Understanding Underuse of Advance Care Planning Among a Cohort of African American Patients With Advanced Cancer: Formative Research That Examines Gaps in Intent to Discuss Options for Care

Abstract

Background:

Advance care planning (ACP), palliative care (PC), and hospice are often underutilized by African Americans (AAs). This study assessed the impact of stage of intent to discuss ACP options as key potential barriers.

Methods:

We examined intent to discuss completion of ACP, PC, and hospice among 22 AA patients with cancer admitted to a local safety net hospital. Participants were asked about intent to discuss an advanced directive or living will (AD/LW), medical power of attorney (MPOA), PC, and hospice with their doctors. Intent to discuss these ACP components was based on the transtheoretical model. Electronic health records were reviewed at various intervals to assess completion of ACP behaviors and survival.

Results:

Participants had colorectal (33%), breast (44%), and lung (23%) cancer, and 82% had stage III/IV disease. Low percentages of patients were in the precontemplation stage for AD/LW completion (4.6%), MPOA completion (13.6%), and PC discussions (27.2%), but 77.2% were in the precontemplation stage for hospice discussions. At 1 year, only 5% completed an AD/LW, 36.4% appointed an MPOA, 42.9% were referred to PC, and 12.5% were referred to hospice. More than half (54.6%) were deceased by the study’s conclusion. Most (81%) of these died within 6 months of their baseline study assessment.

Conclusions:

Despite being hospitalized with advanced cancer and having poor prognosis, intent to discuss ACP options, PC, and hospice in this population was variable, and completion of these activities was low. This formative research is needed to develop education and counseling interventions for this high-risk, vulnerable population.

Introduction

Advance care planning (ACP), palliative care, and hospice are often underutilized in the underrepresented and underserved. Research suggests many reasons for this dynamic, including lack of knowledge, medical mistrust, spiritual conflict, and general preferences for more aggressive treatment.^1–4 African Americans and members of other underrepresented groups are sometimes not informed about options for care and often report they have not talked to anyone about their care goals, values, and preferences.^2,5 Many studies have documented racial and ethnic disparities in ACP and end-of-life care. While multiple reasons for this underuse have been identified, less is known about African American patients’ behavioral intention or willingness to consider ACP or palliative care, both of which may affect their end-of-life care.

Quality of life (QOL) and symptom management have been investigated in African American patients with cancer with varied results. Among some African American cancer survivors, higher social support levels are associated with increased mental health QOL.⁶ Other studies have shown that some low-income, African American patients with cancer report higher emotional well-being and fewer symptoms than non-Hispanic whites.⁷ Additionally, patient–provider communication may influence physical and mental health QOL in African Americans with cancer diagnoses.⁸ Differences in pain and symptom management have also been identified. A study of bereaved family members of patients who died while receiving hospice services found African Americans more likely to express concerns about unmet needs for pain, dyspnea, and emotional support.⁹ Other research has found that African Americans and members of other underrepresented groups express concern about pain management more often than non-Hispanic white patients.¹⁰ Reasons for these differences are likely multifactorial and deserve further investigation.

Palliative care has been shown to improve QOL and symptom burden among patients with cancer, and research suggests racial differences in participation in ACP and enrollment in hospice may improve once members of underrepresented groups receive palliative care consultation.^11–14 Lack of knowledge of palliative care as an option for treatment can act as a significant barrier to end-of-life care, and education about this option for care should be enhanced.¹ The goal of this study was to describe QOL, symptom burden, and stages of intent to discuss ACP, palliative care, and hospice among a cohort of African American patients hospitalized with advanced cancer. There is an abundance of research that identifies differences in completion of ACP and use of hospice among African Americans; consequently, the overarching goal of this study was to examine the factors noted above as a means of conducting the formative research needed to develop interventions to reduce disparities in ACP and end-of-life care in this population.

Methods

Patients were identified via an electronic health record (EHR)-enabled algorithm. The algorithm identified patients by International Classification of Diseases, Tenth Revision codes for their primary cancer diagnosis, race/ethnicity, and other factors.¹⁵ The algorithm then alerted the research team when a patient was admitted to the hospital. Prospective participants’ charts were then reviewed in detail by a board-certified hospice and palliative medicine physician to confirm eligibility. Eligible patients were those who (1) identified as non-Hispanic black; (2) had been diagnosed with advanced (stage III or IV) breast, lung, or colorectal cancer or had stage II disease with significant comorbidities, making them appropriate for palliative care; (3) were able to read, write, and speak English; (4) were competent to give informed consent; (5) were not receiving palliative care or hospice care at the time of the study; and (6) were hospitalized at the time of recruitment. Eligible patients were then contacted during their inpatient hospitalization to gauge their willingness to participate. Identified caregivers were also invited to participate in the study. Parkland health and hospital system is responsible for caring for the county’s uninsured and underinsured. It is a public, integrated delivery system with a comprehensive EHR used across all inpatient and outpatient settings. 1. Low-income residents without insurance are eligible for the Parkland financial assistance program, which covers hospital and ambulatory-based care including cancer treatment, palliative care, and pain medications. We predominantly focused on patients who had advanced disease (stage III/IV) but also recruited some with stage II disease and other significant comorbidities that made them appropriate for possible palliative care consultation. Participants completed an in-person baseline assessment that captured data on sociodemographics, religious affiliation, and symptoms and QOL via the McGill Quality of Life Questionnaire.¹⁶ We assessed patients’ intent to complete an advance directive/living will (AD/LW), complete a medical power of attorney (MPOA), discuss palliative care, and discuss hospice with their doctor or health-care team. Stage in intent to complete ACP components, to discuss palliative care, and to discuss hospice was based on the Transtheoretical Stages of Change model.¹⁷ Our questionnaire consisted of items that related to the specific stage of behavior change for ACP (completion of an AD/LW or MPOA), discussion with providers about palliative care, and discussion with providers about hospice. For instance, stage of intent to complete an AD/LW was determined by asking: “Do you intend to complete an AD or LW?” Response choices were: (1) “No, and I do not intend to in the next 6 months (precontemplation),” (2) “Not right now, but I intend to in the next 6 months (contemplation),” (3) “Not right now, but I intend to in the next 30 days (preparation),” (4) “Yes, and I have within the past 6 months (action),” or (5) “Yes, and I did more than 6 months ago (maintenance).” Charts were abstracted at 1, 3, 6, and 12 months after enrollment to ascertain ACP outcomes, including documentation of a signed AD/LW, MPOA, out-of-hospital do-not-resuscitate (DNR) form, use of palliative care, or hospice (enrollment in or referral to hospice), vital status, and date/location of death if applicable and available. Descriptive statistics were used to characterize the patients and summarize their responses and outcomes. Associations between stage of intent to discuss ACP components and actually completing those behaviors were assessed using the χ² test. All analyses were conducted using Stata/SE 15.1, College Station, Texas, and significance levels of P < .05 were used for all statistical comparisons. The study was approved by the Institutional Review Board of the University of Texas Southwestern Medical Center.

Results

Of the 47 patients identified by the electronic algorithm, 8 were deemed ineligible (1 for delirium, 7 because they were being discharged). Sixteen declined participation, but 23 agreed to participate. One patient was subsequently deemed ineligible because of cognitive impairment identified during recruitment. Median age was 59.5 years and 59.1% of participants were female. One-third of participants (33.3%) had colon/colorectal cancer, 43.8% had breast cancer, and 22.9% had lung cancer. Only 8 (36.3%) participants could identify someone they considered their primary caregiver. Of those caregivers, only 6 completed their initial assessment; consequently, caregiver data were not analyzed in this study. All caregivers were considered informal, as they provided care at no cost to study participants. The majority of patients (81.8%) had either stage III or stage IV cancer, while 18.2% had stage II disease with significant comorbidities (human immunodeficiency virus, congestive heart failure, etc). Most patients (63.6%) received county financial assistance for their health care, which does not cover hospice, and 31.8% had Medicaid or Medicare Part A.

Only 36.4% of participants were seen in palliative care clinic during the study’s duration. At the study’s conclusion, only 1 (4.6%) participant had a signed AD in the chart, 36.4% had a signed MPOA, and 2 (9.1%) participants had a documented out-of-hospital DNR order. At the study’s end, 54.6% of patients had died, and 66.7% of those patients died during an inpatient hospitalization. Median time to death from baseline assessment was 79 days (see Table 1).

Table 1.

Baseline Characteristics and ACP Outcomes of the African American Patients Hospitalized With Advanced Cancer.

Patient Characteristics at Baseline	All (N = 22)
Age in years, median (range)	59.5 (36–78)
Gender
Female	13 (59.1)
Male	9 (40.9)
Primary cancer diagnosis
Colon/colorectal	16 (33.3)
Breast	21 (43.8)
Lung	11 (22.9)
Cancer stage
1	0 (0)
2	4 (18.2)
3	5 (22.7)
4	13 (59.1)
Religion
Protestant	5 (22.7)
Other	5 (22.7)
None	1 (4.6)
Baptist	11 (50.0)
Marital status
Married	4 (18.2)
Divorced	8 (36.4)
Separated	3 (13.6)
Never married	4 (18.2)
Widowed	3 (13.6)
Medicare Part A
Yes	7 (31.8)
No	14 (63.6)
Don’t know	99 (4.6)
Medicaid
Yes	7 (31.8)
No	14 (63.6)
Don’t know	1 (4.6)
County financial assistance
Yes	14 (63.6)
No	7 (31.8)
Don’t know	1 (4.6)
Total household income
US$0-US$10 999	8 (36.4)
US$11 000-US$20 999	3 (13.6)
US$21 000-US$30 999	3 (13.6)
US$31 000-US$50 999	2 (9.1)
US$51 999-US$99 999	1 (4.5)
US$100 000 or more	1 (4.6)
Don’t know	4 (18.2)
ACP outcomes at 12 months
Advance directive/living will in patient’s chart	1 (4.6)
Medical power of attorney in patient’s chart	8 (36.4)
Out-of-Hospital Do-Not-Resuscitate order in patient’s chart	2 (9.1)
Patient seen in palliative care clinic	8 (36.4)
Patient referred to hospice	6 (27.2)
Died	12 (57.1)
Location of death if known
Home	2 (13.3)
Hospital	10 (66.7)
Nursing home	0 (0)
Hospice	0 (0)
Median time to death from baseline assessment, days (range)	79 (20–361)

Abbreviation: ACP, advance care planning.

Patients were asked about their overall QOL and reported symptoms over the previous 2 days (48 hours). Quality of life was assessed on a 0 to 10 scale (0 = very bad, 10 = excellent). Of all participants, 18.2% reported bad overall QOL, while 27.2% reported that their QOL over the previous 2 days had been excellent. The majority reported symptoms over the previous 2 days as well, including pain (85.7%), tiredness or fatigue (80.9%), weakness (76.2%), sleeping problems (71.4%), and shortness of breath (80.9%; see Table 2).

Table 2.

Overall Quality of Life and Symptom Burden.

Question	Response (N = 21)
Overall quality of life
Considering all parts of life, over the past 2 days, the quality of your life has been (0 = very bad, 10 = excellent)
0–2	5 (22.8)
3–5	4 (18.2)
6–7	2 (9.2)
8–10	11 (49.9)
Patient-reported symptoms over the past 2 days
Pain	18 (85.7)
Fatigue	17 (80.9)
Weakness	16 (76.2)
Nausea	10 (47.6)
Vomiting	3 (14.2)
Poor appetite	11 (52.4)
Difficulty sleeping	15 (71.4)
Shortness of breath	17 (80.9)
Constipation	8 (38.1)

Participants were in various stages of intent to discuss completion of an AD/LW, appointment of an MPOA, discussion with providers about palliative care, and discussions with providers about hospice. Only 27.3% intended to discuss completion of an AD/LW with their doctor or health-care team in the next 30 days after baseline assessment (preparation), while 31.8% did not know whether they intended to discuss this. Of participants, 31.8% intended to discuss appointing an MPOA within 30 days of their baseline assessment (preparation), but 27.3% had no intent to discuss palliative care with their doctor or health-care team in the 6 months after baseline assessment (precontemplation). There was less variability in intent to discuss hospice; 77.3% had no intent to discuss hospice with their doctor or health-care team within the next 6 months (precontemplation; see Figure 1). Although there were no statistically significant associations between stage of intent to discuss ACP components and actually completing those behaviors, the association between stage of intent to discuss hospice and actual referral to hospice did approach statistical significance (P = .087).

Figure 1.

Stages of intent to discuss advance care planning issues at baseline. AD indicates advance directive; LW, living will; MPOA, Medical Power of Attorney.

Discussion

This descriptive study examined QOL, symptom burden, and stages of intent to complete or discuss ACP, palliative care, and hospice among a small sample of hospitalized African American patients with cancer. We found that, while many rated their overall QOL positively (49.9% rated their QOL as 8, 9, or 10 on a 0–10 scale), many also reported symptom burden, including pain, fatigue, weakness, difficulty sleeping, and shortness of breath. Despite these reported symptoms, only about one-third (36.4%) were seen in palliative care clinic during the study period. We also found that, despite what we now know were poor prognoses for many participants, three quarters (77.3%) of patients had no intention of getting information about hospice from their doctor or health-care team within the next 6 months. Only 1 participant completed an AD and few appointed an MPOA. Furthermore, two-thirds (66.7%) of patients died during an acute care hospitalization where such decisions would be most useful.

Patients in this study were interviewed during an inpatient hospitalization, and many reported pain and other symptoms. Despite these reports, many rated their overall QOL positively. These findings are contrary to previous research that suggests a negative relationship between symptoms and QOL among some African American patients with cancer.¹² Additional reviews of the literature indicate that overall the QOL for African American cancer survivors is poorer than for Caucasians, though additional research is needed that uses consistent, culturally appropriate measures across studies.¹³ One could argue that QOL is in the eye of the beholder. Some patients with advanced cancer may view their overall QOL positively despite the circumstances of being hospitalized and having some symptom burden because they are, in fact, living. Others may view their overall QOL positively during hospitalization because there is an expectation that their acute medical concerns will be or are being addressed. Unfortunately, we were only able to assess overall QOL, as all data needed to calculate the McGill Quality of Life total scores were not available. Calculation of subscales and the total score may have resulted in different findings. As noted previously, future research with consistent measures should be considered to evaluate QOL in African American patients with advanced cancer.

Despite poor prognoses, many patients in our sample had not completed ACP documents. Furthermore, few were seen in palliative care clinic, and the vast majority (77%) were in the precontemplative stage of getting information about hospice from their doctor or health-care team. A number of studies have evaluated racial/ethnic disparities in advance care planning and hospice enrollment. Some previously cited barriers include lack of knowledge about advance care planning, palliative care and hospice, misconceptions about what ACP entails, and perceived conflicts with patients’ spirituality and the hospice and palliative medicine philosophy of care.^1–4 Other research suggests that racial and ethnic minorities may be disadvantaged if they have serious illness because most have not talked to anyone about their goals, values, or preferences for care.⁵ Another barrier to accessing hospice may come from lack of insurance. In our sample, 63.6% had the county assistance plan, which does not cover hospice care. Only 31.8% had Medicare Part A or Medicaid, which do cover these services. However, all patients had access to inpatient and outpatient palliative care services, and our previous research has found that access to palliative care among those who are underrepresented or underserved may lessen disparities in hospice enrollment—even in underserved populations.¹³

In this study, we used the transtheoretical stages of change model to assess stage of intent to discuss ACP, palliative care, and hospice. Similar models have been used to evaluate behavior change in a number of health behaviors, including ACP; however, these models have been validated in majority populations.^18–20 In our study, there was significant variability in stage of intent to complete an AD/LW or appoint an MPOA. There was also variability in stage of intent to discuss palliative care among participants. There was, however, less variability in stage of intent to discuss hospice. As noted previously, 77% of participants were still in the precontemplation stage regarding discussion of hospice compared to 4.6% for AD/LW, 13.6% for MPOA, and 27.3% for palliative care, respectively. The decreased variability in our hospice findings may indicate that participants had some prior knowledge of hospice and some hesitance in speaking with their providers about it. This could be based on misperceptions about the hospice philosophy of care or limited knowledge of prognosis—factors that have been cited as barriers to hospice enrollment for African Americans throughout the literature. Furthermore, to our knowledge, there are no validated tools that measure knowledge of ACP, palliative care, and hospice. Future research should be dedicated to the creation and validation of measures that assess patient and caregiver knowledge of ACP, palliative care, and hospice among diverse populations.

Our study had certain limitations that must be taken into account. Although the small sample of African American patients with advanced cancer received their care in one safety net hospital limits generalizability, our findings are consistent with other research that cites underutilization of ACP, palliative care, and hospice by members of the African American community. Another limitation is the potential underestimation of AD/LW and MPOA documentation in the medical record; however, charts were reviewed multiple times—at 1, 3, 6, and 12 months—in hopes of accurately capturing these data. Additionally, given losses to follow-up, our estimates of survival might be overly optimistic because deaths not documented in the safety-net health system would have been missed even with our repeated reviews of the medical record. Finally, because of the limited number of caregiver participants and small sample size, we were unable to examine any associations between caregiver involvement or participants’ religious preferences and stage of intent to complete ACP or to discuss palliative care and hospice. Despite these limitations, these findings add to a growing body of literature examining ACP, palliative care, and hospice utilization among members of underrepresented minority groups, who are sometimes of low socioeconomic backgrounds. Furthermore, this study yielded the additional formative research that we need to design and implement culturally sensitive interventions that address barriers to ACP, use of palliative care, and enrollment in hospice for patients who have historically underutilized these services. We plan to use the data obtained from this and other studies we have conducted to develop these interventions, with a longer term goal of conducting multisite interventions to further establish generalizability. This work helps us to recognize even more that for all ACP behaviors, nearly all patients will need more support across all stages of intent, particularly getting beyond precontemplation and contemplation for most end-of-life issues and options.