Recruitment, Retention, and Participant Diversity in the Context of a Heart Failure Self-Care Digital Intervention Decentralized Clinical Trial

Abstract

Background: Low participant accrual is a persistent concern in physiological disease intervention trials, inflating costs and jeopardizing the timeliness and validity of findings. Investigators are increasingly adopting decentralized methods to facilitate participation. Objective: To add to the recruitment evidence base by describing the performance of direct and remote recruitment strategies in a decentralized randomized controlled trial of a digital intervention to improve heart failure self-care behaviors. Methods: We conducted a descriptive analysis of referral, enrollment, and retention rates; cost; and sociodemographic diversity of participants across six recruitment streams. Data were collected from Sept 30, 2022 to June 30, 2025. Results: Decentralized recruitment channels generated 97.5% of enrollments and achieved varying success with respect to sample representativeness. Enrollment rates progressed in accordance with proposed timelines. Retention at 6 months was 82.6%. Conclusions: Decentralized recruitment strategies are feasible, cost effective, and conducive to achieving enrollment targets.

Introduction

Recruitment is a persistent challenge in physiological disease research. Participation in clinical trials is encumbered by a wide range of financial, logistic, cultural, and systemic barriers, and it can be particularly challenging in the context of heart failure (HF), where functional disabilities and exacerbations requiring urgent medical attention may curtail participation. These obstacles, coupled with low levels of public and provider research awareness, can undermine trial enrollment goals, resulting in delays, terminations, and underpowered analyses. Recruitment challenges are well documented in the HF literature, where median monthly enrollment rates of <1 participant/site/month have been reported for years.¹ Sluggish cardiovascular trial enrollment is a global concern, with 41% of failed trials attributing early termination to poor accrual,² but trials in North America have been particularly hampered by lower than average and declining enrollment rates.³ Over the last 15 years, North American participants have accounted for only 8 to 15% of global HF trials enrolling hospitalized participants; this modest representation has been attributed to site-centric challenges to conducting research.⁴ Representation remains a primary concern in HF recruitment, with low participation-to-prevalence ratios among racial and ethnic communities disproportionately burdened by the disease.⁵

New recruitment challenges have emerged as decentralized clinical trials (DCTs) have transformed the research landscape. Decentralized approaches, harnessing digital tools to conduct trial activities without the need for any in-person contact, are increasingly adopted as researchers seek to expand their reach into previously inaccessible communities, accelerate recruitment timelines, and minimize participants’ burdens. Direct comparisons between decentralized and site-centric recruitment strategies have favored digital innovations in each of these domains.⁶ In lieu of traditional recruitment channels grounded in community embeddedness and relationships, investigators have recourse to a wide range of recruitment platforms and services, including large registries, social media, and contracts with specialized clinical research organizations. ResearchMatch,⁷ for example, is a nonprofit national registry, accessible in English, Spanish, and simplified Chinese, that connects investigators with volunteer participants. Investigators can search a pooled, non-identifiable database for health characteristics and share study information over a secure email platform with potential participants who can elect to complete screenings, request information, and share their contact information. Companies like Trialfacts⁸ offer paid recruitment services using data-driven social media advertising, and others, such as StudyKIK,⁹ draw research participants from digital and mobile platforms to a centralized registry where they can be matched with relevant trial opportunities.

Nevertheless, although DCTs are expected to increase the diversity of study samples, their overall impact on representation and participation-to-prevalence ratios remains uncertain. Similarly, although the promise of quicker recruitment and increased retention suggests that decentralized strategies could be financially advantageous, clear cost accounting of decentralized recruitment methods has not been widely reported in the literature, particularly in the context of small-scale behavioral investigations. Experts caution that decentralized elements are not uniformly suitable for all research questions, and their inclusion should be justified in terms of risk minimalization, the appropriateness of task delegation, the accuracy of eligibility vetting, and the potential to introduce sampling bias.¹⁰ DCTs have proliferated in the context of behavioral health interventions; but the same is not true for physiological chronic disease, and the value of these approaches for individuals with advanced chronic disease remains uncertain.

In the present study, we describe the performance of a range of recruitment mechanisms used in a DCT involving digital tools in the context of HF, an advanced chronic disease. Specifically, we investigated the following questions: 1) How do different recruitment mechanisms perform in terms of their respective contributions to the recruitment, enrollment, and retention of participants, and to recruitment costs? 2) Is the sociodemographic diversity of participants recruited with a decentralized recruitment strategy representative of the broader HF population? And 3) Are decentralized recruitment methods efficient in meeting target enrollment timelines? We also explore barriers to remote recruitment, including participant fraud, technological literacy, age, and socioeconomic status.

Methods

Design

This is a descriptive secondary analysis of recruitment data froma decentralized randomized controlled trial evaluating the efficacy of an adaptive sensor-controlled digital game to improve adherence to self-care behaviors among older adults with HF,¹¹ registered at ClinicalTrials.gov (NCT050556129) and supported by the NIH under award R01HL160692. This study was reviewed and approved by the University of Texas Institutional Review Board (STUDY00001438). To minimize the risk of deductive disclosure, de-identified data will be made available upon request only under a data-sharing agreement, with IRB approval.

Participants

A combination of traditional and decentralized strategies was used to recruit participants in 20 U.S. states with the highest disparities in HF: Alabama, Arkansas, Georgia, Idaho, Iowa, Indiana, Kansas, Kentucky, Louisiana, Michigan, Mississippi, North Carolina, Ohio, Oklahoma, Oregon, South Carolina, Tennessee, Texas, Utah, and West Virginia. Traditional recruitment (i.e., direct outreach) was conducted in a cardiac rehabilitation center and at a hospital in central Texas. Decentralized strategies included use of the ResearchMatch registry, geographically targeted Internet Service Provider recruitment in areas with a high prevalence of cardiovascular disease, as well as contracting with StudyKIK and Trialfacts. Advertisements on a “freemium” streaming radio platform were also used to expand recruitment in Mississippi, the state with the highest age-standardized HF death rate in the U.S.¹¹ but with the lowest enrollment in our study. Recruitment began on September 30, 2022. Data reported here were collected from September 30, 2022, through June 30, 2025.

Potential participants were considered eligible if they were at least 45 years old; had a diagnosis of class I, II, or III HF based on the NYHA functional classification system; and had been hospitalized for HF within the last calendar year. They also had to be fluent in English, ambulatory without an assistive device, and capable of passing a short-form cognitive assessment. Exclusion criteria consisted of severe visual or sensory impairment, prior heart transplantation or pacemaker implantation, renal failure, or any end-stage or terminal diagnosis. Initial pre-screening surveys were administered on a web-based REDCap platform. Study staff contacted all individuals who passed the initial pre-screening by phone to complete screening. Participants were required to submit discharge summaries to verify HF diagnoses and hospitalizations. Consent forms were reviewed with eligible participants over the phone before informed consent was collected electronically through REDCap.

Participants were randomized to either the intervention or the control group. Both groups received a smart scale and activity tracker with an app for visualizing data. The control group received educational content related to HF self-care in paper form, and the intervention group received the same educational content electronically, embedded within an app-based sensor-controlled digital game responsive to real-time HF self-management behaviors. Study staff conducted remote training sessions on the installation and use of all devices. Incentive e-gift cards were distributed at each data collection time point (baseline, 6, 12, and 24 weeks). Between September 30, 2022 and September 27, 2024, incentive payments were fixed at $10 across all data collection time points for a total of $40 per participant. Beginning on September 28, 2024, a more substantial incentive was implemented and applied retroactively in response to participants’ feedback, with $25 distributed at baseline, $30 at week 6, $35 at week 12, and $50 at week 24. Given the timeline for this change, data were insufficient to determine whether this change in incentives influenced recruitment or retention rates.

Measures

Referral, eligibility, and enrollment data were obtained from databases maintained for each recruitment stream, which were routinely checked to remove duplicate and fraudulent submissions. Retention rates were calculated at each of the three follow-up milestones in the study. Cost per enrolled participant was calculated using the total amount invoiced by each recruitment stream divided by the accrual attributed to that stream. Demographic and socioeconomic data collected in baseline surveys were used to compare gender, age, race and ethnicity, and financial hardship across recruitment streams. Participant-reported motivations for enrolling in the study were captured as free-form text in pre-screening questionnaires and manually categorized by a study team member. Finally, zip code data were used to calculate the geographic distance between the research team and participants, as well as to characterize participants’ areas of residence as urban, suburban, or rural. Descriptive statistics were calculated with Microsoft Excel.

To summarize, the primary outcomes of interest in this study are the conversion and retention rate for each recruitment stream, the cost per enrolled participant for each recruitment stream, and participants’ characteristics for each recruitment stream. The frequency of reported motivations for participating in the study and the mean and range of participants’ distance from the study center are secondary outcomes.

Results

Recruitment, enrollment, and retention

Between September 30, 2022 and June 30, 2025, a total of 3,947 individuals completed pre-screening questionnaires (Table 1). Trialfacts referrals accounted for 87.26% of all completed pre-screening questionnaires, with 3012 direct referrals and 432 indirect referrals who reported that a friend or family member had shared a Trialfacts advertisement with them. StudyKIK contributed 10.51% of referrals; Spectrum Reach and ResearchMatch contributed less than 2% combined. Direct recruitment, which was attempted at two clinical partnership sites (one cardiac rehabilitation facility and one hospital), accounted for less than 1% of all referrals. Twelve months into recruitment, we trialed a novel recruitment channel using advertisements on a “freemium” streaming radio platform to expand outreach in Mississippi, the state with the highest age-standardized HF death rate,¹² which remained persistently underrepresented in our study. Thirty-second radio spots, broadcast between 6 a.m. and 7 p.m. on a free streaming platform 120 times over a 4-week period, yielded an estimated 188,000 gross impressions (i.e., ads were heard by listeners 188,000 times) but no referrals. The inefficacy of radio as a recruitment platform may have been due to its asynchronous nature, which introduces a cognitive burden by requiring participants to remember information about the study and take action later, in contrast with digital platforms that allow participants to immediately engage through direct links. This experience informed our decision to prioritize digital options that reduce barriers to real-time interaction.

Table 1.

Referrals, eligibility and enrollment by recruitment platform

Recruitment stream	Referrals		Eligible at pre-screening		Enrolled
	n	%	n	%	n	%
Trialfacts	3,444	87.26	1027	29.82	175	5.08
Spectrum Reach	35	0.89	19	54.29	11	31.43
StudyKIK	415	10.51	31	7.47	6	1.45
ResearchMatch	33	0.84	14	42.42	3	9.09
Radio	0	0	--	--	--	--
Clinical partnerships	20	0.50	14	70.00	5	25.00
Overall	3,947	100	1105	27.99	200	5.07

Ineligibility

Of 3,444 total referrals, 2,842 (72.0%) were deemed ineligible for study participation on the basis of at least one pre-screening response. Of them, 194 (6.83%) were ineligible on the basis of age, and 646 (22.73%) lived outside the geographic sampling range of the study. Health concerns and disabilities precluded participation for 661 (23.26%) who were ineligible due to a comorbidity and 538 (18.93%) who could not walk without an assistive device. Within the past year, 1,453 (51.13%) had not been hospitalized for HF. StudyKIK referral reports did not itemize reasons for ineligibility; consequently, 384 referrals recruited through this stream were documented as ineligible without a specified reason.

Of the 1105 who passed the pre-screening phase, 321 (29.05%) were subsequently unreachable by phone after five attempts by study staff. After completing the full screening, submitting discharge summaries to satisfy hospitalization eligibility requirements, and providing informed consent, 200 (18.10%) were ultimately enrolled.

Retention

Due to our staggered enrollment, participants do not complete study activities on the same schedule. As of June 30th 2025, not all participants had been enrolled long enough to reach all data collection milestones. Accordingly, retention at each timepoint shown in Table 2 is calculated as a percentage of the participants who would have been expected to have reached a data collection milestone based on their date of enrollment. The overall retention rate was 89.5% at week 6, 87.7% at week 12, with 82.6% retained throughout the full 24 weeks. Retention rates varied across streams, from as low as 60% to as high as 90.9%, but the imbalance in group sizes precludes meaningful interpretation.

Table 2.

Retention by recruitment stream

	Baseline		6 weeks		12 weeks		24 weeks
	n = 200		n = 199		n = 187		n = 161
	n	%	n	%	n	%	n	%
Trialfacts	175	100%	158	90.8%	144	88.8%	113	83.8%
Spectrum Reach	10	90.9%	10	90.0%	10	90.9%	10	90.9%
StudyKIK	6	100%	5	83.3%	5	83.3%	5	83.3%
ResearchMatch	3	100%	2	66.7%	2	66.7%	2	66.7%
Clinical partnerships	5	100%	4	80%	3	60%	3	60%
Overall	199	99.5%	179	89.5%	164	87.7%	133	82.6%

Recruitment costs

The total cost of recruitment was $111,802.04 (Table 3). Recruitment mechanisms varied in cost from $0 (ResearchMatch, word of mouth) to $92,660.04 (Trialfacts). The cost of site-centric recruitment at clinical partnership was calculated on the basis of recruiters’ time, compensated at a rate of $16.00/hour.

Table 3.

Cost by recruitment stream

Recruitment stream	Months of Active Recruitment	Participants Enrolled	Cost	Average Monthly Cost	Cost per participant
Trialfacts	26	175	$ 92,660.04	$ 3,563.85	$529.49
StudyKIK	4	6	$ 11,174.00	$ 2,793.5	$ 1,862.33
ResearchMatch	28	3	$ 0.00	$ 0.00	$ 0.00
Spectrum Reach	7	11	$ 12,200.00	$ 1,742.86	$ 1,109.09
Radio	1	0	$ 1,000.00	$ 1,000.00	--
Clinical partnerships	8	5	$ 355.00	$ 44.38	$ 71.00
TOTAL	--	200	$111,802.04	--	$ 559.01

Recruitment timelines

The average time elapsed between referral and enrollment was 9.7 days; 85.4% of participants completed baseline surveys within 14 days of their initial referral, with 63.3% taking 7 days or less. Time between recruitment and enrollment was affected by a number of factors, including unforeseen health issues, difficulty locating or obtaining discharge summaries from hospitalizations, travel, and scheduling difficulties. After enrollment, the shipment and installation of equipment required an average of 12 additional days, with 83.4% of participants using study devices within 2 weeks of enrollment. An average of 127 pre-screening questionnaires were completed each month, yielding an average 6.45 monthly enrollments. Paid recruitment was paused during the winter holiday season, when advertisements became prohibitively expensive; consequently, yield was highest in early 2024 and declined at the end of each calendar year. As of June 30, 2025, enrollment had concluded in accordance with the originally proposed timeline.

Participant Diversity

As of June 30, 2025, 199 participants had completed the baseline surveys for sociodemographic data, which are summarized in Table 4: 70.9% were aged ≤64 years; nationally, this age bracket represents 21.5% of all individuals with HF.¹³ The proportion of women enrolled in the study (40.7%) is highly representative of the national disease burden borne by women (41%).¹³ Non-Hispanic White participants are overrepresented (74.4%), but Black/African American (14.6%), Hispanic/Latinx (6%), and Asian (0.5%) participants are not represented at levels that match national HF distribution statistics (29%, 14.6%, and 2.5%, respectively).¹³ The imbalance in group size precludes meaningful between-group analyses. The study captures the financial hardships faced by many people living with HF, with 61.8% reporting difficulty paying for basic needs such as food, housing, and heating, and over a quarter (28.6%) reporting difficulties in arranging transportation to medical appointments.

Table 4.

Characteristics of participants by recruitment stream

	Recruitment Stream
	Trialfacts (n = 175)	StudyKIK (n = 6)	Research Match (n = 3)	Clinical Partners (n = 5)	Spectrum Reach (n = 10)	Full sample (n = 199)
Mean age, years (range)	59.5 (45-89)	55 (49-64)	60.7 (50-71)	48 (45-56)	62.5 (49-74)	59.3 (45-89)
Sex, n (%) Female	72 (41.1)	2 (33.3)	2 (66.7)	1 (20%)	4 (40)	81 (40.7)
Education
High school diploma, or less	53 (30.3)	3 (50)	1 (33.3)	2 (40)	2 (20)	61 (30.7)
College degree or some college	81 (46.3)	3 (50)	0 (0)	3 (60)	5 (50)	92 (46.2)
Graduate / professional degree	41 (23.4)	0 (0)	2 (66.7)	0 (0)	3 (30)	46 (23.1)
Race & Ethnicity, n (%)
American Indian / Alaska Native
Hispanic/Latino	1 (0.6)	0 (0)	0 (0)	0 (0)	0 (0)	1 (0.5)
Non-Hispanic/Latino	0 (0.0)	0 (0)	0 (0)	0 (0)	1 (10)	1 (0.5)
Prefer not to answer	1 (0.6)	0 (0)	0 (0)	0 (0)	0 (0)	1 (0.5)
Asia n	1 (0.6)	0 (0)	0 (0)	0 (0)	0 (0)	1 (0.5)
Black or African. American
Hispanic-Latino	1 (0.6)	0 (0)	0 (0)	0 (0)	0 (0)	1 (0.5)
Non-Hispanic/Latino	26 (14.9)	3 (50)	0 (0)	0 (0)	0 (0)	29 (14.6)
White
Hispanic-Latino	7 (4)	0 (0)	0 (0)	0 (0)	1 (10)	8 (4.0)
Non-Hispanic/Latino	131 (74.9)	3 (50)	3 (100)	5 (100)	6 (60)	148 (74.4)
Prefer not to answer	3 (1.7)	0 (0)	0 (0)	0 (0)	0 (0)	3 (1.5)
Other
Hispanic/Latino	1 (0.6)	0 (0)	0 (0)	0 (0)	2 (0)	3 (1.5)
Non-Hispanic/Latino	1 (0.6)	0 (0)	0 (0)	0 (0)	0 (0)	1 (0.5)
Prefer not to answer	2 (1.1)	0 (0)	0 (0)	0 (0)	0 (0)	2 (1.0)
Financial hardship, n (%)
Difficulty paying for basic needs	108 (61.7)	5 (83.3)	1 (33.3)	5 (100)	4 (40)	123 (61.8)
Transportation Insecurity	51 (29.1)	1 (16.7)	1 (33.3)	2 (40)	2 (20)	57 (28.6)

Geographic range and residential characteristics

The haversine distance formula was used to calculate the distance between each participant and the study team, using zip code data. Participants lived as close as 2 miles and as far as 1,708 miles from the study team. The average distance between participants and the study team was 478 miles. Zip codes were classified as urban, suburban, or rural, using a national predictive model based on housing density, census data, and residents’ subjective characterizations of their place of residence.¹⁴ Given this classification system, 32.66% of participants were categorized as residing in rural areas, 54.77% in suburban areas, and 12.56% in urban areas. More participants in this study lived in Texas (48.2%) and bordering states (9.1%), suggesting that our affiliation with the University of Texas may have influenced participation through name recognition or reputation.

Motivations for participating in research

Among the participants, self-interest was a stronger motivator for study participation than altruism: 25% were motivated by a desire to learn more about HF and another 34% by the desire to directly improve their health outcomes, whereas 29% cited altruistic motivations (i.e. advancing science, helping others with HF); only 3% cited compensation as their motivation to participate. Nineteen participants did not report a motivation, so the full range of motivations driving research participation may not be fully captured in the data.

Discussion

Inclusivity as a key concern in decentralized recruitment

The digital divide presents a significant challenge to equity in decentralized research. To mitigate digital inequities, we offered free mobile phones and data plans to those who could not otherwise participate in the study, ensured that all apps and smart devices were compatible with both Android and iOS platforms, provided real-time tech support, and closely monitored device activity for indications that participants needed assistance in using study technologies.¹⁵ However, the success of these methods is contingent upon participants’ ability to access information about the study in the first place. Strategies are needed to increase awareness of research opportunities in communities that are less digitally connected (i.e., rural, socioeconomically disadvantaged, minoritized, more advanced in age, and less educated). Participants in this study lived, on average, hundreds of miles from the study team, and 32.66% resided in rural areas. Access-driven rural–urban disparities in HF are particularly striking for Black men and women,¹⁶ who in this study were not represented at levels indicative of the relative burden of HF in the Black community. Rural areas may lack the infrastructure for affordable high-speed internet service and may have few public resources that support digital access and inclusivity. Therefore, we are encouraged by the high level of participation from this historically underrepresented group. The average age of participants was under 60, which may reflect the growing number of younger U.S. adults diagnosed with HF but may also reflect greater digital literacy and self-efficacy among younger Americans. Although older adults’ cellphone and smart home device ownership is now on par with that of younger adults, older adults have been slower to adopt wearable devices, and their confidence in digital literacy skills has been found to decrease with age.¹⁷ In consideration of these patterns, the nature of the study itself, grounded in consumer-facing health technologies, may have disproportionately attracted younger individuals more comfortable with these tools.

Considerations to protect the validity of DCT findings

Decentralized recruitment, with no face-to-face interaction between participants and study staff, is inherently at risk of fraud and deception, particularly when financial or material incentives are involved.¹⁸ Precautions against deception are all the more important in recruitment from centralized registries designed to streamline access to study opportunities; in a recent review, the increase in null cardiovascular clinical trials observed between the years 2000 and 2015 was found to be most strongly associated with recruitment from the ClincalTrials.gov registry.¹⁹ In our study, administrators routinely conducted data checks and used automated methods to screen for duplicate submissions (e.g. tracking IP addresses, identical responses), manual fraud, and “bot” activity (e.g., rapid sequences of submissions, submissions with improbable responses). The strongest safeguard against participant deception in this study was the requirement that participants submit discharge summaries from an HF hospitalization to confirm their eligibility. These documents contain technical terminology, corporate letterheads and templating, and often have security features which make them difficult to fabricate convincingly.

Retention

DCTs are thought to improve retention rates by minimizing burdens of participation. To date, there is limited evidence in support of these claims. Miyata et al.⁶ have identified seven DCTs reporting retention outcomes, of which only four presented direct comparisons between decentralized and traditional trial methods. However, retention rates were found to be consistently high in DCT participant cohorts, with decentralized retention rates higher than those found in site-centric study arms. In the HF literature, investigators using a hybrid recruitment strategy reported a 15.58% attrition rate between enrollment and baseline data collection, which involved an in-person site visit.²⁰ The relative ease of completing baseline surveys remotely may be the differentiating factor contributing to our success in retaining 99.5% of enrolled participants at baseline. Our retention rates at 6, 12, and 24 weeks (89.5%, 87.7%, 82.6%, respectively) mirror those reported in a traditional 6-month pilot clinical trial to improve hypertension self-management among low-income African Americans, in which 91.5% of participants were retained at baseline, 88.1% at 3 months, and 83.1% at the trial’s conclusion.²¹ Given the markedly greater symptom burden and hospitalization rate in those with HF, these findings are encouraging. Further examination of digital self-efficacy and patterns of engagement as predictors of attrition is merited in light of the substantial cost and effort associated with participant enrollment.

Limitations

Although this study has generated insights that may be valuable to investigators conducting DCTs, its limitations warrant acknowledgement. Recruitment streams were not all used simultaneously or for the same length of time. Consequently, comparisons of the relative effectiveness of the recruitment streams are limited to monthly averages and may be influenced by time bias. Additionally, the cost accounting in this study was restricted to direct recruitment costs only and did not include the cost of study staff’s time spent on fraud detection, follow-up and screening phone calls, time engaging with recruitment stakeholders, etc. Per-participant costs were calculated on the basis of aggregate yield, and therefore our analysis did not allow us to isolate differentials in recruitment costs across sociodemographic groups. Given the contextual nature of this DCT, the generalizability of our findings may be limited to researchers using digital technologies in their work with older adults diagnosed with a physiological chronic illness. Participants’ motivations for joining the study were self-reported at pre-screening and may be influenced by social desirability bias. Similarly, because we have no way to ascertain the characteristics of all individuals who were informed about the study (e.g., those who heard a radio advertisement or saw an online advertisement), we are unable to speculate on how self-selection bias might have influenced our study sample. Finally, this study presents preliminary data; while recruitment has concluded, the intervention is ongoing and expected to be completed by October 2025.

Conclusion

This descriptive analysis provides empirical evidence that may prove useful to investigators as they develop recruitment strategies. The majority of participants in this trial (87.5%) were recruited via Trialfacts, with 83.8% of those recruited through this stream retained throughout the 6-month trial. Our findings suggest that decentralized recruitment may be effective in overcoming logistical barriers to participation, particularly those linked to financial hardship and transportation challenges, yet more research is needed to identify methods to increase participation in racial and ethnic minority groups in decentralized HF research to address ongoing participation-to-prevalence imbalances. Strategies are needed to engage older seniors in decentralized research as well, because this population has been slower to adopt wearable digital health technologies but stands to gain the most from such innovations given their higher rates of chronic illnesses that require informed, assiduous self-care. Overall, our study demonstrates that decentralized recruitment strategies are feasible, cost effective, and conducive to meeting enrollment targets and timelines.

Grants and funding

Research reported in this publication was supported in part by the National Heart, Lung, and Blood Institute of the National Institutes of Health under award R01HL160692 and the National Institute of Nursing Research of the National Institutes of Health under Award Number T32NR019035. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.