Pilot test of a Multi-Component Implementation Strategy for Equity in Advanced Heart Failure Allocation

Abstract

Advanced heart failure (AHF) therapy allocation is vulnerable to bias related to subjective assessments and poor group dynamics. Our objective was to determine whether an implementation strategy for AHF team members could feasibly contribute to organizational and culture change supporting equity in AHF allocation. Using pretest-posttest design, the strategy included an 8-week multi-component training on bias reduction, standardized numerical social assessments, and enhanced group dynamics at an AHF center. Evaluations of organizational and cultural change included pretest-posttest AHF team member surveys, transcripts of AHF meetings to assess group dynamics using a standardized scoring system, and posttest interviews guided by framework for implementing a complex strategy. Results were analyzed with qualitative descriptive methods and Brunner-Munzel tests for relative effect (RE, RE >0.5 signals posttest improvement). The majority of survey metrics revealed potential benefit with RE >0.5. RE were >0.5 for 5 of 6 group dynamics metrics. Themes for implementation included: 1) promotes equitable distribution of scarce resources; 2) requires change in team members’ time investment to correct bias and change meeting structure; 3) slows then accelerates the allocation process; 4) adaptable beyond AHF and reinforceable with semi-annual trainings. An implementation strategy for AHF equity demonstrated feasibility for organizational and culture change.

INTRODUCTION

Allocation of advanced heart failure (AHF) therapies, heart transplants and ventricular assist devices (VAD), is vulnerable to bias.^1–4 After a patient has been referred to an AHF center, a multidisciplinary team of experts must determine whether a patient has appropriate indications for AHF therapies and lacks contraindications that prevent benefit from AHF therapies. Indications are guided by objective criteria from International Society of Heart and Lung Transplantation guidelines.^5,6 In contrast, many of the guidelines-based contraindications to AHF therapies are relative,^5,6 and the point at which a contraindication becomes inappropriate for AHF therapies is subjective and can vary by center. This contributes to an unequitable AHF allocation process.

National studies have demonstrated that the most significant factors contributing to disparities in AHF allocation by race and gender are related to identifying relative contraindications, which include social bias of AHF team members, unstandardized descriptive assessments of social history, and dysfunctional group dynamics of the AHF team.^1,2,7 Each of these factors are fixable. Evidence-based strategies and/or scientific frameworks exist for each factor,^8–13 but have not been systematically used nor purposefully combined to make the allocation of AHF therapies more equitable. The National Academies of Science, Engineering, and Medicine recently convened a taskforce to review these systemic concerns in U.S. transplantation and made an urgent call for action.¹⁴ Given consistent unequitable allocation of AHF therapies by race, ethnicity, and gender,^15–19 we sought to adapt these evidence-based strategies and scientific frameworks into a pilot test of a multi-component implementation strategy for equity in the decision-making process for allocating AHF therapies.

It is of prime importance to demonstrate feasibility of strategies to perform a complex process such as organizational and culture change. Understanding this process is the focus of implementation science and necessary to make evidence-based interventions accessible in real-life. Therefore, we combined scientific frameworks and evidence-based strategies that reduce the impact of social bias, utilized standardized numerical assessments of social history, and change AHF team environment to facilitate high quality group dynamics in an implementation strategy called, Seeking Objectivity in Allocation of Advanced Heart Failure Therapies (SOCIAL HF). Using pretest-posttest design, we examined whether an implementation strategy for AHF team members could feasibly contribute to organizational and cultural change promoting equity in AHF allocation decision-making process over the course of 1.5 years at an AHF center. We specifically assessed for signals for change in behavior and workplace environment, signals for change in group dynamics, and perceptions for implementing a complex strategy.

METHODS

Study Design

A pretest-posttest pilot study of an implementation strategy for AHF team members was conducted at a U.S. AHF center from 9/2019-3/2021. AHF Team members (i.e. coordinators, nurses, pharmacists, physicians, social workers) were eligible for inclusion if they were official members of the AHF team as reported on allocation meeting attendance sheets typically submitted to United Network for Organ Sharing for auditing purposes. The study included a 5-month pretest evaluation, followed by 8-week implementation strategy, and 6-month posttest evaluation (Table 1). For five months (n=9 meetings, only recording meetings including at least 1 patient evaluation for AHF therapies pretest), weekly in-person AHF allocation meetings were securely audio recorded, transcribed by Health Insurance Portability and Accountability Act-approved transcriptionist, and redacted for personal identifiers by the study team. Coronavirus disease (COVID-19) and staffing related issues resulted in a pause in allocation of AHF therapies that did not resume during study follow-up and led to several months delay of start of the implementation strategy. However, weekly meetings continued throughout the study period, including discussion on management of patients with AHF and patients with prior transplant or VAD via videoconference. One week before implementation of the strategy, all AHF team members were invited to complete a pretest survey via personalized email. Then the SOCIAL HF strategy was conducted over eight weeks via videoconference. This was followed by interviews of five individuals from different disciplines regarding the process of implementing the strategy. Simultaneously, all AHF team members were invited to participate in a posttest survey via email. Secure audio recordings of weekly allocation meetings recommenced immediately following the implementation strategy for six months (n=23 meetings) via videoconference. This study was approved by the University of Arizona Institutional Review Board. Consent was obtained for study participants. Additional details regarding the methods are available in the Supplementary Methods.

Table 1.

Study Timeline

Implementation Strategy: SOCIAL HF

SOCIAL HF is an eight-week interactive training that was provided via videoconference by the principal investigator, PI (KB). SOCIAL HF was designed using a validated implementation science framework for changing behavior [Capability, Opportunity, and Motivation for Behavior Change/Behavior Change Wheel (COM-B/BCW)].²⁰ SOCIAL HF components include: 1) evidence-based bias reduction training tailored for HF^8,9; 2) instruction on use of validated measures of social support and medication adherence, which enhance objectivity;^10,11 and 3) instruction on use of anonymous electronic voting and equitable seating arrangement, which restructures how workplace interactions occur (Figure 1, Supplementary Table 1).^12,13

Figure 1. SOCIAL HF Intervention Strategy.

SOCIAL HF is composed of eight twenty-minute interactive training sessions with all AHF team members, and 1-hour training with AHF team leaders (medical and surgical directors), and 1-hour training with social worker or designees. ARMS indicates Adherence to Refills and Medications Scale; SIPAT, Stanford Integrated Psychosocial Assessment for Transplantation. The inner circle describes the problems contributing to AHF allocation inequity. The next light green oval describes training strategies. The next dark green oval describes the intended audience for each training strategy. The final outer blue oval and blue section describe the methods of evaluation. Overlapping sections represent overlapping problems, training strategies, audience, and methods of evaluation.

The evidence-based bias reduction training portion was adapted from the Breaking the Bias Habit: Bias Reduction in Internal Medicine Program® developed by Dr. Molly Carnes, which significantly increased self-reported bias-reducing behavior with concomitant improvements in the workplace interactions and increased diversity among new faculty hires.^8,9 The SOCIAL HF adaptation included 20-minute installments during the regularly scheduled weekly AHF meetings focused on: 1) introduction to HF racial/ethnic and gender disparities, 2) bias in HF allocation, 3) bias as a habit, 4) identification and labelling of forms of bias, 5) bias reduction & anti-racism strategies, 6) bias reduction and anti-racism strategies specific to the SOCIAL HF Trial, 7) using and interpreting numeric scales for social assessments, and 8) using electronic voting and equitable seating and a commitment to change activity. At week five, an additional 1-hour moderator training was provided to the allocation leaders (cardiologist and surgeon) that reviewed best practices for facilitating group meetings and provided case examples on achieving equitable participation while limiting groupthink and discriminatory practices.

A separate 1-hour interactive training was provided via videoconference by the PI to the social worker/designee on use of evidence-based metrics for social history that enhance objectivity, which supplemented training provided to all AHF members on the same topic. The Stanford Integrated Psychosocial Assessment for Transplantation (SIPAT) has the most comprehensive evidence for assessing risk of adverse events in AHF using a standardized numeric scale.^10,21 SIPAT metrics, scoring, recommendations based upon scores, and standardized reporting to AHF team were reviewed. An evidence-based method of numerically assessing medical adherence was also discussed, Adherence to Refills and Medications Scale (ARMS),¹¹ and methods of assessing and reporting results to AHF team were reviewed. Literature supporting evaluation methods was provided. Additional details regarding validity and reliability of SIPAT and ARMS are in the Supplementary Methods.

Instruction on anonymous voting and equitable seating was taught during the eight-week training and re-emphasized during moderator training to promote improved group dynamics. Anonymous voting and round spatial seating are scientifically established methods to reach quality consensus during group decision-making meetings.^12,13 AHF team members were taught how to use Poll Everywhere to electronically vote for allocation decision via computer or phone following a patient presentation. Real-time results could be used at the discretion of the AHF team to guide allocation. No patient identifiers were used in the Poll Everywhere system. Instructions on use of round spatial seating and dispersion of seating (to avoid clumping by discipline) were provided for return of allocation meetings to in-person mode.

Surveys

An 87 question survey was provided pre and post the SOCIAL HF implementation strategy including stem questions on knowledge, attitudes, behavior and participant demographics (Appendix: Survey, Table 1, Supplementary Methods), previously developed for the Breaking the Bias Habit: Bias Reduction in Internal Medicine Program.^® In this study reporting, we focused on Likert scale questions examining willingness to change behavior related to clinical decision-making (30 questions), workplace environment attitudes (2 questions), and demographics. For each behavior change (recognizing bias, speaking about equity, challenging stereotypes related to clinical decision-making, assessing environment, adopting perspectives, and becoming better acquainted with an individual with a different background) participants were asked to rate level of agreement with performing each behavior (desire to perform behavior, confidence in behavior, perceived benefit in behavior, perceived risk in behavior, regular performance of behavior) on 7-point Likert scale (strongly agree to strongly disagree). Workplace environment was evaluated with 5-point Likert scale questions (extremely to not at all) on whether overall clinical workplace environment enforces stereotypes and the level of satisfaction with team leaders’ efforts to create collegiality and support. Survey invitations were sent via email using REDCap and took approximately 10-15 minutes to complete.

Transcripts for Group Dynamics Metrics

Group dynamics metrics were evaluated from transcripts of AHF allocation meeting audio recordings pre and post the SOCIAL HF implementation strategy (Table 1). A validated scoring system for group dynamics designed for clinicians, de Groot Critically Reflective Diagnoses Protocol (DCRDP),^22,23 evaluates six metrics associated with group dynamics including: challenging groupthink (welcoming new ideas that differ from the dominant viewpoint), critical opinion sharing (sharing opinions that can be openly questioned), research utilization (discussing research), openness to mistakes (willingness to admit faults), asking and giving feedback (requesting and offering suggestions), and experimentation (considering how to test an idea). Each metric was scaled from 1-4 (best to worst group dynamics), 1: demonstrating interaction (dialogue between group members) and reflection (discussing viewpoints shared), 2: reflective on an individual basis (single person discussing viewpoint shared without dialogue), 3: non-reflective and non-interactive, and 4: restricted (cutting off individuals from sharing viewpoints).^22,23 Restricted scores are associated with clinical uncertainty,²² which is associated with biased care of minoritized racial and ethnic patients.^24,25 Restricted scores have been associated with lower likelihood of allocating AHF therapies to women across multiple and racial and ethnic backgrounds.⁷

Interviews for Strategy Implementation

Anonymous interviews with five individuals from different disciplines were conducted post the SOCIAL HF implementation strategy (Table 1). Interview questions were guided by a framework for evaluating and implementing a complex strategy, normalization process theory.^26,27 Using the four core constructs of normalization process theory, we examined: 1) coherence: what is SOCIAL HF, how does SOCIAL HF compare to prior methods of allocation; 2) cognitive participation: who performs SOCIAL HF, what type of investment is needed to perform SOCIAL HF; 3) collective action: how does SOCIAL HF impact time duration of the allocation meeting; and 4) reflexive monitoring: how is SOCIAL HF perceived, how should SOCIAL HF be adapted or changed (Appendix: Interview Guide).^26–28 Interviews were conducted by a trained member of the study team(RHY) via videoconference and lasted approximately 15 minutes per interview.

Outcomes

We evaluated feasibility of implementation strategy to create organizational and culture change through 1) signals for change in behavior and change in workplace environment via surveys, 2) signals for change in group dynamics via transcripts of audio recorded AHF allocation meetings, and 3) understanding how the strategy can be implemented via interviews.

Statistical Analyses

Survey responses were compared pretest versus posttest using the Brunner-Munzel Test.^29,30 The summary statistic for the Brunner-Munzel Test is the Relative Effect(RE), which measures the probability that an observation from one group is in a higher category than an observation in the other group, making this ideal for the analysis of ordinal response variables.³¹ Values of RE range from 0 to 1. The direction of the effect is determined by whether RE is less than or greater than 0.5. To aid in interpretation, all metrics were scaled such that the more beneficial response had the higher value; thus, a RE >0.5 signaled improvement posttest compared to pretest and RE <0.5 signaled worsening posttest compared to pretest. P-values were considered statistically significant at p-value <0.05. All analyses were conducted in R version 3.6.3. Brunner-Munzel tests were conducted using the R package ‘lawstat’.³²

Group Dynamics scores were independently calculated by analysts (NP, MCH) for each DCRDP metric for each meeting using allocation meeting transcripts. Averages were taken when scores varied by one point between analysts. Score differences greater than one point resulted in review of exemplar quotes, discussion with the PI, and consensus on score. Scores for each metric were compared pretest versus posttest using the Brunner-Munzel Test as performed for survey responses by the biostatistician (SK).

Interviews were analyzed using inductive qualitative descriptive techniques³³ by two independent analysts (NP, MCH). Open coding was performed for the entire dataset. The analysts collaborated to combine codes and create major and minor themes. The PI reviewed and verified the coding framework and themes. Trustworthiness criteria including credibility, dependability, and confirmability were addressed through triangulation (comparing results across data and team members), reflexivity (analysts’ memos), and maintenance of an audit trail and codebook throughout the study.³⁴

RESULTS

Participant Demographics

Among 35 eligible AHF team members, on average 81% ($\stackrel{-}{x}$=29, standard deviation 3.2) were present for each of the 8-week SOCIAL HF training sessions (Supplementary Table 1). The proportion was similar to the attendance pattern for AHF allocation meetings prior to study initiation. Surveys were completed by 63% (n=22) of AHF team members pretest, and 60% (n=21) of AHF members posttest (Table 2, Figure 2). Approximately half of participants identified as female for both survey time periods (50.0% pre, 52.4% post). Excluding gender, a third or more of AHF team members self-identified within a minoritized group (pretest 45.5%; posttest33.3%). Participants included a wide range of disciplines.

Table 2.

Participant Demographics

Characteristic	Pretest N=22	Posttest N=21
Gender
Male	10 (45.5%)	10 (47.6%)
Female	11 (50.0%)	11 (52.4%)
Unstated gender	1 (4.5%)	0 (0.0%)
Minoritized Representation*
Not an Underrepresented Group	12 (54.5%)	14 (66.7%)
Underrepresented Race, Ethnicity, or International Resident (Not U.S. Citizen)	6 (27.3%)	5 (23.8%)
Person with a Disability/Lesbian, Gay, Bisexual, Transgender, Queer+/Underrepresented Religion	3(13.6%)	3 (14.3%)
Woman in Male-Dominated
Work group or Man in Female-Dominated
Work group	3 (13.6%)	5 (23.8%)
Unstated Underrepresented group	1 (4.5%)	0 (0.0%)
Discipline
Physician	8 (36.4%)	6 (28.6%)
Nurse Practitioner/Physician Assistant/Nurse Coordinator/Nurse	7 (31.8%)	5 (23.8%)
Social Worker/Dietitian/Financial Coordinator/Pharmacist	2 (9.1%)	4 (19.0%)
Other Discipline	4 (18.2%)	6 (28.6%)
Unstated Discipline	1 (4.5%)	0 (0.0%)

^*Participants with minoritized representation could select more than one category. The survey included the following self-selected minoritized representation categories: not an underrepresented group; underrepresented race; underrepresented ethnicity; international resident; person with a disability; lesbian, gay, bisexual, transgender, queer+; underrepresented religion; woman in male-dominated work group; and man in female-dominated work group. Data were aggregated for groups with one representative to maintain anonymity, with the exception of unstated groups.

Figure 2. SOCIAL HF Enrollment Schematic.

Enrollment schematic for pretest and posttest implementation science study are presented according to Consort Standards of Reporting Trials flow diagram.

Survey Metrics

Across the majority of measures of behavior change, potential benefit was observed posttest (Figure 3a, Supplementary Table 2). Signals of benefit with RE greater than 0.50 were found across all behaviors changes for recognizing bias and speaking about equity, meaning higher signals of the following: want to do, confidently perform, would benefit, less risky, and regularly do this. The majority of behavior changes for challenging stereotypes related to clinical decision-making, assessing environment, and becoming better acquainted with individuals from different backgrounds also had signals of benefit. A minority of behavior changes demonstrated signal for benefit for adopting perspectives of others. However, none of the behavior changes reached statistical significance. Signal of worsening for workplace environment was found with overall workplace environment enforces stereotypes (RE 0.46) and did not reach statistical significance. Signal of benefit for workplace environment was found with satisfaction with leaderships efforts to create collegiality and supportive workplace, which also reached statistical significance (RE 0.69, p-value =0.008 Figure 3b, Supplementary Table 2).

Figures 3a and 3b. Surveys Pre and Post Strategy.

Behavior change and workplace environment were reported on 7-point Likert scale (strongly agree to strongly disagree), and 5-point Likert scale (extremely to not at all), respectively. The 7-point Likert responses are represented by bar graph colors with strongly agree as blue and strongly disagree as red; the 5-point Likert responses are represented by bar graph colors with extremely as green and not at all as red. #, indicates statistical significance with p value ≤ 0.05. Direction of the study signal is illustrated with bold box indicating relative effect >0.5 (positive signal of the strategy) and no box indicating relative effect <0.5 (negative signal of the strategy).

Group Dynamics Metrics

From pretest to posttest, the majority of DCRDP metrics for group dynamics showed signals of improvement (Figure 4, Supplementary Table 2.). Relative effects (RE) greater than 0.5 represent signals of benefit post training as observed with critical opinion sharing (RE 0.67), asking and giving feedback (RE 0.68), challenging groupthink (0.65), openness to mistakes (RE 0.65). Research utilization has a positive signal (RE 0.76, p-value =0.012) and was the only metric reaching statistical significance. Only the metric of experimentation showed a signal of worsening post training as observed with RE less than 0.5 (RE 0.36).

Figure 4. Metrics of Group Dynamics Pre and Post Strategy.

Group dynamics scores were rated 1-4 (best to worst) per the de Groot Critically Reflective Diagnoses Protocol and are represented by bar graph colors with best indicated by green and worst with red. #, indicates statistical significance with p value ≤ 0.05. Direction of the study signal is illustrated with bold box indicating relative effect >0.5 (positive signal of the strategy) and no box indicating relative effect <0.5 (negative signal of the strategy).

Themes for Strategy Implementation

Participants’ coherence of the implementation strategy, SOCIAL HF, was summarized as promoting equitable distribution of a scarce resource in a thoughtful, objective, methodical, and equal manner. Participants believed that the strategy would benefit patients by creating more fairness in the allocation process.

“I think the benefits are that, if we, as a team, are recognizing bias in each other and with our patients, we’re less likely to be [biased] and more —there’s more fairness in our selection process and in our general interactions with patients and each other.”

“It’s obviously an advancement in how we operate, and more standardization can only help us study and improve things.”

Participants’ cognitive participation revealed that SOCIAL HF is a worthwhile investment. Overall, participants were prepared to invest time to make SOCIAL HF work for their center but expressed that additional personnel may be needed to effectively examine bias and change structure of meetings.

“I think it—for some team members, it will increase their involvement and hopefully give them a voice. For other members, it may, uh, let’s say, decrease their influence on the outcome, both of which are probably a good thing.”

“I assume it entails just making sure quiet people are talking. And if someone gets interrupted or calling people out on that and maybe quieting some of the more vocal speakers. And then the changes are also with the social worker and how making the score explicit in what the score means, what the recommendations are, and then having a clear—and then adding the compliance evaluation tool. So that’s in addition. And then clearly specifying what the intervention is and then retesting the patient at a specified time and redoing the score.”

“I think that…to do this well, you need more social work support, and probably psychology support.”

Participants’ collective action demonstrated an appropriate fit of SOCIAL HF with the clinical practice. Participants believed that SOCIAL HF would slow down the allocation process initially and later accelerate allocation.

“So I think, initially, it’s gonna slow it down because we are going to need to be very mindful of our word choice. And like learning anything, it’s gonna take time. I think if we implement this on a monthly and a weekly basis over a long period of time, I think it’ll speed up our allocation process. I anticipate it’s gonna slow things down in the beginning.”

Participants’ reflexive monitoring revealed desires to adapt SOCIAL HF to other areas of clinical care beyond AHF allocation. Participants also recommended that electronic/paper handouts and verbal reminders of key concepts be provided several times a year.

“I do think that what we’re learning here can be adapted into other areas of medical care. It can and it should. If I had to change or if I had to adapt this, I would like everyone to make independent assessments of the patient, beforehand or separate from selection meeting because, sometimes, crowd think, or hierarchy can change people’s minds or influence people.”

DISCUSSION

In a pilot test, we demonstrated how a complex implementation strategy may feasibly contribute to organizational and cultural change in a U.S. AHF center. SOCIAL HF, centered on evidence-based strategies that reduce bias, use standardized numerical assessments of social history, and enhance group dynamics. Pretest-posttest assessments revealed signals of improvement in behavior, workplace environment, and group dynamics. SOCIAL HF was perceived as an improvement on prior allocation processes by promoting equitable delivery of AHF therapies in a more objective and methodical manner. An investment in time and resources was acknowledged. The center believed that additional support for the social worker or psychologist would be helpful to make long-term changes. Implementing SOCIAL HF was perceived to potentially make meetings initially longer until the strategies became habitual, eventually resulting in acceleration. The center perceived that SOCIAL HF could be adapted to other disciplines and reinforced by sending reminders of training semi-annually.

System-level organizational change has been achieved with similar strategies to SOCIAL HF. First, evidence-based bias reduction training programs have demonstrated significant change in self-reported behavior of organizations, resulting in increased diversity of faculty hires^8,9 and medical students,³⁵ and increased satisfaction with career and environment.^8,9 The pilot study of SOCIAL HF also demonstrated multiple signals of behavior change that may contribute to equitable allocation of AHF therapies. Second, training on use of standardized numerical metrics rather than unstandardized descriptive assessments has led to better decision-making and enhanced group dynamics in tumor board literature where multiple disciplines must decide about most appropriate therapy for patients with malignancy.^36,37 The SOCIAL HF recommendation for use of a free, readily available standardized numerical assessments of social history^10,11 may have contributed to signals of improved group dynamics. Third, organizations have observed higher quality decision-making among teams that have equitable opportunity to voice their diverse opinions, and structuring the environment to allow for seating that engages each individual more equitably has increased speed and efficiency of decision-making.^12,13 Similarly, signals of positive change in leadership environment were observed with SOCIAL HF.

This study differs from others by combining multiple evidence-based strategies to address equity in allocation of AHF therapies. Using national studies to identify the most important behaviors for allocation equity,^1,2,7 we mapped behaviors to an evidence-based framework for behavior change components: Capability, Opportunity, and Motivation for Behavior Change/Behavior Change Wheel (COM-B/BCW).²⁰ Each behavior change component has known evidence-based strategies for successfully changing behavior.²⁰ After determining which components matched each behavior, we matched components with evidence-based strategies and scientific frameworks for behavior change per the COM-B/BCW. Then we searched the literature for strategies that either had evidence for use in AHF or may be adaptable for AHF allocation. Thus, changing social bias required mental capacity for change (psychological capability) and reflex-like motivation (automatic motivation); changing unstandardized descriptive assessments of social history required a known toolset (physical opportunity) to change metrics, and group dynamics required change in the social environment (social or environmental opportunity). This combination of strategies may potentiate each other, increasing likelihood of mechanistically changing the factors that contribute to inequity in the allocation process.

The COVID-19 pandemic revealed another major unplanned strength of the SOCIAL HF strategy, adaptability for hybrid in-person/ virtual use. Prior to the pandemic, many AHF centers’ allocation meetings were performed in-person via teleconference in order to accommodate out of office members. Thus, there is familiarity and availability of conference room technology to communicate effectively as a hybrid in-person/virtual meeting. The entire training can be provided virtually even if centers return to in-person allocation meetings. Although one of the strategies is primarily for in-person use, round table setup; the concept is utilized de facto with virtual formats. Virtual formats remove seating hierarchy and may enhance group dynamics in a similar way as round table setup and should not adversely impact pilot findings.

After establishing feasibility, next steps should include a type II hybrid effectiveness-implementation trial in the form of a randomized controlled cluster trial. A powered randomized controlled cluster trial would help identify whether the SOCIAL HF strategy is effective in increasing allocation of AHF therapies to populations that are traditionally less likely to receive therapies such as minoritized racial/ethnic groups and women. We are actively enrolling centers to perform such a study (https://www.clinicaltrials.gov/ct2/show/NCT05390411) that also examines mechanisms for behavior change and the implementation process so that the strategy can be further tailored to achieve AHF equity.

Limitations

As a pilot pretest-posttest study, this study was not powered to assess for effectiveness and could only assess for feasibility and signals of change. In addition, this study took place during COVID-19 at a small AHF center where allocation of AHF was paused prior to initiation of the strategy but decisions regarding care of AHF patients continued. This limited the ability to assess how the strategy might impact implementation during the allocation process in the long-term, particularly with no control site. Group dynamics were assessed using audio rather than video recording given known association with Hawthorne effect and lower consent rates using video versus audio recording.³⁸ Therefore, this study could not assess the role of non-verbal communication such as body language, which could confound findings. During COVID-19, meetings transitioned from in-person to virtual. This may have impacted the evaluation of the decision-making process, but another study revealed that clinical decisions are similar in the in-person and virtual settings.³⁹ Furthermore, the survey responses represented approximately 60% of eligible participants. While this may bias results, this study is strengthened by evaluations at center and individual levels. The interviews represent a sample of multiple disciplines. The group dynamics assessment represent all members of the AHF team. Overall, the pilot study demonstrated potential for benefit and will be better understood in a large multi-center type II hybrid effectiveness-implementation trial.

CONCLUSIONS

Allocation of AHF therapies is vulnerable to bias, subjectivity in social history assessments, and dysfunctional group dynamics of the AHF team. In a pilot study, feasibility was demonstrated in implementation of a multi-component strategy centered on evidence-based bias reduction, standardized numerical assessments of social history, and changes to enhance group dynamics. AHF members found the strategy beneficial, believed that it would improve the process, and recognized resources needed to maintain the strategies. Analyses of the AHF team decision-making processes and individual surveys demonstrated signals of improved behavior, workplace environment, and group dynamics. Future investigation will include a powered randomized controlled cluster trial of the strategy.

List of Abbreviations

AHF

advanced heart failure

AHF therapies

heart transplants and ventricular assist devices

ARMS

Adherence to Refills and Medications Scale

COM-B/BCW

Capability, Opportunity, and Motivation for Behavior Change/Behavior Change Wheel

COVID-19

coronavirus disease

DCRDP

de Groot Critically Reflective Diagnoses protocol

LVAD

left ventricular assist device

RE

relative effect

SIPAT

Stanford Integrated Psychosocial Assessment for Transplantation

SOCIAL HF

Seeking Objectivity in Allocation of Advanced Heart Failure Therapies

Data Availability:

Deidentified data that support the findings of this study are available from the corresponding author upon reasonable request. Because of the sensitive nature of the data collected for this study, request to access the data must come from qualified researchers trained in human subject confidentiality protocols.