes and relapse management. ⁴¹

Figure 2 shows a flow diagram of our coaching which is broken into initiation (Months 0–3), assimilation (Months 4–12) and sustainability (Months 13–18) phases. Participants receive the most coach support during initiation. Coaches and participants review heart rate data and exercise diaries, discuss symptoms experienced during exercise, dyspnoea management techniques, goal setting, barriers to exercise and barrier mitigation strategies. Importantly, coaches assist participants in developing an individualized plan for return to exercise if relapse occurs. During assimilation, individualized sessions are reduced to 15 min every other week, and group sessions are added on alternate weeks. During Months 13–18 (sustainability), participants continue to sync heart rate data with our web application and submit exercise diaries.

Intervention fidelity monitoring

The design and implementation of this study incorporates specific strategies to address intervention fidelity as recommended by the National Institutes of Health (NIH) Behavior Change Consortium Treatment Fidelity Workgroup. ⁴² The intervention fidelity plan, revised from our previous work, addresses the training of those interacting with the participants, the delivery of the intervention and the receipt and enactment of the intervention by the participants. ⁴³

Outcomes

Table 2 presents study variables, instruments and data collection times. The primary outcome of this study is adherence to moderate intensity exercise at the primary study endpoint of 12 months and a sustainability endpoint at 18 months. Adherence is defined as achieving ≥120 min of moderate intensity exercise (40%–80% of HRR) per week. Definitions and measures of exercise adherence are inconsistent across HF exercise trials. Session attendance is often used as a surrogate adherence measure rather than objectively measured minutes of exercise at a prescribed frequency, intensity, time, type, volume and progression. ⁷ , ¹⁴ , ²² This surrogate measurement of adherence limits examination of adherence‐related dose‐response effects or comparison with other clinical outcomes.

Table 2.

Outcome measures and data collection times.

Variable	Instrument	Collection time				Data source
		Baseline	6 M	12 M	18 M
Primary outcome
Adherence	Heart rate sensor	X	X a	X a	X a	Investigator‐developed, web‐based application
Secondary outcomes
HF‐related health status	Kansas City Cardiomyopathy Questionnaire (KCCQ‐23) 44 , 45 , 46	X	X	X	X	Online survey or interview
Dyspnoea severity	PROMIS—Dyspnea Severity–10a 47	X	X	X	X
Global symptoms	PROMIS‐29 Profile v2.1 48 , 49	X	X	X	X
Quality of life	EQ‐5D 50 , 51	X	X	X	X
Physical function	6 min walk test (6MWT) 52	X	X	X	X	EHR or study visit
Physical activity	Accelerometer	X	X	X	X	Study visit
Inflammatory markers	Mesoscale discovery V‐plex human cytokine panel	X	X	X	X	Plasma
Interim clinical events a	Hospitalizations, ED visits, injuries and unplanned procedures	X a	X a	X a	X a	EHR or interview
Behaviour change theory‐defined constructs
Self‐efficacy	Barriers Self‐Efficacy Scale 53	X	X	X	X	Online survey or interview
Outcome expectations	Multi‐dimensional outcome expectation for exercise 54	X	X	X	X	Online survey or interview
Attitudes	Attitudes towards physical activity/exercise 55	X	X	X	X	Online survey or interview
Self‐regulation	Physical Activity Self‐Regulation Scale 56	X	X	X	X	Online survey or interview
Social support	PROMIS—Social Support 49	X	X	X	X	Online survey or interview
Participant perceptions	Qualitative interview			X		Semi‐structured interview

Abbreviations: CKD, chronic kidney disease; DM, diabetes; EHR, electronic health record; EQ‐5D, European Quality of Life, 5‐Dimension; HTN, hypertension; MCID, minimal clinically important difference; PAD, peripheral vascular disease; PROMIS‐29, Patient‐Reported Outcomes Measurement Information System; RA/CT, rheumatoid arthritis/connective tissue.

^a Interim Clinical Events are collected on an ongoing basis for 18 months. For all events, we record relationship to exercise, days of provider‐recommended withdrawal from exercise and/or days of elective withdrawal from exercise. In addition, we collect the following for each type of event—hospitalization: admitting diagnosis, length of stay, discharge disposition; emergency visits: chief complaint, intervention, discharge disposition; injury: type and location of injury, type of care sought, if any.

Secondary outcomes include exercise adherence at 6 months, clinical outcomes including 6 min walk test distance and physical activity, and patient‐reported outcomes including HF‐related health status, QOL, dyspnoea severity, physical function, anxiety, depression, fatigue, sleep disturbance, social roles and pain. We will also measure inflammatory biomarkers and measures of behaviour change theory‐defined constructs (e.g., self‐efficacy, outcome expectations; shown in Table 2) at baseline, 6, 12 and 18 months. We collect demographic information (e.g., age, biological sex, gender identity and race) at baseline and prognostic indicators of HFpEF (e.g., comorbidities, current medications\ and supplements) at all timepoints from the medical record with participant confirmation. Interim clinical events including hospitalizations, emergency department visits, injuries and unplanned procedures are collected on an ongoing basis throughout the 18 month study period. The 23‐item Kansas City Cardiomyopathy Questionnaire (KCCQ) assesses patient‐reported health status and importantly has a well‐established minimal clinically important difference of 5 points change over time. ⁵⁷ All measures are reliable and valid with psychometric properties reported in the cited references and previously used by our team. The collection of questionnaires is managed by trained, blinded graduate assistants via REDCap (Research Electronic Data Capture) survey sent by email link or in interview format with direct entry to the study REDCap database, according to participant preference.

Incentives

At the end of the study participants keep the exercise watch and heart rate monitor (approximate value of $250). Additional incentives of $20 are provided for each complete data collection. Participants who complete all data collections are given an extra $20 incentive at 18 months.

Sample size estimates

The sample size for this study (n = 300) was determined based on the primary outcome, testing group differences in adherence to exercise over 12 months (intervention endpoint). The power analysis assumes the proportion of participants adherent to exercise (achieving ≥120 min of moderate‐intensity exercise per week) of 42% and 14% drop‐out rate in the intervention groups, compared with 14% adherence and 12% dropout rate in the EUC group. ¹⁶ Using Bonferroni correction for the two comparisons of the intervention groups to the EUC, a total sample size of 285 (114:114:57) can achieve 90% power using proportion test at significance level 0.05, accounting for attrition. For the non‐differential comparison between the two treatment groups, if there is truly no difference between the standard and virtual treatment, then with 90% probability, the upper limit of a one‐sided 95% confidence interval will exclude a difference in favour of the standard group of more than 21%. We plan to enrol 300 participants (120:120:60) to account for a higher‐than‐expected rate of attrition or if the effect size is lower than anticipated. For the secondary outcome associated with a minimal clinically important difference in HF‐related health status, we performed a simulation‐based power analysis assuming the true minimum adherence level is 120 min, and it shows that there is 85% chance to obtain the estimated change point within ±20 min. For mediation analyses, previous literature shows that an analytic sample size of 200 can have >80% power to detect at least small to moderate effect size between intervention and mediator and at least small to moderate effect size between mediator and outcome.

Statistical analysis

For all analysis, we use the significance level at 0.05 with Bonferroni correction for multiple comparisons. Multiple imputations ⁵⁸ will be performed to handle missing data, and inverse probability censoring weighting ⁵⁹ will be used to handle early drop‐out. Bootstrap method will be used to construct confidence interval and P value.

Primary outcome

Descriptive statistics will be generated for baseline variables and adherence over time by the three groups. Group differences will be compared by t‐test and χ ² test, respectively. Baseline variables that are potentially related to adherence (e.g., sex and race) will be adjusted to improve the efficiency of estimating the average treatment effect. ⁶⁰ Generalized linear mixed effect models will be used to incorporate the correlation between adherence at 12 and 18 months. Interaction between time and treatment groups will be tested. ⁶¹ Separate Wald tests will be conducted to evaluate whether the adherence levels are the same among the three groups at 12 and 18 months. Sensitivity analysis will be performed by weighting each observation by the length of the period that the participant is available for exercise for that measurement. Further exploratory analysis will be performed to model the monthly adherence profile until 18 months.

Secondary outcomes

Descriptive statistics will be created for all secondary outcomes over time by the three groups, and each outcome will be plotted over the minutes of moderate intensity exercise.

Minimal clinically important difference health status . We will fit longitudinal segmented regression models of HF‐related health status on exercise minutes separately to obtain the breakpoint and then compare the effect with the established MCID for the KCCQ (change ≥ 5 points). ⁶² Baseline confounding variables will be adjusted in the regression model. Goodness of fit tests will be performed and constrained generalized additive models with isotonic component longitudinal marginal regression will be fitted if there is a lack of fit. If no difference in total effect comparing HEART Camp and HEART Camp Connect is detected, we will pool the two intervention groups for the MCID analysis. Causal mediation analysis with longitudinal mediators and outcomes will be performed. ⁶¹ , ⁶³

Behaviour change theory‐defined constructs and interim clinical events . Our primary analysis will evaluate the direct effect of treatment (combined intervention groups vs. control) on adherence and the indirect effect of treatment on adherence mediated by changes in of each our theory‐defined constructs with and without the adjustment of interim clinical events (e.g., hospitalizations and injury, as shown in Table 2). Behaviour change theory‐defined constructs include self‐efficacy, outcome expectations, attitudes, self‐regulation and social support. We will handle induced time‐varying confounding (interim clinical events) between the mediators and the adherence using g‐formula, ⁶⁴ marginal structural model ⁶⁵ or their combinations, ⁶⁶ which will allow us to calculate the pathway‐specific effects (such as the effects of intervention on outcome through a specific mediator or the adherence at prior timepoints, and prognostic indicators of HFpEF) with or without the potential impact of intervention on clinical events removed. Although our primary analysis assumes that the mediation effects for the two intervention groups are the same for better power, exploratory analysis allowing different effects of the type (in‐person or virtual) of treatment groups will be conducted to test whether the intervention type modifies the effect of a certain pathway, and the interaction term will be added as needed.

Cost analyses

We hypothesize that HEART Camp Connect and HEART Camp are more cost‐effective than EUC. Cost analyses will determine overall programme costs and average costs per participant. We use activity‐based costing to allocate staff time and direct and indirect expenses attributable to the activities associated with the interventions and EUC. Resources not requiring budgetary outlays are included in the activity‐based costing analysis as indirect costs. We will relate intervention costs to benefits and compare costs versus benefits in improving participants' physical function (6MWT), symptoms (PROMIS‐29) and QOL (EQ‐5D). Based on responses to the EQ 5D, the D1 valuation model, developed by Shaw et al., ⁶⁷ for US populations, will be used to determine utility tariffs (preference‐based index of quality of life). Utility tariffs range in value from 1.0 (i.e., full health) to −0.001 (i.e., extremely poor health). These utility tariffs are used to compare QOL from participants both over time using baseline and follow‐up questionnaires and between the interventions and EUC groups. Incremental cost‐effectiveness ratios will be constructed to compare differences in average per‐participant costs of each intervention and EUC to differences in average QOL.

Discussion

Regular exercise is a non‐pharmacological strategy that improves outcomes (exercise intolerance and QOL) for individuals with HFpEF; however, long‐term adherence to exercise interventions is difficult and the patient‐reported outcomes and inflammatory response to exercise are unclear. This prospective, randomized clinical trial tests our central hypothesis that behaviour change theory‐informed coaching strategies delivered virtually will promote long‐term adherence to exercise in adults with HFpEF and drive clinically meaningful and cost‐effective improvements in physiological and patient‐reported outcomes. Our previous work in adults with HF demonstrated that reversing negative attitudes, improving exercise self‐efficacy and managing relapse mediated long‐term adherence to exercise in an in‐person setting. ⁴¹ In this study, we test all mediation effects from our theoretical underpinnings to determine what behaviour change theory‐informed constructs may mediate adherence in HFpEF. We incorporate coaching strategies to address outcome expectations as a new theory‐informed construct. Qualitative data will help us to better understand the drivers of exercise adherence. We will also examine the impact of interim clinical events (e.g., hospitalization and injury), adherence at prior timepoints and prognostic indicators of HFpEF as treatment induced time‐varying confounders between mediators and adherence.

Prior research has demonstrated the potential of exercise interventions to drive short‐term improvements in physical function in adults with HFpEF. ³ , ⁴ , ⁸ However, these trials were not designed to promote sustained adoption of exercise behaviour over time. Few of these trials tested patient‐reported outcomes. ³ , ⁶⁸ Also, recent HFpEF exercise trials focused exclusively on in‐person intervention delivery during regular business hours, which limits access to participants with scheduling or transportation barriers. Interventional cost has not been examined and may be necessary for future translation to clinical practice and to support federal policy changes in Medicare reimbursement for exercise programmes for adults with HFpEF.

The proposed study incorporates several innovative features. We are the first to test the long‐term and sustained effects of a behaviour change theory‐informed intervention to promote adherence to exercise behaviour in HFpEF, and explore cost effectiveness, and only the second team to measure long‐term adherence using objective measures of frequency, intensity and time. ⁶⁹ To our knowledge, there are no other studies designed to define a benchmark of minutes of moderate intensity exercise that correspond to a minimal clinically important difference in a patient‐reported outcome for HF. We also include (a) several technical innovations including the use of a web application for real‐time data capture from heart rate sensors that triggers automated motivational messaging to HEART Camp Connect participants falling below adherence targets, (b) innovative analytic techniques including causal inference and assessment of interim clinical events and prognostic indicators of HFpEF as mediators of long‐term adherence to exercise and (c) use of a multiplex electrochemiluminescence immunoassays to simultaneously assess patterns in inflammatory markers in response to chronic exercise (a priority of NHLBI). These innovations challenge the current paradigm of exercise interventions and help to reach new horizons in HFpEF science using MCID and inflammatory markers. This proposal holds promise for efficacious and sustainable exercise interventions to improve long‐term outcomes for those with HFpEF.

We expect the HEART Camp Connect and HEART Camp groups to achieve greater long‐term adherence compared with EUC. If this occurs, we will test HEART Camp Connect and HEART Camp for equivalence on adherence. We expect to find equivalence between the interventions, which we will interpret as suggesting that in‐person intervention achieves similar outcomes to virtual in adults with HFpEF. Equivalence allows individuals' preference for in‐person or virtual exercise and delivery cost to be considered when recommendations to initiate and maintain exercise are incorporated into the clinical plan of care. We also expect to identify a threshold for minutes of moderate intensity exercise to potentially serve as a new adherence benchmark in HFpEF that differs from the 120 min of exercise in our current definition. Defining this cutoff in the context of MCID is critical for future studies that aim to improve outcomes most important to adults with HFpEF. We expect several inflammatory markers to change or stabilize in response to long‐term exercise. Behaviour change theory‐informed constructs may mediate adherence outcomes in in‐person versus virtual interventions. We know from qualitative interview data collected in our prior work that although social support was not a significant mediator, it is a key interventional strategy and may be a critical differentiator between virtual and in‐person coaching. Lastly, we expect our interventions to be more cost‐effective than EUC.

We acknowledge potential limitations in our protocol including those typical of prospective longitudinal trials. We selected secondary sites that care for a larger number of underrepresented groups to improve the representativeness of our sample and generalizability of our findings. We acknowledge the possibility of performance bias in the enactment of the interventional protocol across our coaches and sites. We reduce this risk with our comprehensive coach training and intervention fidelity plans. We acknowledge that in long‐term studies, attrition and missing data can be a challenge and contribute to bias. To this end, we request contact information from at least one family member or friend of each enrolled participant and incorporate regular data integrity checks to capture missing data. Access to smart devices and broadband could be an issue for participants in remote areas. We provide tablets with cellular data to anyone that does not have access to a smart device or high‐speed internet. Despite these potential limitations, our protocol has several key strengths including objective measures of exercise adherence facilitated by our innovative data capture platform. HFpEF is described as ‘the greatest unmet need in cardiovascular medicine today’ and is identified as a high‐priority target for clinical research. ²¹ , ⁷⁰ The findings resulting from this study hold great promise for efficacious and sustainable exercise interventions to improve long‐term outcomes for those with HFpEF.

Trial status

The trial is actively recruiting participants from sites at the University of Nebraska Medical Center/Nebraska Medicine in Omaha, NE, and Henry Ford Health System in Detroit, MI. Enrolment at the time of this submission is 72 participants.

Funding

Research reported in this manuscript was supported by the National Heart Lung and Blood Institute of the National Institutes of Health under award number R01HL163288. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Conflict of interest statement

None declared.

Alonso, W. W. , Bills, S. E. , Lundgren, S. W. , Keteyian, S. J. , Norman, J. , Fisher, A. L. , Zheng, C. , Kupzyk, K. A. , Wilson, F. A. , Dudley, T. J. , and Pozehl, B. J. (2025) HEART Camp Connect—Promoting adherence to exercise in adults with heart failure with preserved ejection fraction. ESC Heart Failure, 12: 3388–3398. 10.1002/ehf2.15341.