Barriers to recruitment into a multi-site randomized controlled trial comparing the effectiveness of specialty outpatient versus nurse-led telephonic palliative care of older adults with advanced illness

Background Emergency department (ED) visits among older adults are common near the end of life. Palliative care has been shown to reduce ED visits and to increase quality of life among patients, but recruitment into these programs is often challenging. Methods A multi-site randomized controlled trial investigating two modes of community-based palliative care delivery for patients in the ED who are discharged home. Research coordinators monitored factors that prevented eligibility and enrollment by screening electronic health records to identify and recruit eligible patients and their caregivers. Reasons for ineligibility and non-participation were documented. Results Among the 7,091 patients who met the age and disease qualier, 4,292 (60.5%) were deemed ineligible due to hospital admission. Additional reasons for ineligibility included previous palliative care (919 [13%]) or residence in a nursing facility (462 [6.5%]). Of the 695 patients who were eligible, 298 were enrolled, yielding an enrollment rate of 43% (range of 9% to 80% across all sites). Of the 397 eligible patients who did not participate, 180 (45.3%) refused due to barriers related to illness severity. Patients commonly refused due to misconceptions/stigma related to palliative care (118 [29.7%]). One-hundred fteen patients (29%) refused due to the mode of palliative care delivery. Conclusions Hospital admission from the ED proved to be the most signicant barrier to eligibility. Eligible patients often refused participation due to the nature of their illness, misconceptions about palliative care, and concerns about being able to attend clinic visits if assigned to the outpatient arm. This study is a ve-year, randomized controlled trial designed to compare the effectiveness of two modes of community-based palliative care delivery for patients with advanced illness who are recruited from the ED or observation unit: nurse-led telephonic supportive care and facilitated, outpatient specialty palliative care. One year into the study, there have been signicant barriers to both eligibility and recruitment at nine different ED sites across the United States. The focus of this paper is to assess the barriers to enrolling patients with advanced illness into a palliative care randomized controlled trial in the ED setting.


Background
Multiple studies have shown that palliative care, which may be initiated in the emergency department (ED) and delivered in conjunction with life-prolonging treatment, improves quality of life in patients with advanced illness. [1][2][3] Additionally, patients receiving palliative care utilize fewer resources, have reduced ED visits, and fewer admissions, leading to reduced cost of medical care and higher satisfaction for patients and families. 2,4 In spite of this, there are signi cant barriers to enrolling patients in palliative care programs. These barriers include patient, family, and provider refusal, di culties in communication regarding end-of-life care, and lack of training in palliative care amongst providers. [5][6][7][8] Recruiting patients into palliative care research in the ED creates additional challenges due to the fast-paced environment, lack of an ongoing relationship with the patient, and high symptom burden that brought patients there to begin with. 9,10 Engaging older adults with advanced illness in research is challenging for multiple reasons. Patients are often too sick or burdened by their illness to participate, and limited life-expectancy among this population complicates data collection. 11,12 In spite of these issues, the research that does exist highlights the overwhelmingly positive bene ts of palliative care in patients with serious illnesses. 1,2,13 This knowledge, combined with the lack of understanding about how best to deliver palliative care to patients, warrants the need for additional research.
This study is a ve-year, randomized controlled trial designed to compare the effectiveness of two modes of community-based palliative care delivery for patients with advanced illness who are recruited from the ED or observation unit: nurse-led telephonic supportive care and facilitated, outpatient specialty palliative care. One year into the study, there have been signi cant barriers to both eligibility and recruitment at nine different ED sites across the United States. The focus of this paper is to assess the barriers to enrolling patients with advanced illness into a palliative care randomized controlled trial in the ED setting.

Methods
Research coordinators monitored factors that prevented patient enrollment into Emergency Medicine Palliative Care Access (EMPallA), a multi-site randomized controlled trial (RCT) comparing two methods of delivering community-based palliative care. 14 EMPallA currently recruits patients and their caregivers from nine ED sites across the country. Eligibility criteria for patients include those who are 50 years of age or older who present to the ED with a qualifying illness. Qualifying illnesses include advanced cancer or end-stage organ failure (de ned as New York Heart Association Class III or IV heart failure, End-Stage Renal Disease with glomerular ltration rate < 15 ml/min/m 2 , or Global Initiative for Chronic Obstructive Lung Disease Stage III, IV, or oxygen-dependent Chronic Obstructive Pulmonary Disease) among patients who are scheduled for ED discharge or observation status. Patients must also have health insurance, speak English or Spanish, and reside within the geographic area. Patients are excluded from the study if they have previously received hospice or palliative care, are admitted to inpatient services, have dementia, or live in a skilled nursing facility (SNF) or similar assisted living facility (ALF). Caregivers who are at least 18 years old, have access to a working telephone, and speak English or Spanish are also eligible to enroll in the study. The study was submitted to and approved by the NYU School of Medicine Institutional Review Board Ethics Committee.
Research coordinators screened electronic health records from 7am-12am, six days a week to identify potentially eligible patients. If a patient met study criteria, research coordinators approached the patient and obtained informed consent at bedside. If a caregiver was present at the time, they were also invited to enroll in the study. Every patient or caregiver who enrolled and completed the baseline survey received at $40 gift card as study compensation. All study procedures have been approved by the Institutional Review Board at each institution.
During the daily electronic health record screens, research coordinators documented reasons for noneligibility and refusals. For patients who met one or more of the disease quali ers but were determined ineligible, the reasons were documented. Patients could be ineligible for more than one reason (e.g. have dementia and live in a SNF). For patients who were eligible but chose not to participate in the study, they were asked to provide a reason for their refusal. Research coordinators then recorded their response using a pre-determined list of common reasons for refusal. Patients were permitted to provide more than one reason. Research coordinators recorded all data in Research Electronic Data Capture (REDCap), a secure, web-based application that served as the research database. In order to track eligibility and enrollment, refusal data was organized into a Consolidated Standards for Reporting of Trials (CONSORT) diagram (see Appendix A). [15][16][17]

Results
To date, 7,091 patients with a qualifying illness have been identi ed, of which 6,196 were deemed ineligible for the study (87.3%). Table 1 provides a comprehensive list of reasons for patients who met the disease quali er were deemed ineligible for the study. Across all sites, 4,292 (60.5%) patients were ineligible due to hospital admission from the ED (range: 38-82.1% across all sites). Other common reasons included previous exposure to palliative care (13%) and residence in a SNF or ALF (6.5%). A smaller proportion of patients spoke a primary language other than English or Spanish (4.9%) or were diagnosed with dementia (4.1%). Other barriers included living outside the treatment area, lack of insurance, and current hospice care, but these were less common (< 2%). Of the 695 eligible patients, 298 (43%) were successfully enrolled. Enrollment rates varied by siteranging from 9-80%. Sample sizes were widely variable (see Table 2). Caregiver enrollment was also tracked by site. Of the 298 patients enrolled, 85 also had a caregiver enrolled, resulting in a caregiver enrollment rate of 29% (range: 0-75% across all sites). Reasons for refusal were categorized as follows: (1) barriers related to illness severity, (2) misconceptions/stigma related to palliative care, (3) mode of palliative care delivery, (4) general research barriers, and (5) family/caregiver barriers. Results are based on the number of patients that cited each reason, and patients were allowed to give more than one reason. reported that they did not want to make a decision without a family member or a caregiver present.

Main Findings
Multiple barriers prevented both patient eligibility and recruitment into community-based palliative care within the context of a multi-site, ve-year RCT. Hospital admission from the ED was the main reason for ineligibility. This exclusion criterion is essential in order for EMPallA to target a speci c group of patients who may not otherwise have access to palliative care programs. Notably, palliative care services are available to admitted patients, but few resources for patients who are discharged home from the ED. [18][19][20] Although admission from the ED prevented many patients from being eligible for the study, this criterion is essential in order to address this gap in palliative care services. As shown in previous studies of populations with advanced illness, patients commonly declined participation for reasons that were related to living with an advanced illness, such as feeling overburdened with their current care. The FamCope intervention, which aimed to test the feasibility of nurse-led, family-coping-oriented palliative home care in patients with advanced cancer, also reported a high refusal rate related to the characteristics of their target population. 5 Many of these patients refused because their illness burden was too great, often citing a "lack of energy" or being "too sick." It is well documented that research with palliative care populations is inherently di cult due to the aforementioned reasons, so these ndings did not come as a surprise. 11,22,23 Among the eligible patients in our study population, some felt that they were satis ed with their current care. This could be explained by the chronic nature of their illness, which may require that these patients see several specialists, attend frequent appointments, and receive in-home care.
Recruitment barriers related to misconceptions or stigma regarding palliative care were also common. Patients often associated palliative care with hospice or death. Although it is well documented in the literature that palliative care improves quality of life and symptom burden for patients and their caregivers, patients often had preconceived notions about palliative care that could not be alleviated even after research coordinators described palliative care and its potential bene ts.

Future Implications
Mode of palliative care delivery was an in uential factor for patients' decisions to participate in the study.
Of the 115 patients who did not participate for this reason, 99 (86.1%) of them indicated that they did not want outpatient palliative due to inability or unwillingness to make it to the clinic visits. In contrast, only 16 (13.9%) said that they did not want to receive telephonic palliative care. These ndings suggest that patients may be more open to receiving palliative care if they did not have to attend outpatient clinic visits, which will have important implications for scalability.
In contrast to previous studies which have shown that caregivers often act as gatekeepers and refuse on behalf of the patient, barriers due to family or caregiver refusal were among the least common for our study. 23,24 This may be explained by the fact that many of our patients were alone at the time they were approached. Research coordinators typically approached patients during the weekdays, so caregivers who have jobs may be less likely to be present. Geographic variation likely also plays a role in these ndings. Many of our sites are located in urban areas with public transportation or public assistance programs, making it easier for patients to access healthcare without a caregiver.
Furthermore, studies have documented that physician gatekeeping is a common barrier to patient recruitment, particularly with palliative care populations. 10,22 On the contrary, this study has not yet encountered physician refusal. This may be a direct result of the requirement that physicians opt out at the beginning of the study, as per study protocol. It may also re ect changing attitudes toward palliative care. A study analyzing the perceptions of palliative care among healthcare providers before and after implementation of a palliative medicine division found increased attendance in educational activities and increased con dence in palliative care. 25 Researchers also noted that providers who favored comanagement with palliative care held core values that aligned with current concepts, such as advanced care planning. 25 As the number of palliative care programs in hospitals across the country increases, we expect to see a similar increase in understanding and acceptance of palliative care amongst providers. 26

Strengths/limitations
The strength of this paper is the generalizability of the results across different geographic ED contexts and patient populations with advanced illness globally. Given that this is a descriptive paper that is exploratory in nature, one of the limitations is that the results cannot be used to draw de nitive conclusions. We report on data that could be hypothesis-generating for researchers and clinicians in the eld. Additionally, the small sample sizes at some of our ED sites limit our ability to comment on these results. For example, one of the study sites has a fairly high enrollment rate at 68% (n = 13 of 19). However, with less than 20 eligible patients to date, it is likely that this number will change as the study progresses. Table 3 Reasons that eligible patients did not participate (n = 397