Implementation Barriers and Recommendations for a Multisite Community Health Worker Intervention in Palliative Care for African American Oncology Patients: A Qualitative Study

Abstract

Background:

Palliative care remains underutilized by African American patients with advanced cancer. Community health workers (CHWs) may help improve palliative care outcomes among this patient population.

Objectives:

To explore barriers to success of a proposed CHW intervention and synthesize design and implementation recommendations to both optimize our intervention and inform others working to alleviate palliative care disparities.

Design:

Semi-structured qualitative interviews.

Setting/Subjects:

Key informants were health care professionals across clinical, leadership, and community health fields. Participants were recruited through purposive sampling from Baltimore, Maryland; Birmingham, Alabama; and Salisbury, Maryland.

Measurements:

Interviewers used an interview guide grounded in established implementation science models. Data were analyzed through a combined abductive/deductive approach by independent coders. A framework methodology was used to facilitate thematic analysis.

Results:

In total, 25 professionals completed an interview. Key informants discussed multiple barriers, including at the patient level (lack of knowledge), clinician and facility level (decreased workflow efficiency), and health system level (limited funding). Recommendations related to the intervention’s design included high quality preintervention CHW training and full integration of CHWs into the care team to “bridge” divides between outpatient, inpatient, and at-home settings. Intervention delivery recommendations included clearly defining care team roles and balancing flexibility and standardization in CHW support approaches. These recommendations were then used to adapt the planned intervention and its implementation process.

Conclusions:

Clinicians, cancer center leaders, and CHWs identified multilevel potential barriers to the intervention’s success but also described recommendations that may mitigate these barriers. Key informant input represents an important step prior to initiating CHW-based interventions.

Key Message

This qualitative study investigated a proposed community health worker intervention to improve palliative care outcomes among African Americans with advanced cancer. Health care professionals identified multilevel potential barriers to success and recommended design and delivery modifications based on their diverse experiences across three clinical sites.

Introduction

Palliative care focuses on improving quality of life, alleviating symptoms, and ensuring goal-concordant care. The American Society of Clinical Oncology guidelines recommend that all patients with advanced cancer receive dedicated palliative care services.¹ However, palliative care remains underutilized, especially among African Americans and other minority groups within the United States.^2–5

The medical community’s growing acknowledgment of the importance of palliative care has led to interest in improving its provision. A 2018 review noted over 60 multicomponent palliative care interventions developed since 1980.⁶ Despite this growth, substantial gaps in care persist and new approaches are necessary, especially in light of palliative care workforce shortages.⁷ Community health workers (CHWs) are lay individuals who combine a deep cultural understanding with job-related training in their work on the front lines of health care.⁸ CHWs have previously been studied within oncology, including among minority populations. However, most CHW interventions have focused on cancer screening and prevention rather than support of patients with advanced cancers.^9–11 More recently, both a pilot study of a lay navigator-led coaching intervention designed to improve distress and quality of life among African American and/or rural caregivers of individuals with advanced cancer¹² and a CHW-led intervention focused on the goals of care documentation and quality of life among low-income racial and ethnic minority adults newly diagnosed with cancer have shown encouraging results.¹³

Our current series of investigations seeks to improve palliative care outcomes among African American patients with advanced cancer. Our previous work piloted an intervention that incorporates CHWs to enhance palliative care outcomes through patient education, support, and empowerment.^14,15 In brief, the intervention begins with a standardized assessment of social determinants of health and access to care and then allows individualized CHW support in accordance with patient and caregiver needs moving forward. Following the principles of the Consolidated Framework for Implementation Research (CFIR) and Reach, Effectiveness, Adoption, Implementation, and Maintenance (RE-AIM) framework,^16,17 we designed our intervention to be feasible and effective across multiple settings. Prior to beginning a multisite randomized controlled trial, known as the Dissemination and Implementation of a Community Health Worker Intervention for Disparities in Palliative Care trial,¹⁸ we sought to identify any remaining potential barriers to success and gather recommendations on our intervention’s design and implementation from professionals at each site. We then adapted our intervention’s design and implementation accordingly. In this report, we share our findings to inform other palliative care and health care disparities investigators working to design community-based programs that may face similar barriers or employ analogous processes.

Methods

Study design, participants, and setting

This study comprised 25 semi-structured professional key informant interviews that explored the potential integration of CHWs into care teams treating African American patients with advanced cancer. Key informants were classified into four categories: (1) CHWs or other care navigators; (2) palliative and hospice care providers; (3) medical, surgical, or radiation oncology providers; or (4) cancer center leaders and administrators. Informants were recruited from the three trial sites: the Johns Hopkins Hospital in Baltimore, Maryland; the University of Alabama at Birmingham Hospital in Birmingham, Alabama; and TidalHealth Peninsula Regional in Salisbury, Maryland. We employed purposive sampling based on the recommendations of site principal investigators (PIs) and added a snowball approach for further participant recommendations. Potential participants were approached via email to offer information about the study and a $50 gift card honorarium for interview completion.

Each interview was conducted by one of three authors (A.S., A.P.W., or O.M.). All interviewers were female, trained as a medical doctor, and affiliated with Johns Hopkins University as a researcher. Interviewers had completed coursework and/or instruction in qualitative research methods from experienced qualitative researchers (S.M.F. and J.O.). No participants had a previously established relationship with their interviewer.

Data collection

Interviews were conducted from October 2022 to February 2023 using a secure version of the Zoom videoconferencing platform (Zoom Video Communications, Inc., San Jose, CA). A copy of the consent form was provided for advance review, and verbal consent was obtained immediately before the interview. Each interview lasted for 60–90 minutes. Participants were given the option to use video or audio only; most elected to use video.

Interviews were conducted using a semi-structured interview guide created collaboratively by the interviewers, the PI, and two experienced qualitative researchers. Topics and questions were grounded in the domains of the CFIR and RE-AIM frameworks. The CFIR is a commonly used typology within health services research and focuses on five major domains: the intervention, inner setting, outer setting, individuals involved, and implementation process.¹⁶ RE-AIM is a methodology to evaluate multilevel public health interventions across five categories: reach, efficacy, adoption, implementation, and maintenance.¹⁷ The same guide was used for each category of professional key informants, with questions adapted as needed (Supplementary Appendix). Interviews explored key informants’ experience with palliative care, patient navigation, and CHWs; opinions on their institutions’ abilities to implement change; and thoughts on barriers, facilitators, and recommendations related to a proposed intervention. Demographic survey questions were asked at the interview’s conclusion. Demographic data were entered into REDCap electronic data capture tools hosted at Johns Hopkins.^19,20 Field notes were made after interviews to reflect on interview logistics, describe the interviewee’s setting, and capture the interviewer’s own thoughts on the interview process (incorporating the principle of reflexivity). Study site team visits were performed by the central research team to obtain general knowledge of each site prior to final data analysis.

Data analysis

Interviews were professionally transcribed and reviewed by the research team for accuracy. The coding approach combined abductive and deductive elements, seeking to balance inductive codes derived from source data with both theoretical knowledge and deductive codes based on known themes from the interview guide.²¹ A preliminary codebook was developed based on initial transcripts from several key informants and then refined after independent preliminary coding and subsequent comparison among analysts (A.P.W., O.M., and S.M.F.) to ensure consistency. Primary and secondary coding were then completed for each transcript by two analysts. The coding authors regularly met to discuss newly coded data to maintain rigor in our analytic process. Through an iterative process of discussion and review of analytic tables, our team determined that data saturation was reached given that no new major or minor themes emerged. MAXQDA 2022 software (VERBI Software, Berlin, Germany) was used for data management during the coding process.

Coded data were analyzed using the framework method.²² First, the codes most relevant to the analysis were determined by consensus. These codes fell under four major content areas that later became the thematic framework for our analytic table: (1) participant attitudes about the CHW intervention; (2) gaps and challenges related to the intervention’s potential implementation; (3) site-specific considerations; and (4) specific recommendations for the design, implementation, and delivery of the intervention. All interview segments with relevant codes across all participants were combined into a summary document. This document was rereviewed for content and each coded segment was summarized in bullet point format. Potential illustrative quotations were also identified in this phase. Finally, summaries were condensed further into the analytic table to facilitate comparison between participants. Barriers and recommendations were then derived from the analytic table.

Human subjects protections and reporting guidelines

This study was approved by the Johns Hopkins University School of Medicine Institutional Review Board (#00283002). It is reported using the Consolidated Criteria for Reporting Qualitative Research (COREQ) guidelines.²³

Results

Participant recruitment and characteristics

In total, 33 potential key informants were approached via email. Interviews were completed with 25 key informants. The eight informants who did not complete an interview did not reply to initial or scheduling emails (n = 4), referred us to a contact perceived to be more relevant (n = 2), or possessed areas of expertise for which recruitment was already sufficient (n = 2). Participant characteristics are displayed in Table 1.

Table 1.

Key Informant Characteristics

Site, n (%)
Johns Hopkins Hospital	8 (32)
TidalHealth Peninsula Regional	8 (32)
University of Alabama Birmingham Hospital	9 (36)
Key informant category, n (%)
Community health worker or care navigator	8 (32)
Palliative care clinician	6 (24)
Oncology clinician	6 (24)
Cancer center leader or administrator	5 (20)
Age in years, median (IQR)	46 (40–60)
Gender, n (%)
Female	20 (80)
Male	5 (20)
Race/Ethnicity, n (%)
Non-Hispanic White	15 (60)
African American	5 (20)
Other (Hispanic White, Asian, other, multiple)	5 (20)
Education level, n (%)
High school	1 (4)
Associate’s degree or “some college”	5 (20)
Bachelor’s degree	2 (8)
Graduate degree	17 (68)
Years of experience in current role, median (IQR)	7 (3–10)
Years of experience in health care, median (IQR)	21 (14–29)

IQR, interquartile range.

Barriers to successful implementation of the CHW intervention

Barriers at the patient level

Key informants identified several barriers to recruitment for and acceptance of the intervention at the patient level; illustrative quotations are shown in Table 2a. One major barrier described was lack of knowledge, including low overall health literacy as well as specific confusion surrounding palliative versus hospice care. Another barrier was emotional readiness to focus on symptom management as a major goal alongside or in lieu of disease treatment. Participants also identified logistical barriers, including perceived financial burden and the demands of managing cancer care among competing priorities such as childcare and work. Negative community and personal experiences with the health system, whether occurring in the past or as an ongoing issue, were also described as barriers. Finally, key informants anticipated a challenge in identifying patients who felt sick enough to be interested in discussing palliative care but remained well enough that they could complete study follow-up.

Table 2.

Barriers to Implementation: Illustrative Quotes from Key Informants

a. Patient-level barriers
Lack of knowledge: “Knowledge of it is probably the biggest [barrier]. Nobody really knows that it exists or that it’s out there until their oncologist or their inpatient hospitalist says, ‘I’m going to consult palliative care … And they get scared to meet with us. Or, you know, there is a huge misconception that it’s the same as hospice.”—Palliative care clinician (participant 13)
Readiness to focus on symptom management: “I’m thinking about a particular instance where the patient, very advanced, came in in a wheelchair. Had been admitted, I don’t know, it was many, many times … And they had heard about palliative care from multiple other doctors … So I think that the barrier was probably that the denial, you know, that that was where that needed to be.”—Oncology clinician (participant 16)
Logistical barriers and competing priorities: “They already have so many appointments. They’re going here and there and everywhere, and this is just another thing that they have to go do, and, ‘Why can’t my oncologist manage all these things? Why do I have to go talk to another person about the same issue?’”—Cancer center leader (participant 11)
Negative experiences within the health system as a community or individual: “Whether it is implicit or not, unconscious, or not, we know that black patients and families are not getting the same care, period. And so where is that happening in the system? I think a place that I see it is around decision making and sort of what is being offered and assumptions that are being made around what social support people have or what resources people have.”—Palliative care clinician (participant 19) “It could be any patient who has a background in drug use and they will naturally assume that they are drug-seeking. You know, they are not looking at the fact that they have cancer. Or any other illness that may cause them to be in a lot of pain.”—Community health worker (participant 4)

b. Clinician and facility-level barriers
Concerns about potential for disruption to existing clinical workflows: “We were in the way. We were in the way with the social workers … The doctors—we did develop relationships with the doctors one-on-one. But the barrier was that you’re in the way … And even the front desk, we were in their way. ‘You’re slowing up progress … Just wait. Let me get this person checked in.’”—Community health worker (participant 21) “I think if they cause me more work because they didn’t understand, and they didn’t have the education they needed or the training they needed to be successful, and they were causing me chaos, then … the program would fail. But if you had a really good [CHW] who really provided good support and was able to think independently and, you know, figure things out, then yeah, I think it’d be great.”—Oncology clinician (participant 9)
Addressing complex medical and psychosocial needs: “Some counseling element, people always have lots of issues and anxiety, some just distress to some degree about all the different aspects of cancer—‘How do I manage this with my job? What about my caregiving?’…some training around those aspects I think would be key. Providing just some oncology background I think obviously would be helpful. That’s a challenging part because you can’t do hundreds of [hours]—I mean how much is enough? I don’t know.”—Oncology clinician (participant 18) “I had a patient’s husband ask me to call back yesterday and I called back, the patient had fallen … the longer I spent on the phone with him … what he was describing is what I think of as failure to thrive. The woman also has more than one cancer … So I don’t know that a lay navigator would have necessarily picked up—she may have known to recommend the emergency room, but she may have not picked up on what to communicate to the physician level.”—Care navigator (participant 20)
Oncology provider knowledge and acceptance: “We used to have an oncologist—he still works there, but he’s better now with palliative care—who would not refer anybody to us, and still we have to text him and message him before we see a patient and ask his permission, but at least now he lets us see people … he has made the comment to my physician and to several of us, ‘Well, that's just giving up.’”—Palliative care clinician (participant 15)
Potentially underutilized navigation programs may already exist: “We had a lay navigator briefly [and] she actually would come and sit down in our office. We don’t really even know 100 percent what she does. There were lay navigators. One thing I think they're supposed to do is find out the patient's distress level up front and I'm not sure who they notify of that because those aren't coming to me.”—Care navigator (participant 20)

c. Health system-level barriers
Challenges with funding and cost-effectiveness: “I think we need to show that it may be more cost, because it will be more cost, but [that] in the long run it’s cost-effective, and it promotes quality of life. It’s going to decrease health care utilization because … I mean unfortunately that’s the system we have. It boils down to what it’s going to cost us … I think until we demonstrate with data and comparative effectiveness, I think it’s going to be very hard to show the value to the health care system.”—Cancer center leader (participant 24) “If we had more staff, [patients] could be seen more quickly because [providers’] schedules wouldn’t be already so full. So I think that’s one thing, and obviously, that comes with funding and justifying the resource. A lot of times, these services are not reimbursable, and so it’s not often revenue generating, and non-revenue generating activities are often very hard to resource.”—Cancer center leader (participant 23)
Limited access to palliative care services: “So many patients with solid tumors that are admitted are admitted for palliative care issues … The single palliative care team is just extraordinarily overwhelmed, and I think that it really contributes to a lot of suffering and a lot of delay in care. The volume is so much that the palliative care team can only go by and see the patient and titrate like once a day because they’re—they’ve got a huge number of people to take care of … and then that’s magnified by the acuity of the problem, right? It’s not just like, ‘Hey, I’ve got some nausea that I’m not handling well,’ but it’s like it’s so bad that a person has to be hospitalized. Plus, we are not able to make as quick adjustments as we might want to make.”—Oncology clinician (participant 8)
Care fragmentation: “The other thing is integrated care. Our health care is very fragmented, very fragmented, and then you just do referrals. This patient goes there, and what I see a lot from where I sit in the leadership of the cancer center, is the patient does not have a quarterback, and who is going to be the quarterback for that patient to be sure to integrate out that information? … So I think that’s from the system level.”—Cancer center leader (participant 24)

Barriers at the clinician and facility level

Many implementation barriers at the clinician and facility level were rooted in preserving and working within the usual clinical flow (Table 2b). Clinicians generally perceived themselves and others to be highly receptive to the intervention’s goals but also expressed hesitancy about a program that could increase their workload or decrease their efficiency. In addition, some key informants were concerned that lay CHWs would be unable to sufficiently address complex medical and psychosocial needs and relay their concerns to treating clinicians efficiently and effectively. Less commonly, oncology provider knowledge and acceptance of palliative care was felt to be a barrier. Key informants also expressed concern about potential overlap with existing programs. Though it was not explicitly labeled a barrier, clinicians at times knew that a navigation program existed elsewhere within their hospital but were consistently unable to describe its details and usage patterns, suggesting that existing navigators may be underutilized.

Barriers at the health system level

One major system level barrier to implementation noted by key informants was the lack of ongoing funding for CHWs beyond the study period (Table 2c). Several informants expressed skepticism that the intervention could ultimately be cost-effective. Cancer center leaders emphasized a lack of reimbursement for many palliative care services. A second major barrier described was access to palliative care providers; many key informants reported extended wait times in their current settings and described a general shortage of professionals trained in palliative care. Additional barriers identified included care fragmentation, such as when disparate specialists did not discuss a patient’s global picture or when patients received aspects of their care within other systems that were not study sites (e.g., patients receiving chemotherapy closer to home who had initially traveled to see a surgical specialist).

Key informant recommendations

Intervention design recommendations

Key informants hoped to fully integrate CHWs into the care team and have them serve as a “bridge” between the inpatient and the outpatient setting; illustrative quotes are displayed in Table 3a. Many informants, including CHWs themselves, identified CHWs’ ability to visit patients in their own homes as an advantage that was otherwise missing from existing care processes. Informants expressed support for the intervention’s design, which included a baseline assessment of social determinants of health and access to care, followed by individualized support thereafter.

Table 3.

Implementation Recommendations: Illustrative Quotes from Key Informants

a. Intervention design
Leveraging the ability of CHWs to perform home visits: “[Inside their homes,] I think most people give you more feedback, and being right there in a person’s presence is better than over the telephone. That way, they can see you. You can see them and their body language.”—Community health worker (participant 12) “If they kept missing appointments, then the doctor would have me go to the house and find out what are the barriers that’s keeping them from coming to the clinic … let me give you an example. I had a patient who kept coming in and was saying that her medicines were missing … we went to the house, and found out that her son, her daughter’s boyfriend, and the daughter were taking her medicine. So, we went to the house and got him connected with the program, you know, so he could stop taking her medicine. So sometimes you go to the house, you find out it's more than just the patient who needs help.”—Community health worker (participant 4) “There’s a huge need for home palliative care that looks different than it would in the hospital or in the clinic, even … there’s still a huge need for community health workers. Again, it gets back to, their experience is different, their skillset is different, and I think it would just supplement, very nicely, the work that palliative care does of just supporting the patient and family.”—Palliative care clinician (participant 5)
Preintervention training for CHWs: “I think knowledge of the resources and maybe the red flags for when someone needs more urgent triaging, for especially depression, might be important. Kind of what are some of the things that might say I need—we need assistance ASAP for this patient? Otherwise, I think really just training and responding to emotion. Because so much of angst is related to just dealing with the emotion of all the ups and downs of illness, and treatment, and changes in functional status. That’s something that anyone in the health system can do, is recognize, and validate, and respond to emotion. That may actually help patients get to a more logical place where they can ask the questions that they need to ask. So, I think that could be something that would certainly be—I think that's helpful for any provider to be trained in, is recognizing and responding to emotion. Kind of knowing when questions are emotional, knowing when questions are more logical, and cognitive, and things like that.”—Palliative care clinician (participant 22)
Limiting caseload size: “You can spend hours with a patient when they’re at end of life and their family member who also need that support to be there for their loved one. It’s not a few minutes. It’s not a few minutes—therefore having an extensive caseload I would say will limit to an efficient [rather than thorough] job. I think there should be a limit placed on the patients that the community health advisors are advising or advocating for.”—Community health worker (participant 21)
Ensuring additional support for CHWs beyond initial training: “Provide them with equipment that they will need to be efficient … ‘We’re going to provide you with,’ we will say, ‘an incentive of $10 an hour, $20 an hour’ or whatever … but offering them a wage. And offering them the equipment, the technology that they would need. Possibly instead of using their own phone, which I have advocated for that—a community health advisor should not be using their own cell phones to communicate.”—Community health worker (participant 21) “I would say … every two to three months like refresher courses for [CHWs] just to continue to guide the CHWs, or to factor in something new, because you might have a new education piece that needs to be incorporated. So whatever you include [that is] new into the program, hold the training for the CHWs.”—Community health worker (participant 1)
Consideration of a disease-specific design: “We are trying to create what we call ‘disease-based pods,’ and within these pods are the nurse navigators, the disease-based social worker, intake specialist—all the people that can help a patient get through that cancer journey—and I think this person could fit very naturally into that … I think it would be really nice to see them as disease-based because I think each of the different diseases carries with it a different set of needs and pitfalls that could occur in those patients, and having someone kind of be more familiar with what would happen to that patient because it is going to be so different for each different disease. I think that would be really helpful.”—Cancer center leader (participant 23)

b. Intervention delivery
Importance of clearly defining roles and integrating into clinical workflow: “I think we all work very well together and there is that level of respect. I do know that on the flip side of it, there have been times that [navigators] have expressed and I have felt myself that they tend to be the so-called dumping ground … it got to be so much at one time that we were almost like minions who were running around delivering what needed to go where and that’s the truth. It just kind of got out of control. It snowballed. So, this one needs a meal bag. This one needs a blood pressure cuff and they felt used, devalued, and insignificant and we had to really work to get that off of us, if you will because we want to make sure that we have the respect that everybody else has in health care or anywhere else. So, you don’t want anybody to feel devalued or they’re less than or let them do the running, that kind of thing, because we are the eyes and the ears out there.”—Care navigator (participant 14) “There’s just a [need for] delineation of what the resource is that’s available, and making it clear of how to contact that [CHW], but also who or what are the appropriate things to escalate to that person. It’s all-around education of I think just awareness that that role exists, and then how best to utilize that role … I think it just gets back to making sure that the communication is there, that it exists, and how to use it appropriately.”—Palliative care clinician (participant 5)
Setting for CHW introduction and assessments: “Bring whoever that CHW is going to be on their care team, let them meet them in the hospital, and you introduce them to the CHW as well as to the role of the CHW. And the CHW can make a follow up visit right then and there. So you won’t be getting a random call: [makes surprised face] ‘Nah, you can’t come to my house!’”—Community health worker (participant 1) “Normally I would say part of the time they could probably work from home, but it’s always helpful to also introduce yourself face to face at some point with the client who you’re working with, and usually earlier than later.”—Care navigator (participant 20)
Balancing adaptability with standardization: “I think it’s fine for it to be adaptable, but I don’t think it should feel adaptable when you put it into place. I think that for the person in the job and the person that they’re taking care of … it should feel very concrete and discrete. And then you can, every three months, or every period of time, reassess and consider what roles and responsibilities you might want to add on or take away. But I think it should not feel vague, and it should not feel open-ended, in practice.”—Oncology clinician (participant 8)
Incorporating both expressed and observed patient needs: “After you meet the patient at their house, you’re doing observation of their house, not just what they’re telling that they need help with. Okay? ‘Miss Betty,’ she’s 75 years old. She’s having problems getting up and down the steps. What resource can I connect her with to get a lift in her house? And I’m telling you the Internet is part of a CHW’s resource. But if you’ve been in the CHW [world] … for some time, like you already know, well, ‘let me reach out to so-and-so and get her connected with this,’ and then you’ve got to follow up. Have they come out and [done] the assessment yet? Did they say that you have to pay anything towards it? Let me see who I can get you funding from. Like, good follow-up. That’s what the CHW role would be.”—Community health worker (participant 1)

CHW, community health worker.

Though specific areas of suggested emphasis varied, key informants expressed hope that there would be substantial preintervention didactic training for CHWs in palliative and hospice care, oncology, and recognizing psychosocial distress. The appropriate caseload for a CHW in our intervention was not well-defined by the informants in absolute terms, though they acknowledged that managing a caseload of patients with potentially high needs would require strong organizational skills and appropriate boundaries and that caseloads may need to be limited based on CHW experience and patient acuity. Ensuring adequate emotional support for CHWs, fair remuneration, and ongoing training in providing psychosocial support and making use of local community resources were also felt to be important. Some key informants recommended a disease-specific pilot program, in which CHWs would initially be integrated into a subdivision of the cancer center.

Intervention delivery recommendations

Both CHWs and other key informants expressed the paramount importance of clearly defining roles, responsibilities, and expectations for CHWs and the involved clinicians prior to beginning the intervention (Table 3b). Training the clinical team to absorb an extra provider into its workflow was also considered very important. Key informants envisioned CHWs assisting patients in their homes, over the telephone, as a supportive presence during appointments, and as visitors during hospitalizations. In general, informants were receptive to having CHWs use personal judgment on the frequency and nature of follow-up contacts. However, some informants expressed a preference for a degree of standardization. CHW assessments were seen as a way to help clinicians by enhancing knowledge about barriers to care, and informants felt that assessments should include both expressed and observed patient needs.

Discussion

Our study adds substantially to the very limited qualitative literature on the use of CHWs in palliative care²⁴ and oncology patient navigation²⁵ interventions within the United States. In semi-structured interviews that included clinicians, cancer center leaders, and CHWs, we explored a proposed CHW-driven palliative care intervention for African American patients with advanced cancers. Informants agreed that current palliative care experiences for this patient population could be improved and expressed strong enthusiasm for the intervention’s goals. Informants shared their perspectives on potential challenges, including barriers at the patient, clinician, facility, and health system level. We also explored recommendations to optimize the intervention’s design and implementation rooted in the diverse perspectives of our key informants. On the basis of our findings, the intervention and its implementation process were adapted in multiple ways, summarized in Table 4. Our findings also provide information that may aid others working to alleviate palliative care disparities or implement CHW-based interventions.

Table 4.

Aspects of the Intervention That Were Added, Adapted, or Confirmed on the Basis of Identified Barriers or Recommendations

Barrier(s) or recommendation(s)	Corresponding aspect(s) of the intervention
Lack of knowledge about palliative care among patients or clinicians Low health literacy among patients Concern about ability of CHWs to address complex needs and relay concerns to clinicians effectively Recommendation for CHW didactic training	Rigorous preintervention CHW training was designed by the study team in concert with subject matter experts in palliative care, oncology, patient navigation, health disparities, and motivational interviewing. Training incorporated a synchronous training component delivered over one week via videoconferencing, which included a combination of didactic and problem-based learning. Subsequently, there was an asynchronous training component, delivered over four weeks through a series of weekly modules. Modules covered clinical communication, working within care teams, patient activation, health disparities, and understanding the role religion can play for some patients. A final experiential training component included shadowing palliative care providers, identifying community resources, and participating in a site-specific orientation.26
Logistical barriers and financial burdens to receiving care	A standardized initial screening questionnaire was developed for CHWs to use as they began to engage with patients. Based on the Protocol for Responding to and Assessing Patients’ Assets, Risks, and Experiences (PRAPARE) toolkit,27 the questionnaire evaluates access to care, emotional support, vision and hearing, social program eligibility, transportation, housing, financial resources, food security, and additional needs. After identifying individual patient and caregiver barriers to health care access and associated social determinants of health, CHWs tailor their approach and support to each patient or patient-caregiver dyad.
Negative community and personal experiences within the health system	CHWs were recruited from within the communities served. The hiring process emphasized a history of local community engagement, ability to establish positive relationships, and comfort with discussing patients’ and caregivers’ medical, psychosocial, and spiritual concerns.
Limited access to palliative care providers	The study team developed partnerships with local palliative care staff and organizations. CHWs received dedicated training on palliative care from palliative care providers and shadowed providers to understand their daily operations and establish relationships with members of the team. Each site also established a Palliative Care Liaison from within their palliative care department to serve as an internal “champion” for the intervention and act as resource for CHWs.
Lack of ongoing funding after randomized controlled trial	Trial secondary endpoints include detailed resource utilization metrics, including number of hospital days, intensive care unit days, emergency department visits, hospital readmission rates, health care costs, and time spent with CHWs. Social return on investment (SROI) analysis was added and we engaged a health care economist collaborator to lead our SROI team.
Recommendation to train the clinical team to absorb CHWs into the workflow	CHW training included shadowing oncology and palliative care providers, allowing CHWs to understand preexisting clinic workflow. Oncology and palliative care teams received a presentation on the intervention prior to implementation. A short video explaining the role of the CHW is being produced to provide asynchronous information for clinicians.
Recommendation to balance standardized versus individualized follow-up frequency	CHWs began with an initial standardized evaluation and screening questionnaire to assess access to care and social determinants of health. Follow-up was then individualized after this point at the discretion of CHWs. Regular all-site CHW meetings occur on a biweekly basis, facilitating consistency and dissemination of best practices across sites. CHWs also receive ongoing mentorship from the study’s lead CHW, creating an opportunity for individualized advice and support from a highly experienced colleague.

One theme of our findings was the educational barrier for both patients and clinicians surrounding the definition and role of palliative care, especially as it compares to hospice care and to cancer-directed therapies. Patients, and even some clinicians, were perceived as having little knowledge of palliative care or as possessing a misperception that it represented “giving up” or abandonment of the patient and his/her goals. These qualitative findings align with others’ quantitative observations: survey-based data of community-dwelling New York State residents found that a majority had a misperception about palliative care, and <10% saw it as applicable to any phase of illness.²⁸ International studies have also confirmed resistance to palliative care among some oncologists;^29,30 encouragingly, however, oncology provider education has been associated with improved self-efficacy toward primary provision of palliative care or use of specialist services.^31–33 Taken together with the existing literature, our findings suggest that focused education about palliative care may support both clinicians and patients with advanced cancer in their pursuit of optimal cancer care. Within our intervention specifically, these findings supported an emphasis on the educational role of CHWs during our training program and helped us describe and frame our intervention appropriately to clinicians at each site.

The importance of high-quality CHW training was another major theme that emerged as both a potential barrier and a recommended area of focus in the intervention’s design and implementation. Informants collectively suggested that didactic training should cover diverse medical topics, including palliative and hospice care, cancer-directed therapies and oncology, and chronic diseases commonly comorbid with cancer. Informants also discussed the importance of job skills training, such as triaging psychosocial and medical emergencies, engaging community resources, providing motivational and emotional support, and ensuring CHWs’ personal safety and boundaries while working in the community. These findings directly informed our previously described preintervention CHW training curriculum.²⁶ Our results align with those of Patel et al., who recently found that health care professionals identified appropriate CHW training, retention, and organizational support as key to achieving acceptance and effectiveness of a CHW-led advance care planning and cancer symptom management intervention.²⁴

Finally, ensuring the intervention’s sustainability through ongoing funding and proof of cost-effectiveness was another important theme, particularly among cancer center leaders. Key informants described a lack of appropriate reimbursement for many palliative care services but hoped that a corresponding decrease in health care utilization might offset increased investment. In discussing the reimbursement landscape, some informants seemed concerned primarily that specialist palliative care usage would increase but was not cost-effective given its reimbursement level and the time and expertise required. Other informants focused on the lack of reimbursement for CHW services. Within the current reimbursement setting, reviews of the existing literature are inconclusive on the cost-effectiveness of hospital-based³⁴ and at-home³⁵ palliative care in general. Data on CHW interventions specifically are very limited. Encouragingly, two recent studies of a lay health worker symptom screening intervention among patients with cancer found an association with lower total costs of care.^36,37 In addition, Medicaid coverage for CHW services is now offered by more than half of U.S. states,³⁸ and Medicare coverage is anticipated to begin in 2024.³⁹ Alongside examining health care utilization data in our upcoming trial, we will also define our intervention’s social return on investment, a methodology that expands traditional cost–benefit analysis to incorporate the social and socioeconomic value generated by an intervention.⁴⁰

Our study must be interpreted in the context of its limitations. First, though we explored the perspectives of diverse types of professionals across three unique institutions, this analysis did not include patients or caregivers and thus cannot comment directly on their perspectives. We anticipate that a separate qualitative investigation of patient and caregiver perspectives will be reported in future work. Participants were selected by purposive sampling, which may have favored the inclusion of those likely to be enthusiastic about our intervention. Our interview data were not directly corroborated for factuality. However, we focused our data interpretation on major themes expressed by multiple informants and completed site visits to obtain general knowledge of each site. Next, our interviews were conducted by three different interviewers. However, differences among interviewers were mitigated by adhering to an interview guide and standard operating protocols and were also critically examined as an inherent part of this qualitative study through the principle of reflexivity.⁴¹ Finally, by design, our findings are highly relevant to our intervention. However, they may not necessarily be generalizable to other settings or populations.

Conclusions

We explored implementation barriers and recommendations related to a proposed CHW intervention to improve palliative care outcomes among African American patients with advanced cancer. Clinicians, cancer center leaders, and CHWs not only identified potential barriers across the patient, clinician, facility, and system levels but also described ways to mitigate these barriers through intervention design and implementation. Our findings provide key insights that were used to adapt our specific intervention and may also inform other investigations related to CHWs and palliative care.

Funding Information

This work was supported by a grant from the National Institutes of Health (National Cancer Institute award #R01CA252101). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH and supporting institutions.

Author Disclosure Statement

The authors have no conflicts of interest to disclose.