Theory-Based Development of an Implementation Intervention Using Community Health Workers to Increase Palliative Care Use

Abstract

Context.

Opportunities for the use of palliative care services are missed in African American (AA) communities, despite Level I evidence demonstrating their benefits.

Objectives.

Single-institution and stakeholder-engaged study to design an intervention to increase palliative care use in AA communities.

Methods.

Two-phased qualitative research design guided by the Behavior Change Wheel and Theoretical Domains Framework models. In Phase 1, focus group sessions were conducted to identify barriers and facilitators of palliative care use and the viability of community health workers (CHWs) as a solution. After applying the Behavior Change Wheel and Theoretical Domains Framework to data gathered from Phase 1, Phase 2 consisted of a stakeholder meeting to select intervention content and prioritize modes of delivery.

Results.

A total of 15 stakeholders participated in our study. Target behaviors identified were for patients to gain knowledge about benefits of palliative care, physicians to begin palliative care discussions earlier in treatment, and to improve patient-physician interpersonal communication. The intervention was designed to improve patient capability, physician capability, patient motivation, physician motivation, and increase patient opportunities to use palliative care services. Strategies to change patient and physician behaviors were all facilitated by CHWs and included creation and dissemination of brochures about palliative care to patients, empowerment and activation of patients to initiate goals-of-care discussions, outreach to community churches, and expanding patient social support.

Conclusion.

Use of a theory-based approach to facilitate the implementation of a multi-component strategy provided a comprehensive means of identifying relevant barriers and enablers of CHWs as an agent to increase palliative care use in AA communities.

Introduction

The management of patients at the end of life raises difficult challenges for health care systems in maintaining the focus on high-quality value-based care.^1–3 Palliative care has been shown to ameliorate these challenges and improve performance on quality measures like satisfaction scores, intensive care and emergency room admissions, and hospital readmissions—all while reducing costs.^4–7 Patients with cancer who receive palliative care early in treatment have also been observed to have better symptom management, higher overall satisfaction with treatment, and potentially improved survival.^8–10 Based on this evidence, the American Society for Cancer Oncology, National Comprehensive Cancer Network, and European Society of Medical Oncology recommend early use of palliative care in the course of disease.^11–13 Despite these guidelines, some studies report usage rates as low as 8.5%-10% in patients diagnosed with cancer.^14,15 In addition, when patients are referred to palliative care services, it often happens much later than guidelines recommend.^16,17

Within African American (AA) communities, palliative care usage is even lower. One 2016 report found that only 8% of eligible AAs received hospice-based palliative care in 2016, compared with 87% of whites.¹⁸ Barriers to palliative care use in AA communities have been found in multiple studies to be multifactorial and include sociocultural (religion, family, and cultural attitudes about death),^19,20 systemic (history of mistrust of medical system),^21,22 and provider-level issues (lack of recommendation by care team and poor patient-clinician communication).²³ Interventions to improve palliative care for minority patients must therefore be delivered through mechanisms that are focused, culturally informed, and sociohistorically sensitive. One such mechanism is through community health workers (CHWs). CHWs are health workers trained to educate patients on health conditions while also helping them access health care and related social services.²⁴ CHWs have previously been espoused in minority communities to deliver critical clinical implementation of evidence-based care in multiple disease settings (cancer, HIV/AIDS, and cardiovascular health) with variable means of delivery (palliative care, screening, and treatment).^25–30 In the palliative care realm, Fischer et al.²⁷ demonstrated in a randomized controlled trial that Latino patients with a CHW were more likely to engage in advance care planning (ACP) (47% vs. 25%) and have earlier pain management discussions (79% vs. 54%) than those without. Similarly, a 2018 randomized clinical trial showed that the incorporation of CHWs increased goals-of-care documentation and patient satisfaction, while reducing health care costs at the end of life.³¹ Besides a 2017 mixed-methods study that showed the benefits of a patient-centered approach to patient navigation,³² there remains a gap in the literature on the base components of any CHW intervention necessary to show effectiveness.

The integration of conceptual frameworks is a unique and contemporary approach to designing reasoned pragmatic interventions to implementation uptake problems.³³ Theories and conceptual frameworks elucidate the interactions between individuals and their environment to yield specific behavior patterns and have been shown to provide critical insight regarding the adjustment of target behaviors.^34,35 Unfortunately, there is a gap in the palliative care literature on using applied conceptual frameworks in the development of interventions to improve the use of palliative care in clinical settings.³⁶ In this article, we describe an approach to a theory-designed intervention to improve palliative care use in AA communities. We combined two theoretical models—the Behavior Change Wheel (BCW) model and the Theoretical Domains Framework (TDF) to both investigate the myriad of factors that can hinder or promote palliative care utilization and also map interventions to address these factors.

Methods

Theoretical Frameworks

The BCW is a framework developed by Michie et al.,³⁷ as a model to synthesize many behavior change theories. At the hub of the BCW framework lies three critical components: capability (C), opportunity (O), and motivation (M), all necessary to produce a behavior (B)—aggregately known as the COM-B model. These three components of the COM-B model form an interacting system comprising an individual’s capability (physical and psychological), opportunity (social and physical), and motivation (reflective and automatic) for behavioral change. Surrounding this central hub are nine intervention functions, and external to that are seven population-level policy categories (Fig. 1). As such, the BCW identifies facilitators and barriers of behavior through the COM-B and links them to multicomponent functions that enable these facilitators and barriers to be manipulated.

Fig. 1.

Behavior Change Wheel, which highlights the capability (C), opportunity (O), and motivation (M), all necessary to produce a behavior (B) model (green), TDF (yellow), and intervention functions (red).³⁶ TDF = Theoretical Domains Framework.

Related to the BCW, the TDF is an implementation science framework that was developed through a consensus process from multiple behavior change theories.^37–39 It comprises 14 domains: knowledge; skills; memory, attention, and decision processes; behavioral regulation; social role and identity; beliefs about capabilities; optimism; beliefs about consequences; intentions; goals; reinforcement; emotion; environmental context and resources; and social influences.³⁹ Each of these domains corresponds to a component of the BCW (Fig. 1), and thus, combining the two theories allows for expanding the COM-B components into specific useable domains.⁴⁰ We chose BCW and TDF over other implementation models because they directly link targets for behavior change with interventions that can bring about the change of interest (e.g., a particular behavior linked to legislative interventions vs. social planning).

Study Setting and Collection of Qualitative Data

During the course of 12 months, five stakeholder focus groups were conducted at Johns Hopkins Hospital. Each focus group comprised 13–15 participants and was attended by the same individuals, except for the replacement of two patients who passed away after earlier participation. As per the seven Ps taxonomy by Concannon et al. for stakeholder engagement,⁴¹ each stakeholder engagement session comprised the following members: providers (surgical and medical oncologists, CHWs, and palliative and hospice clinicians); patients (with advanced solid organ malignancies) and their caregivers; purchasers/hospital administrators; policy makers; and principal investigators (quality, implementation, and palliative clinical care researchers). We did not include the remaining two Ps of the taxonomy (payers and product makers) for this study as their participation was not deemed necessary to the scope of the question. Participants were invited by the lead author or their oncologist to participate across longitudinal stakeholder engagement sessions. Key informants were identified based on their preidentified clinical and academic involvement in palliative care, cancer care, or patient navigation. A semistructured guide was prepared in advance by the study team. In line with the aims of the study, the guide focused on identifying the behaviors that most consistently impact palliative care use and the viability of CHWs as a mechanism to address these behaviors in AA communities, culminating with stakeholder input regarding the developed implementation interventions. Focus groups continued until saturation was reached, and this occurred during the course of five engagement sessions. Engagement sessions were 45–90 minutes long, audio recorded, and transcribed verbatim by study team members (B. K. and K. T. L.). Anonymization and deidentification of participant comments were accomplished by assigning a numerical code to each stakeholder during the transcription process. The study was approved by the Johns Hopkins Medical Institution Institutional Review Board.

Coding and Analysis

Thematic analysis was performed on the transcripts of the five audiotaped engagement sessions by two investigators (B. K. and Z. O. E.) using consensual qualitative research methods.⁴² Grounded theory was then used to expound on identified psychosocial phenomenon based on the thematic analysis and create larger abstract themes and theories about promoters and barriers to palliative care. We chose to use grounded theory because this article’s topic is a research area where major gaps in understanding exist, and grounded theory uses concepts and categories to explain patterns of a psychosocial phenomenon when this is the case.^43,44 NVivo (NVivo qualitative data analysis software; Version 10, 2014; QSR International Pty, Ltd., Burlington, MA) was used to facilitate data management and organization. Differences in analysis between independent coders were reconciled by consensus; and to enhance credibility and validation of data analysis, member checking was conducted.⁴⁵

Implementation Intervention Development

We organized our findings from transcripts to map to the three prediscussed components of the COM-B.^33,46,47 Based on the intervention functions of the BCW (Fig. 1), 10 behavior change strategies were chosen to influence capability, opportunity, and motivation of both patients and providers, as illustrated in Table 1. After deliberation of strengths and weaknesses of proposals, each stakeholder was asked to rank the 10 implementation strategies in order of most important (Number 1) to least important (Number 10) during the fifth focus group discussion (FGD). The data were then compiled, and each ranking was weighted, culminating in an aggregate score for each named strategy in Table 1. This fifth FGD was the culmination of the first four sessions, focusing directly on the intervention after gaining theme saturation from the previous four sessions. For complete and transparent reporting of our research, we used the 21 Standards for Reporting Qualitative Research checklist to guide our reporting, and it is included as an appendix.⁴⁸

Table 1.

Selection of Behavior Change Techniques and Implementation Change Strategies, With Stakeholder Sample Quotes

COM-B Factor From the BCW	Modifiable Determinant of Behavior, i.e., Barrier or Facilitator	Sample Quotes	Theoretical Domain and Techniques for Behavior Change from TDF-Guided Analysis	Implementation Strategy
Capability	Barrier: Unaware of the differences between palliative care and hospice care	Everybody thinks when you say palliative care, they usually think the patient is at the seven stages of death	Knowledge  • Provide information about differences between palliative care and hospice care	• CHWs will distribute brochures developed to explain palliative care and its differences from hospice care
	Barrier: Physician unfamiliarity with CHW roles may limit respect for recommendations	A part of creating a structure is empowering [CHWs] and having the respect for them from the clinical side	Behavioral regulation  • Demonstrate a willingness to empower and respect CHW opinions during clinical encounter	• CHWs will provide assessment of patient satisfaction with physician care and communication and deliver feedback with focus on patients’ goals of care
	Barrier: Lack of physician ability to effectively navigate conversations with patients about palliative care	I’ve seen some of the best clinicians stumble at what really comes down to you’re only going to be here for another week so what do you want to do this week?	Cognitive and interpersonal skills  • Rehearse the use of communicative strategies  • Model and set goals for the use of communication strategy in palliative care conversations  • Solicit feedback from CHW regarding communication strategy	• Three-part communication strategy for patients  1. Ascertain patient’s goals of care and wishes  2. CHW to assist patient/caretaker to set agenda in consultations with physicians  3. Establish plans for follow-up and advocacy via motivational interviewing if empowerment unsuccessful
Opportunity	Barrier: Inability of patients to prioritize health care because of other more pertinent life needs	I’ve had members tell me I do not really care for my health today because my BG&E is about to get cut off	Environmental context & resources  • Provide social service resources to low-income patients and patients with weak social support	• CHW to provide resources in concordance with their standard roles by assessing social determinants of health via a checklist
	Facilitator: Being surrounded by a social circle that encourages the use of palliative care, especially in religious and familial settings	Sometimes it’s not patients, it could be their mother or their husband. My experience has always been in church that ministers forever have preached the fact that Jesus is the doctor	Social influences  • CHW empowerment of church leaders to discuss palliative and end-of-life care  • Physician discussion of palliative care with family members present	• Conduct monthly educational outreach visits with CHWs during church and community activities
Motivation	Barrier: Low physician confidence to cease disease modifying treatment when necessary and deal with hesitant family members	I’ve got something else for you, I’ve got something else for you, and so in a way we’re waiting for the patient to wave the little white flag	Beliefs about consequences  • Empowering patients to share decisions regarding their health	• Train CHWs to initiate ACP conversations and track uptake through number of completed ACP documents and ACP conversation with family and caregivers
	Barrier: Patient unwillingness to discuss palliative care because of fears about the implication of conversations for prognosis and subsequent treatment	As far as everybody is concerned that’s like a sentence to go die. So that’s a no-no. Then the patient thinks that’s like let me go die at my house	Beliefs about consequences  • Use goal setting, problem solving, and education for reassurance	• Begin palliative care discussions within first two visits postreferral to CHW (i.e., prognostic awareness, goals of care)
	Barrier: Unrealistic patient expectations about treatment and prognosis reduce patient motivation to use palliative care	They have this type of mistaken belief in their God or pastor or something	Optimism  • Action planning for patient treatment (setting early goals of care and incorporating palliative care as a viable part of that care)	• CHW meetings with families to discuss their understanding and beliefs and relay to physicians and/or obtain necessary palliative and end-of-life care resources

COM-B = capability (C), opportunity (O), and motivation (M), all necessary to produce a behavior (B); BCW = Behavior Change Wheel; TDF = Theoretical Domains Framework; CHWs = community health workers; ACP = advance care planning.

Results

Participant Characteristics

Participants included three AA patients with advanced malignancies (lung, liver, and colon primaries); one caregiver (AA); three hospital administrators (one AA); four clinicians (two hospice and palliative medicine) and two oncologists (one Latina); and three CHWs (two AAs and one Latina). About 53% of stakeholders were males, and 46% of the participants were AAs.

Themes and Constructs

Regarding barriers of palliative care, we organized emergent themes under the sources of behavior in the COM-B and their relevant domains in the TDF and present them later. The intervention functions targeting these themes are presented in Table 1, along with the themes to which they map.

Capability: Knowledge, Behavioral Regulation, and Cognitive and Interpersonal Skills

Various stakeholders expressed the concern of a general lack of knowledge and conflation of palliative care with hospice care, citing that palliative care is often mistakenly described as synonymous with end-of-life care by both physicians and patients alike. Participants discussed how these views, although broadly a barrier for palliative care, still affected cultural groups differently. For example, AA patients who conflated palliative care with hospice were described by stakeholders as more likely to be concerned that the health care system was trying to withhold care from them and subsequently more likely to be resistant to palliative care. Stakeholders reported similar confusion among care providers and discussed that this was a significant health system-level barrier, as care providers were often responsible for referring patients for palliative care. Physician stakeholders shared that they sometimes developed an emotional attachment to long-term patients, which led to a sense of guilt and abandonment when these patients became candidates for palliative care referral. Conversely, physician lack of emotion and the personal touch required for these conversations were also seen by patients and caregivers as barriers to palliative care. These comments, as well as comments regarding physician propensity to rush through the softer sides of medicine, revealed that inadequate cognitive and interpersonal skills was a significant barrier for palliative care. Stakeholders linked this lack of empathy to time constraints placed on physician interactions with patients. Because of this, stakeholders emphasized flexibility as a key characteristic of individuals selected and trained as CHWs. Seeing patients beyond regular work hours and pitching in however possible, according to stakeholders, were all important steps to begin bridging knowledge gaps and enhancing behavioral regulation toward a desired behavior.

Opportunity: Environmental Context and Resources As Well As Social Influences

Family, religion, and lack of sociohistorical relatability between physicians and patients arose as themes for barriers to palliative care. Stakeholders emphasized that for palliative care implementation to be effective, the entire family has to be considered in interventions. Comments revealed that attitudes about palliative care were often informed by larger social institutions like religion, with some patients being reported to ascribe to the philosophy that Jesus is the real doctor. Stakeholders also reported past history with research as an especially important barrier to palliative care for AA patients, with one citing the Henrietta Lacks story as an example of medical mistreatment of marginalized communities.⁴⁹ Overall, comments from stakeholders highlighted the need to work a little bit harder with AA patients who often had slightly more resistance in establishing the opportunity for palliative care conversations. Stakeholders felt that it was equally important that CHWs understand both the clinical issues that a patient is facing, as well as their social environment, to best help them through the complex medical system and maximize opportunities for palliative care use. The word community in CHW was emphasized in many interviews. It was deemed vital that a CHW be someone familiar with the community and embedded in the community and thus aware of the history and culture of the patient population.

Motivation: Beliefs About Consequences and Optimism

Low provider motivation to cease disease-modifying treatment because of a fear of damaging the therapeutic relationship arose as a barrier to palliative care. Comments revealed that providers often relied on families and caregivers to wave the little white flag before proceeding to palliative care referral. Furthermore, patient belief that palliative care discussions implied a sentence to go die was also identified by stakeholders as an important barrier in palliative care use. The desire to maintain high optimism was described by patient, physician, and CHW stakeholders to lead to unrealistic patient expectations regarding prognosis and available treatment options. This was seen by stakeholders as especially important in palliative care interventions for AA patients. They hypothesized that historically informed trust issues about medicine in AA communities could be alleviated by a CHW who is not a nurse or a clinician. For stakeholders, a CHW that could culturally understand, relate to, and advocate for the patient was critical in modulating unrealistic expectations from patients. Comments expressed that navigators could help eliminate patient misconceptions about palliative care while simultaneously providing an educational benefit for physicians by helping expand their general understanding of the communities within which they work.

Stakeholder Assessment of CHW Intervention

Behavior change strategies were chosen to influence capability, opportunity, and motivation of both patients and providers, as illustrated in Table 1. To increase capability, the intervention will improve patient knowledge, expand CHW roles, and improve physician communication and respect for palliative care recommendations. Capability will also focus on increasing the confidence of physicians in the initiation of these discussions. Opportunity to use palliative care will be increased through CHWs engaging directly with patient social circles through family and churches and CHWs assisting patients with navigating extraclinical problems. Furthermore, strategies to increase opportunity will include timely check-ins with patients to identify palliative care need as early and frequently as possible. Goal setting, auditing, and feedback of patient improvement (or lack thereof) will be used to increase motivation of both patients and providers for palliative care discussions. Notably, these strategies have been combined to be delivered in a titrated manner through practice and small tests of change facilitated by CHWs. These approaches were chosen because as a constellation, they target multiple aspects of the BCW framework that influence human behavior and subsequent patient outcomes across different settings.

After deliberation of strengths and weaknesses of all 10 intervention proposals (Table 1) during the fifth FGD, the five most important strategies according to stakeholders were as follows: 1) three-part communication strategy for patients; 2) CHW meetings with families to discuss their understanding and beliefs and relay to physicians and/or obtain necessary palliative resources; 3) CHWs should begin and complete palliative care discussions within first two visits postdiagnosis; 4) train CHWs to initiate ACP conversations and track the number of completed ACP documents; and 5) CHWs will develop and distribute brochures explaining palliative care and its differences from hospice care.

Discussion

Using a stakeholder-engaged process, we were able to refine an intervention that targets the 14 theoretical domains of behavior change. Our data revealed that patient recognition that palliative care is not mutually exclusive from therapy and treatment is a critical facilitator of their ability and willingness to heed provider recommendations for palliative care use. It also identified the lack of understanding of the pillars of palliative care as a barrier. As such, one strategy of our intervention is for CHWs to disseminate simple but informative brochures that would outline the specific differences between palliative and hospice care to be discussed with patients deemed suitable for palliative care referral. Similarly, lack of provider emotional intelligence and interpersonal communication skills were perceived as critical barriers to their capability to recommend palliative care and address patient hesitancies. Thus, a part of our intervention dispatches CHWs to augment these limitations in communication using a three-part communication strategy as described in Table 1. In bridging this emotional, and sometimes, cultural chasm between providers and patients, CHWs would increase both the opportunities for discussion of palliative care and the rate at which existing opportunities are capitalized on.

The importance of strong physician interpersonal and cognitive skills in palliative care discussions—an emergent theme from our stakeholders—is well corroborated in the literature. For example, a 2005 mixed-methods study found that ineffective communication by physicians was a major barrier to the acceptance of patients and caregivers of an irreversible condition.²³ Our proposed intervention has several innovative features, such as improving the quality of provider communication through CHW consultation and recommendations as well as the aforementioned communication strategy. In an ideal world, basic palliative care competencies would be achieved during medical school and residency,⁵⁰ but our intervention recognizes shortcomings in educational infrastructure and augments failures in training with the implementation of our three-part communication strategy (Table 1). Although our strategy gives CHWs the bandwidth to initiate and provide strong recommendations for palliative care discussion, it still allows providers to maintain leadership and autonomy in discussions about palliative care, as this has been found to develop and sustain motivation⁵¹ and will therefore keep physicians engaged in palliative care dialogue.

The environmental context (culture, religion, and family) of a patient and the resources available to them can either hinder or facilitate opportunities for palliative care use. Given the low representation of AA practitioners in our physician workforce,⁵² it is unrealistic for AA patients needing palliative care to be only managed by AA physicians. By using CHWs who hail from backgrounds similar to the patients, we envision our implementation strategy expanding physician knowledge on race, religion, and culture to capitalize on the benefits that come with racial concordance in a care team.^53,54 Furthermore, by conducting outreach visits at churches and community activities, CHWs will expand opportunities for group-scale education about the importance of palliative care in the continuum of care. This element of our implementation strategy may begin to reduce some of the aforementioned disparities in AA communities, who have higher symptom burden, caregiver distress, and medical costs.⁵⁵

Patient belief about the consequences of palliative care conversations is a well-described barrier to palliative care use, especially in AA communities. However, recent data have shown that the earlier palliative care consult is associated with greater family satisfaction and can help reduce anxiety regarding death.⁵⁶ Based on the BCW behavior source of Beliefs about Consequences, our intervention seeks to promote prognostic awareness both by initiating palliative care discussions within the first two visits with a CHW and by encouraging family goal setting. Because a false belief about consequences inherently implies a deficit in knowledge, there will be an education component in all meetings with a CHW at this stage. Thus, the burden of education about palliative care, often thought to rest on physicians,⁵⁷ will be shared with CHWs who can spend more time and have more access to educate patients in a tailored nonintimidating manner.

We chose to use CHWs to deliver our multicomponent implementation strategy because of the versatility and flexibility of their socioprofessional identity, providing us with increased bandwidth to apply our proposed strategies to dynamic clinical settings. Although our proposed utilization for CHWs expands on traditional roles, existing data support our confidence that this is a viable and adaptable form of implementation delivery.⁵⁸ Still, our study has several limitations. Our sample consists only of a small number of stakeholders and thus their observations may not be representative of entire populations, although it is well corroborated by the literature describing palliative care experiences in large-scale academic practices. This literature not only indicates that palliative care is underused but cites social, emotional, and cognitive factors as the major barriers to this end.^59,60 In addition, given the size of our study population and number of subgroups represented (e.g., patients, providers, CHWs), we were not able to do any comparison across groups that would offer a more nuanced perspective on the highlighted results and conclusion in this study. There is also potential transcriber bias as well as the potential of a recorder influencing what participants share on tape.

Conclusions

Interventions to increase palliative care use are urgently needed, especially in AA communities. In this study, we used the TDF and BCW to systematically investigate the barriers to use of palliative care services, identify patient and physician target behaviors for change, and design a pragmatic intervention to be implemented based on these target behaviors. By using a multicomponent implementation strategy, we hope that the interventions will be better tailored and more sustainable in increasing use of palliative care services for AA patients.

Key Message.

In this stakeholder-engaged qualitative study, community health workers were found to be viable delivery agents of a theory-informed implementation strategy to increase palliative care use in African American communities. The benefits of early palliative care referral, well described in the literature, can thus be conferred to historically marginalized communities in a more equitable manner.