‘The Woman Gives’: Exploring Gender and Relationship Factors in HIV Advance Care Planning among African American Caregivers

Abstract

Aims and objective:

Advance care planning (ACP) is the communication process of documenting future health care preferences in case patients are unable to make health care decisions for themselves. Research suggests ACP discussions among persons living with HIV (PLHIV) are infrequent overall and may differ by gender.

Background:

Previous literature has displayed that African Americans are less likely to use advanced care planning, palliative care, and hospice, but does not conclusively account for ACP among PLHIV. African American PLHIV rely on informal care that may be gendered and an important pathway to increase ACP.

Design:

Participants completed self-report surveys and partook in in-depth interviews, resulting in a mixed method study design.

Method:

Participants completed self-report surveys (N=311) and were interviewed (n=11). Poisson regression (quantitative) and grounded theory analyses (qualitative) were implemented, using COREQ checklist principles.

Results:

Less than half of the quantitative sample had discussed ACP (41.2%; N=267). More ACP knowledge predicted 76% lower likelihood of ACP discussions among women. Men who spent more time caregiving in a given week were nearly 3 times more likely to discuss ACP. Women were more likely to be caregivers and also expected to serve in that role, qualitatively described as “being a woman.”

Conclusions:

The present study is one of few studies exploring ACP planning among caregivers in African American populations hardest hit by HIV. Results suggest that ACP skill-building and education are critical for African Americans living with HIV- ACP discussions with their caregivers. Indexes were low overall even when healthcare had recently been accessed. Support reciprocity and gender-specific communication skill-building may facilitate ACP in African American HIV caregiving relationships. Results underscore need for ACP education which includes healthcare providers, given African Americans’ strong cultural preference for life-sustaining treatments at end-of-life.

INTRODUCTION

Advance care planning (ACP) refers to ongoing conversations with family and care providers about future healthcare if individuals are unable to speak or make decisions autonomously (De Caprariis, Carballo-Dieguez, Thompson, & Lyon, 2013). ACP facilitates documentation of care preferences such as advance directives (AD), living wills, and designation of healthcare proxies in healthcare decision-making. ACP is associated with improved quality of life at end-of-life among individuals with chronic life-limiting illnesses, such as HIV (Koss & Baker, 2017). The 1991 Patient Self Determination Act (PSDA) mandated provision of written information about advance directives to admitted or enrolled patients at any Medicare- or Medicaid-funded healthcare centers due low ACP rates among adults in the US (Koss & Baker, 2017; Kamal et al., 2017). Private sector organizations like the American Bar Association have since led educational campaigns to increase advance directives completion, although increases have not been seen among African Americans who remain half as likely as Whites to have advance directives (De Caprariis et al., 2013; Koss & Baker, 2017; Kamal et al., 2017; Maragh-Bass, Zhao, Isenberg, Mitchell, & Knowlton, 2017). Increasing participation in ACP documentation among African Americans, particularly those living with HIV, can improve quality of life, long-term engagement in healthcare, and ultimately better health outcomes for African Americans living with HIV (De Caprariis et al., 2013; Koss & Baker, 2017; Kamal et al., 2017; Maragh-Bass, Zhao, Isenberg, Mitchell, & Knowlton, 2017).

ACP among African Americans

Research suggests that ACP rates may be low in the US partly due to healthcare advancements, which have reduced mortality rates from chronic illnesses such as HIV (Hansen et al., 2019). However, African Americans have benefitted less ACP efforts, much like they have benefitted less from HIV care advancements than other racial groups, even when controlling for factors such as socioeconomic status and/or interest in ACP (Maragh-Bass et al., 2017; Mitchell et al., 2018). African Americans comprise 12 percent of the US population but 45 percent of persons living with HIV (PLHIV; Mitchell et al., 2018; Centers for Disease Control & Prevention [CDC], 2016). Aging PLHIV may have particularly complex healthcare needs compared to their non-PLHIV counterparts (Kamal et al., 2017; Maragh-Bass et al., 2017; Mitchell et al., 2018; CDC, 2015). Therefore, research which identifies strategies to increase ACP for older African Americans with HIV is urgently needed.

Factors associated with ACP among PLHIV

Socio-demographics factors associated with increased likelihood of ACP include older age (CDC, 2015; Hong, Yi, Johnson, & Adamek, 2018; Clark, Person, Gosline, Gawande, & Block; 2018) which may be due in part to progressed illness severity (Clark et al., 2018). Access to and engagement in healthcare may be also associated with ACP; however, providers may not perceive need for ACP discussion and patients are not informed enough to inquire themselves (Maragh-Bass et al., 2017; Mosack & Wandrey, 2015). Providers avoid discussion of ACP with patients because of limited time during clinical encounters, but also because they are undertrained in counseling PLHIV on ACP (Mosack & Wandrey, 2015).

In urban settings, communication is further strained by patient-provider racial discordance and providers being ill-equipped to navigate the sensitivities of ACP, culture, and medical mistrust that is deeply rooted in African American communities. Cultural norms in the African American community reinforce mistrust of the healthcare system, and this barrier to ACP may be exacerbated by pervasive HIV stigma in the African American community (Schickedanz et al., 2009; Geter, Sutton, & Hubbard McCree, 2018). The role of social support as a facilitator to ACP is documented, although more research is needed to understand its role among older African Americans living with HIV with complex care needs such as history of substance use (Schickedanz et al., 2009; Geter et al., 2018; Taylor et al., 2018; Warren-Jeanpiere et al., 2017; Schnall, Hirshfield, Liu, Siegel, & Gradilla, 2018).

Informal Care and ACP in African American PLHIV

Informal care refers to unpaid care provided to individuals living with chronic illnesses, and is generally performed by the individual’s social network (e.g., spouses, children, friends; Maragh-Bass et al., 2017; Geter et al., 2018; Taylor et al., 2018; Warren-Jeanpiere et al., 2017; Schnall, Hirshfield, Liu, Siegel, & Gradilla, 2018; Nguyen et al., 2019). HIV informal caregiving includes multiple dimensions: emotional (empathy in coping with HIV); instrumental (tangible aid such as transportation to medical appointments); and informational (advice from caregivers related to HIV management; Mitchell et al., 2018; Schickedanz et al., 2009; Geter et al., 2018; Taylor et al., 2018; Warren-Jeanpiere et al., 2017; Schnall, Hirshfield, Liu, Siegel, & Gradilla, 2018). Due to socioeconomic disadvantage and medical mistrust, African American PLHIV may be more likely than other racial groups to depend on and/or provide informal care to PLHIV in their social networks which makes them a critical partner to increasing ACP in this population (Taylor et al., 2018; Warren-Jeanpierre et al., 2017; Schnall et al, 2018; Mitchel et al., 2018; Cruz-Oliver et al., 2019).

Previous research suggests several factors which may impact ACP discussions among African American PLHIV and their caregivers. HIV illness severity may be related to knowledge of ACP topics and likelihood of ACP (Maragh-Bass et al., 2017). Women are more likely than men to be informal caregivers; therefore, caregiver gender may be differentially associated with ACP likelihood (Taylor et al., 2018; Knowlton et al., 2015). African Americans are more likely than other groups to perform informal caregiving when they are managing chronic conditions themselves. Therefore, caregivers’ own health may be associated with their ACP knowledge and willingness to discuss with care recipients (Nguyen et al., 2019; Lyon et al., 2020). Finally, reciprocity (mutual giving and benefit) may be associated with quality of informal caregiving relationships, and potentially with ACP, where caregivers may be highly engaged in their care recipients’ medical treatment (Mitchell, Robinson, Wolff, & Knowlton, 2014). Recent findings by Lyon and colleagues (2020) suggest that caregivers are accurate sources of ACP knowledge about care recipients, even when preferences change over time.

Current Considerations

Compounding effects of COVID-19 and HIV, and low levels of ACP uptake creates syndemic consequences for African Americans (Poteat et al., 2020; Lyon et al., 2019). Historical and existing inequities for African Americans have left them particularly vulnerable to poor COVID-19 outcomes which may be particularly true for older African Americans living with HIV and other co-morbid conditions that can compromise immune status. Current data estimate that African Americans make up 30% of COVID-19 cases among states with available racial data (Poteat et al., 2020). The disproportionate prevalence of COVID-19, coupled with the uneven occurrence of underlying conditions, African Americans have experienced disproportionate COVID-19 morbidity and mortality (Poteat et al., 2020; Millett et al., 2020). HIV burden remains highest among African Americans, making them synergistically more vulnerable to COVID-19 health consequences. With many public health services shifted due to the pandemic, people living with HIV may lose access to essential HIV services, potentially harming medication adherence and increasing viral loads. Uncontrolled viral loads among people living with HIV, and disproportionate COVID-19 mortality among African Americans creates urgency for uptake of ACP particularly among older African Americans living with HIV at disproportionate risk of poor HIV health outcomes and simultaneous COVID-19 risk.

Synthesis and Purpose

ACP rates among African Americans remain lower than all other groups. HIV informal caregivers may be crucial conduits to increasing ACP discussion among African American PLHIV, particularly among PLHIV who are socioeconomically disadvantaged and have a history of substance use (Maragh-Bass et al., 2017; Hong et al., 2018; Clark et al., 2018; Taylor et al., 2018; Warren-Jeanpierre et al., 2017; Schnall, Hirshfield, Liu, Siegel, & Gradilla, 2018; Mitchel et al., 2018). The present study explored ACP among a sample of predominantly African American informal caregivers of African American PLHIV to understand: (a) the role of gender in likelihood of ACP discussion with care recipients; (b) factors associated with ACP discussions with care recipients; and (c) caregivers’ own knowledge and preferences related to ACP.

METHODS

Overview of study

Informal caregivers completed baseline surveys and interviews as part of the larger Affirm study, the purpose of which was to: identify caregiver, recipient, and network factors related to HIV caregiving continuity; assess effects of informal and formal HIV care on caregivers and care recipients; and explore treatment goals and end-of-life care preferences with mutual understanding between PLHIV, caregivers and healthcare providers (Maragh-Bass et al., 2017; Mitchell et al., 2018). Inclusion criteria for main participants (care recipients) were: (a) age of 18 years or older; (b) documented HIV seropositive status; (c) use of HIV medication defined in the prior 30 days. Inclusion criteria for caregivers were: (a) age of 18 years or older; and (b) invitation for participation from a care recipient. Care recipients were recruited through community sampling, and an HIV-specialty clinic (Maragh-Bass et al., 2017; Mitchell et al., 2018). Many care recipients and caregivers were individuals who had previously injected drugs, though this was not an inclusion criterion.

Data collection

Data were from the Affirm study and were collected through a mixed methods approach, which included qualitative in-depth interviews and quantitative surveys. Affirm study data were collected between August 2014 and October of 2015, at a research center located off-campus near an academic hospital. This academic center is located in the Southeast US, in an urban predominantly African American city. Written consent was obtained from all participants, none of whom had personal relationships with the study team. The principal investigator (ScD, MPH) and three trained researcher team members (all MPH-level) conducted interviews and focus groups. The study was approved by the institutional IRB.

Qualitative data collection.

In-depth interviews elicited views about participants’ current health status including personal and care recipients’ history of substance use; HIV informal caregiving relationship factors; and knowledge of ACP. The research team also specifically asked about the Medical Orders for Life Sustaining Treatments (MOLST), because the state had recently mandated completion of the MOLST for all patients discharged to long-term care facilities (Maragh-Bass et al., 2017; Maryland MOLST, 2010). The MOLST includes preferences for cardiopulmonary resuscitation and other life-sustaining treatments, which are transferred across an individual’s care spectrum (Maryland MOLST, 2010). We concluded with discussion of caregivers’ own end-of-life preferences and interest in participation in future ACP interventions. All researchers minimized deviation from interview questions based on biases or assumptions. Most interviews (10 out of 11) lasted between 60 and 120 minutes. One interviewee requested to terminate the interview early due to a family emergency, therefore this interview was only 30 minutes in duration.

Quantitative data collection.

Participants completed computer-assisted personal interviews with trained interviewers (Nova Research Company, 2006).

Measures.

Dependent variable.

Participants were asked: “Have you ever talked with [care recipient participant] about what kinds of medical treatments s/he would want if s/he was unable to speak for her/himself?” Responses were binary, where 0 = never discussed ACP with care recipients, versus 1 = have discussed ACP with care recipient.

Independent variables.

Correlates assessed were: (a) socio-demographic factors: age, gender, health-related quality of life; (b) caregiving relationship factors: relationship to care recipient, care recipient-related HIV care challenges, time spent weekly caring for care recipient, care reciprocity; and (c) knowledge and attitudes related to ACP: knowledge of ACP-related topics, views on whether loved ones should be included in ACP, whether care recipient has a medical designee, and views on the best time to discuss ACP.

Sex was measured via, “Are you:” where 1 = Male, 2 = Female. The decision to refer to sex in the present study as gender interchangeably is because (1) over 90% of the sample were cisgender and heterosexual; and (2) qualitative participants used sex and gender interchangeably themselves, for example in discussion of ‘women’ versus ‘females.’ Next, age was assessed as a continuous variable, with “How old are you,” where participants reported their age in years. Health-related quality of life was assessed using the 12-item Medical Outcomes Study Physical Functioning Measure (SF-12; McDowell & Newell, 1996). Responses were on 3-point Likert scale, ranging from “Not at all” to “A lot”. Items 1, 8, 9, and 10 were reverse-scored and summed as a continuous scale, such that higher scores indicated higher levels of physical functioning (McDowell & Newell, 1996; Hagell & Westergren, 2011).

Relationship to care recipient was assessed via “What is your relationship to [care recipient]?”, and a variable was created where 0 = Spouse or current or former partner, 1 = biological child, grandchild, or parent, 2 = biological sibling, and 3 = Other (Friend/Other relatives/Other). Care recipient-related HIV care challenges were assessed via items gathered from several existing validated scales (Knowlton et al., 2015; Knowlton et al., 2011; Strauss, 1979; Pearlin, Mullan, Semple, & Skaff, 1990). The five-item scale included “In the last few months, how often has [care recipient] taken you for granted?” and “In the past few months, how often has [care recipient] taken his/her anger out on you?” Responses were on a 3-point Likert scale from “Never” to “A lot.” Based on the distribution of responses, all five items were summed, and a 4-level categorical variable was created, where 0 = no challenges, 1 = few challenges, 2 = some challenges, and 3 = many challenges.

Time spent caring for care recipient per week was assessed via “In the past few years, what was the most time you spent helping out [care recipient]…? About how many hours a week…?” Responses were categorized where 0 = 0 to 3 hours a week, 1 = 4 to 9 hours a week, 2 = 10 to 29 hours a week, and 3 = 30 or more hours a week. Care reciprocity was assessed via the Non-Reciprocal Social Relations Questionnaire (Chandola, Marmot, & Siegrist, 2007). Items included “In the past year, how much would you say [care recipient] expressed affection and appreciation to you?” Responses were on 4-point Likert scales. Based on the distribution of responses, items were summed and categorized where 0 = low, 1 = medium, and 2 = high reciprocity.

Knowledge of ACP-related topics was assessed via: “Have you heard of an advance directive or MOLST form?”, “Have you ever heard of hospice?”, and “Have you ever heard of palliative care?” Responses where summed (0 = knowledge of 0 topics, 1 = knowledge of any 1 topic, 2 = knowledge of any 2 topics, and 3 = knowledge of all 3 topics; Fong, 2014). “Inclusion of loved ones in ACP” was assessed via “Family and doctors can consider their own interests when making healthcare decisions for someone.” Responses were categorized, and on a 5-point Likert scale ranging from “Strongly disagree” to “Agree.” Medical designation was assessed with “Has [care recipient] ever named a medical decision maker or signed a document naming such a person?” Views on the best time to discuss ACP was assessed via “When do you think is the best time to talk about critical care decisions?” Responses were dichotomized at the median, where 0 = Before getting sick (still healthy), versus 1 = When first diagnosed/First sick/First hospitalized/at end-of-life.

Data analyses

Quantitative analyses.

Univariate frequencies were generated for all variables (N=311). Principal component analyses were conducted on the health-related quality of life, care recipient-related HIV care challenges, and care reciprocity scales. One-factor solutions were obtained with good reliability on all three scales (Cronbach’s α = 0.79, 0.80, and 0.70, respectively; McDowell & Newell, 1996; Hagell & Westergren, 2011). Variables marginally significant (p<.10) or significant (p<.05) at the bivariate level were entered into a multiple Poisson regression model, to regress ACP discussion on socio-demographic, caregiving relationship, and ACP knowledge-related variables (Gordon, 2012). Poisson regression is appropriate for binary outcome event with prevalence greater than ten percent, which in this case was over 40 percent (Gordon, 2012). Robust standard errors were used to account for wide dispersion in the data (Dean & Lawless, 1989). Adjusted analyses were only conducted on African American participants due to theoretical significance (N=267). Health-related quality of life and views on the best time to discuss ACP were retained as control variables in the final model, despite non-significance. Quantitative analyses were conducted only on complete cases (Bennett, 2001). Analyses were run in STATA Version 14.0 (StataCorp, 2015).

Qualitative analyses.

Analyses of qualitative data were conducted after statistical analyses, to contextualize quantitative findings. Completed interviews were professionally transcribed verbatim, and coded by the same graduate-level researchers who conducted data collection, using grounded theory and constant comparison methods and overseen by the principal investigator (StataCorp, 2015; Glaser & Strauss, 1967). Grounded theory is a qualitative approach where theory is inductively derived from social context; constant comparison refers to the iterative means by which the theory is formed (Glaser & Strauss, 1967). This process has three phases: (1) open coding, when tentative codes are assigned; (2) axial coding, to identify possible relationships between codes; and (3) selective coding, where data are distilled to a single, grounded theory (Glaser & Strauss, 1967). The research team met multiple times to ensure consensus coding and inter-coder consistency. Upon thematic saturation, analyses were discontinued. A finalized codebook was created, coded text was analyzed for recurrent themes to inform quantitative analyses, and exemplary quotes were then extracted. Analyses were conducted in Atlas.ti 7.0 (Scientific Software, 2012). Given that no current EQUATOR checklists exist which focus on mixed methods research, the present research utilized the Consolidated criteria for reporting qualitative research (COREQ) checklist in completion of data analyses and manuscript writing (Tong, Sainsbury, & Craig, 2007) (Supplementary File 1). After quantitative analyses, the study team to discuss thematic findings which were salient to quantitative analyses and constructs which had been collected in survey measures with caregivers (i.e., socio-demographic factors, caregiver relationship factors, and ACP knowledge). Each of these were examined as they arose in the qualitative codebook and theme generation, and coding reports were generated from relevant codes, and with focus on gender roles as related to HIV caregiving.

RESULTS

Quantitative results

Table 1 reports overall sample demographics and chi-square bivariate analyses stratified by gender. Forty-one percent of individuals had discussed ACP previously with care recipients (n=128). This finding was similar between men (42.2%) and women (40.2%). Table 2 reports adjusted correlates of ACP stratified by gender.

Table 1.

Socio-demographics of caregivers by gender (Affirm Study)

Characteristic	Men (N=147) N(%) or Mean (SD)	Women (N=164) N(%) or Mean (SD)	Total (N=311) N(%) or Mean (SD)
Race
African-American or Black	140 (95.2)	159 (97.0)	299 (96.1)
White	5 (3.4)	4 (2.4)	9 (2.9)
Other	2 (1.4)	1 (0.6)	3 (1.0)
Education
Less than high school diploma	62 (42.2)	65 (39.6)	127 (40.8)
High school diploma/GED	45 (30.6)	58 (35.4)	103 (33.1)
Some college/above	40 (27.2)	41 (25.0)	81 (26.1)
Relationship to Care recipient**
Spouse/Current or Ex-partner	49 (33.3)	34 (20.7)	83 (26.7)
Biological (Grand)child or Parent	24 (16.3)	50 (30.5)	74 (23.7)
Biological Sibling	18 (12.2)	25 (15.3)	43 (13.8)
Friend/Other Relatives/Other	56 (38.2)	55 (33.5)	111 (35.7)
Knowledge of ACP topics
0 topics	17 (11.6)	10 (6.1)	27 (8.7)
1 topic	56 (38.1)	61 (37.2)	117 (37.6)
2 topics	56 (38.1)	71 (43.3)	127 (40.8)
3 topics	18 (12.2)	22 (13.4)	40 (12.9)
Care recipient-related care challenges**
None	35 (23.8)	47 (28.7)	82 (26.4)
Few	24 (16.3)	46 (28.1)	70 (22.5)
Some	48 (32.7)	32 (19.4)	80 (25.7)
Many	40 (27.2)	39 (23.8)	79 (25.4)
Most time spent helping care recipientǂ
0–3 hours/week	34 (23.5)	50 (31.1)	84 (27.5)
4–9 hours/week	29 (20.0)	44 (27.3)	73 (23.9)
10–29 hours/week	38 (26.2)	31 (19.2)	69 (22.5)
30+ hours/week	44 (30.3)	36 (22.4)	80 (26.1)
Consider own wants in loved ones’ ACP
Strongly disagree/Disagree	67 (45.6)	62 (37.8)	129 (41.5)
Neither agree nor disagree	7 (4.8)	11 (6.7)	18 (5.8)
Strongly agree/Agree	73 (49.6)	91 (55.5)	164 (52.7)
Ever discussed ACP with care recipient
No	85 (57.8)	98 (59.8)	183 (58.8)
Yes	62 (42.2)	66 (40.2)	128 (41.2)
Care Support/Reciprocity
Low	52 (35.4)	49 (29.8)	101 (32.5)
Medium	57 (38.8)	57 (34.8)	114 (36.6)
High	38 (25.8)	58 (35.4)	96 (30.9)
Care recipient has Medical designee
No	93 (70.4)	92 (61.3)	185 (65.6)
Yes	39 (29.6)	58 (38.7)	97 (34.4)
Best time to discuss ACP
Diagnosed/sick/hospitalized/EOL	67 (45.6)	63 (38.4)	130 (41.8)
Before getting sick, while healthy	80 (54.4)	101 (61.6)	181 (58.2)
Health-related Quality of Life	18.4 (2.7)	18.9 (2.9)	18.7 (2.8)
Age (years)	51.8 (10.2)	51.1 (11.8)	51.4 (11.1)

Significant or marginally significant (χ²) at:

^ǂ(p<.10)

^*(p<.05)

^**(p<.01)

^***(p<.001)

Table 2.

Correlates of ACP discussion with care recipients by gender (African Americans only)

	Men (N=123)				Women (N=144)
	IRRa	CIc	AIRb	CI	IRR	CI	AIR	CI
Relationship to Index
(Grand)child/Parent	0.92	(0.55, 1.53)	0.56	(0.26, 1.17)	0.60*	(0.37, 0.98)	0.73	(0.44, 1.21)
Sibling	0.68	(0.33, 1.37)	0.54ǂ	(0.28, 1.04)	0.56ǂ	(0.29, 1.06)	0.49*	(0.27, 0.89)
Friend/Oth Kin/Others	0.64ǂ	(0.40, 1.02)	0.55*	(0.32, 0.95)	0.63ǂ	(0.39, 1.00)	0.92	(0.47, 1.79)
(ref: Spouse/Partner)	1.00		1.00		1.00
Knowledge of ACP
1 topic	0.95	(0.46, 1.97)	0.87	(0.44, 1.73)	0.44*	(0.23, 0.85)	0.24***	(0.11, 0.53)
2 topics	1.31	(0.65, 2.62)	1.25	(0.64, 2.45)	0.77	(0.44, 1.36)	0.31**	(0.13, 0.70)
3 topics	1.25	(0.56, 2.83)	1.76	(0.79, 3.91)	0.79	(0.40, 1.56)	0.29**	(0.12, 0.70)
(ref: 0 topics)	1.00		1.00		1.00
Care recipient challenges
Few	0.43ǂ	(0.16, 1.13)	0.32*	(0.13, 0.82)	0.83	(0.48, 1.42)	0.78	(0.41, 1.48)
Some	1.01	(0.61, 1.66)	0.74	(0.40, 1.34)	0.96	(0.54, 1.68)	1.09	(0.55, 2.17)
Many	1.13	(0.69, 1.85)	0.81	(0.43, 1.53)	1.17	(0.71, 1.91)	1.34	(0.59, 3.04)
(ref: None)	1.00		1.00		1.00
Most time help care recipient
4–9 hours/week	1.95ǂ	(0.89, 4.30)	1.42	(0.60, 3.36)	0.99	(0.53, 1.84)	1.26	(0.67, 2.37)
10–29 hours/week	2.71**	(1.32, 5.59)	2.62**	(1.30, 5.30)	1.69ǂ	(0.97, 2.93)	1.78ǂ	(1.00, 3.17)
30+ hours/week	2.19*	(1.05, 4.57)	1.39	(0.67, 2.87)	1.68ǂ	(0.99, 2.86)	1.88*	(1.04, 3.38)
(ref: 0–3 hours/week)	1.00		1.00		1.00
Cons wants in loved ones’ ACP
Neither Agree/Disagree	0.51	(0.09, 2.84)	0.71	(0.20, 2.45)	1.57	(0.86, 2.86)	2.16**	(1.22, 3.82)
Agree/Strongly agree	0.74	(0.50, 1.10)	0.60*	(0.39, 0.93)	1.00	(0.66, 1.51)	1.12	(0.73, 1.71)
(ref: Str dis/Disagree)	1.00		1.00		1.00
Care Support/Recip
Medium	0.98	(0.59, 1.61)	0.92	(0.55, 1.53)	0.91	(0.52, 1.59)	1.07	(0.59, 1.94)
High	1.58*	(1.00, 2.49)	1.71ǂ	(0.99, 2.95)	1.55ǂ	(0.97, 2.47)	1.83ǂ	(0.98, 3.41)
(ref: Low)	1.00		1.00		1.00		1.00
Care recipient has Med designee
Yes	1.47*	(1.00, 2.16)	1.04	(0.70, 1.57)	2.80***	(1.84, 4.26)	2.57***	(1.63, 4.04)
(ref: No)	1.00		1.00		1.00		1.00
Best time discuss ACP
Diag/sick/hospital/EOL	0.90	(0.61, 1.35)	1.02	(0.69, 1.49)	1.30	(0.86, 1.99)	0.71	(0.44, 1.14)
(ref:Before getting sick)	1.00		1.00		1.00
Health-related QoL	1. 01	(0.95, 1.09)	0.99	(0.91, 1.07)	1.01	(0.95, 1.08)	1.00	(0.92, 1.05)
Age (years)	0.99	(0.97, 1.01)	0.97*	(0.95, 0.99)	1.01	(0.99, 1.03)	1.02ǂ	(1.00, 1.03)

^ǂp<.10

^*p<.05

^**p<.01

^***p<.001

^aIRR = Incidence Rate Ratio

^bAIR = Adjusted Incidence Rate Ratio

^cCI = 95% Confidence Interval

Descriptive characteristics.

Participants were mainly African American (96.1%), and just over half were women (52.7%). Two-fifths had less than a high school education (40.8%). The mean age was 51.4 years (Standard Deviation = 11.1). Regarding ACP, nearly half of participants had no or little familiarity with topics such as advance directives (46.3%). Roughly half of participants agreed or strongly agreed that family should consider their own wishes regarding their loved ones’ ACP needs (52.7%). Nearly two-thirds reported that to their knowledge, care recipients did not have a medical designee (65.6%), and nearly 60% stated that the best time to discuss ACP was before getting sick (58.2%). Regarding gender differences as reflected in chi-square analyses (Table 1), women were nearly twice as likely as men to report that their relationship to the care recipient was biological child, grandchild, or parent (30.5% vs. 16.3%; p<.01). Compared to men, women were less likely to report frequent challenges when caring for care recipients (43.3% vs. 59.9%, p<.10).

Correlates of having ACP discussion with care recipients (Men).

Regarding negative correlates – friends had a 45% lower likelihood of ACP discussions with care recipients compared to spouses/partners (Adjusted Incidence Rate Ratio [AIR] = 0.55; 95% Confidence Interval [95% CI] = 0.32, 0.95; p<.05). A similar association was found among siblings as compared to spouses/partners (AIR = 0.54; 95% CI = 0.28, 1.04; p<.10). Compared to men who reported no challenges when caring for care recipients, men who reported a few challenges had nearly a 70% lower likelihood of ACP discussions with care recipients (AIR = 0.32; 95% CI = 0.13, 0.82; p<.05). Compared to men who disagreed or strongly disagreed that one should consider their own wants regarding loved ones’ ACP, men who agreed or strongly agreed had a 40% lower likelihood of ACP discussions (AIR = 0.60; 95% CI = 0.39, 0.93; p<.05). Increased age among men had a slight correlation with lower likelihood of ACP discussions with care recipients (p<.05). Regarding positive correlates - men who spent a maximum of 10 to 29 hours in a week caring for care recipients had nearly 3 times the likelihood of ACP discussions with care recipients compared to men who spent no more than 3 hours a week caring for care recipients (AIR = 2.62; 95% CI = 1.30, 5.30; p<.01). Compared to men who reported low care reciprocity from care recipients, men who reported high support and reciprocity had a 70% higher likelihood of ACP discussions with care recipients which was marginally signficant (AIR = 1.71; 95% CI = 0.99, 2.95; p<.10).

Correlates of having ACP discussion with care recipients (Women).

Regarding negative correlates - women who were siblings with care recipients had half the likelihood of ACP discussions as women who were spouses/partners of care recipients (AIR = 0.49; 95% CI = 0.27, 0.89; p<.05). Next, women’s increased familiarity with ACP topics was association with between 70 to 75 percent lower likelihood of ACP discussions with care recipients. Women who were familiar with 1 ACP topic had a 76% lower likelihood of ACP discussions with care recipients (AIR = 0.24; 95% CI = 0.11, 0.53; p<.001). Similarly, women who were familiar with all 3 ACP topics (advance directive, hospice, palliative care) had a 71% lower likelihood of ACP discussions with care recipients compared to women who had no familiarity with these topics (AIR = 0.29; 95% CI = 0.12, 0.70; p<.01).

Regarding positive correlates - women who had spent 10 to 29 hours in a week helping care recipients had nearly an 80% higher likelihood of ACP discussions than women who had spent no more than 3 hours a week helping care recipients (AIR = 1.78; 95% CI = 1.00, 3.17; p<.10). Similarly, women who had spent over 30 hours in a week helping care recipients had nearly 90% higher likelihood of ACP discussions with care recipients (AIR = 1.88; 95% CI = 1.04; 3.38; p<.05). Next, women who were neutral on whether one should consider their own wants in loved ones’ ACP had more than 2 times the likelihood of ACP discussions with care recipients (AIR = 2.16; 95% CI = 1.22, 3.82; p<.01). Compared to women who reported low care support and reciprocity from care recipients, women who reported high support and reciprocity had an 83% higher likelihood of ACP discussions with care recipients (AIR = 1.83; 95% CI = 0.98, 3.41; p<.10). Women who reported that care recipients had a medical designee had nearly 3 times the likelihood of ACP discussions with care recipients (AIR = 2.57; 95% CI = 1.63, 4.04; p<.001). Finally, increased age among women had a slight correlation with higher likelihood of ACP discussions which was marginally significant (p<.10).

Qualitative results

Qualitative analyses identified five themes: (1) Knowledge related to ACP; (2) Knowledge of care recipient’s ACP preferences; (3) Caregiver’s own preferences; (4) Perceived Need and Benefits of ACP discussion; and (5) Caregiving-related coping and social support. Table 3 reports example quotes corresponding to all 5 themes.

Table 3.

Qualitative themes and example quotes (Affirm study)

Theme	Quote
Knowledge related to ACP	Knowledge of MOLST: Well no actually, I haven’t of those forms, per se. What R_____ was—- he had mentioned it…his family was real adamant, you know… I’m not quite sure if he actually had a Maryland MOLST or whatever. What I do know is that there were a lot of friends that were there…I wasn’t that directly involved in it…Well I want to look more into this MOLST form, because that’s something that…it sounds to me that everyone living with HIV and/or AIDS, or anyone period, should be not only aware of but to take advantage of. I think that once we discuss that—- for me for the most part, I think that even though it can be trying to say the least, discussing medical issues with my family, I think that when we make it real clear, one thing that’s really clear, is that the goal is to help and not hurt. And once we have clarity in what it is that I want, and if you can help me then fine, if not, you know… – 45-year-old Man Knowledge of advance directives: Yeah, advance directives actually…we have attorneys that were able to assist most of our support group with advance directives. And they actually were binding legal documents that we were able to provide them with, just for future care. And so it was helpful to many of them who no, you know…even myself I have one. – 45-year-old Man [I never heard about those forms with Care recipient] Not with him, but I’ve seen forms [like that] and I’ve been exposed to them because like I said, my adopted son, his kidneys burst so he was in the hospital so that’s why I got to know about-- the lawyer talking about [advance directives]. – 65-year-old Woman Knowledge of palliative care: Well, I really haven’t heard about it [palliative care] until you just said it. I think I might have experienced it with my sister…like I guess the quality of life was going down with her, so, you know what I mean, it was more or less we just let her go, because that’s what she wanted. If that’s palliative care, tried to make her comfortable until eventually they left. She didn’t want to be all chemoed on this, that, and the other. She didn’t want that. – 53-year-old Man Are you talking about what they was going to do with him? They was going to put him in a place [like hospice? If not, no I’ve never heard of palliative care, but]… I see what you’re saying. I think I would be okay like that just say for instance they want to put me in a nursing home or whatever because…nobody can be able to take care of me at home. As long as my family came to see me every day then I will go into a nursing home. Or this place that you’re speaking about as long as my family came and see me every day nobody because I don’t trust the nursing home to give me as well or that place you talking about take care of me as well as my sister’s because they could wash me up. Or they could take me out on the porch, get me some air. They could bring me fruit. They can make sure I’m eating until I pass. – 59-year-old Woman
Knowledge of Care recipient’s ACP preferences	Care recipient’s care preferences: Mm-hmm. [Care recipient] jokes a lot, and because him and I have gone to not only family, his family and my family, his mom just recently died, and I was to go with him all the way to Boston, Massachusetts to bury her. And so going with him, discussing little thi—- you know, sometimes we discuss, it’s an opportunity sometimes to discuss future plans…I do know for a fact that he does have advance directives. I am not clear as to what they are …but I do know that he has family that’s involved in his aftercare. And that’s clear, and I’m sure <laughs> because of the conversation that we did have, that she probably is clear about what it is he would like. – 45-year-old Man …He talked about if somethin’ was to happen to him, he just wanted me to go on and I did-- you know, just like, we’d be talkin’ and he’d be like, “Oh, I don’t wanna hear that…he just want me to go on with my life and…that he wanted a funeral but his brother and them said that he-- they told him-- them that he wanted to be cremated…he always had told me his wishes was to be-- have a funeral but he said his brothers-- you know, family said that later on, he said he didn’t want me to know that. So that’s what he was. Other than that, that was, you know, pretty much it. He never really talked about passin’. – 65-year-old Woman My cousin, he hasn’t discussed that yet [with me]. Now that I think about it, maybe we need to discuss it just in case [because I don’t know his wishes.] don’t know if it would be easy to talk to him about it. This is talking about if anything happens to him so I don’t know if there would be an easy way to talk to him about it. But we will talk about it though because I think I need to know or somebody at least needs to know his wishes at least. – 56-year-old Woman Care recipient’s healthcare decision-maker preferences: I don’t got to go there [discussing MOLST and advance directives].. My husband [living with HIV] is 54 years old. He going to live with me. I take care of him. My mother she _________ and my niece will take care of her and I will take care of her so I don’t have to deal with nobody going to no nursing home. That’s not in my story…A lot of things with H.I.V. A lot of new things coming up. I don’t have no involvement in it because I don’t have that problem. I don’t have that problem for my husband. I don’t have that problem for my mom. So I don’t have no problem with somebody going to a nursing home and all that. I don’t have no problem about that because I don’t know nobody in my family that I had to sign to a home. Because I will take care of them before I put them in a home…if I would send them to a home they would get mistreated so I don’t go through that. – 36-year-old Woman Well, oh, sometimes we might not talk in depth about it, but he said if something happened, “---“ take care of everything.” That’s all he said, “Take care of everything. What you see fit, you do.”… when he was working, he put me down in his will of just that I’m his main contact person. Before his brothers and sisters, he got me as being the main contact person. – 53-year-old Man
Main supporter’s own preferences	Yeah. I don’t want to be on no machine. Just let me go…Me and my mom and dad. We talked about that. I told them, I just-- I don’t want the machine. Let us go. – 49-year-old Woman My wife. She knows what I want…[It’s not documented but we] Talked about it, yeah, whatever…After you send me out with the trash, have a memorial service. Don’t give me that, you know what I mean? The money is in the bank. Don’t spend it on no funeral…If my quality is there-- if my quality of life is going to go down, let me go, y’all. I don’t want you having to be dealing with me every day and I can’t do nothing. I don’t know what’s going on. I’m just alive. Yeah, let me go, let me go, and you move on. You might say that, but ain’t no leaving you in the bed in the hospital for 10 hours and ain’t nothing going to improve, or they send me home and you’ve got to take care of me. – 53-year-old Man [They] talked about it when I was in the hospital. They talked to me about it as far as when I was in the hospital…This has been about five years ago. [They said] That if I wanted to sign it if I needed someone to take care of things for me if I wasn’t able to do it myself. [I listed] My daughter. Well I thought of I just decided myself. They [the doctors] didn’t help me decide on who to pick. I decided on it myself… I wanted for them to do every measure possible to revive me. Don’t give up on me – 56-year-old Woman …No [no life support if in a coma], if I have dementia, God forbid, no I don’t want to put that on my family, on my children. They wouldn’t want to see me that way. No. I wouldn’t want to go through that. I don’t know what they could do about that part, though. But surely, if I have to be hooked up on the breathing machine to keep me living and while they hoping and praying I might come back I might because I know people. Even L_____ came back come to think of it. Even he came back. They put him on it because he was in a coma and he woke up and he came out of it and he was hooked up and screaming and they had to rush in there and unplug him and everything. And he came back. Ain’t that something? I just remembered that. And he even came back off of the machine. Some people come back and some people don’t but I don’t want my family worrying over me every day hoping I come back and all of this stuff. – 59-year-old Woman
Perceived Need and Benefits of ACP Discussion	[If you discuss ACP with others] well they would at least know what you want, at least, your preferences and everything like that. How to take care of you or what to do… don’t know if it would be easy to talk to him [Care recipient] about it. This is talking about if anything happens to him so I don’t know if there would be an easy way to talk to him about it. But we will talk about it though because I think I need to know or somebody at least needs to know his wishes at least… it’s difficult. It’s talking about if something happens to you. Me, myself would think if something happens to me, I don’t want to talk about if something happens to me. I’m hoping nothing would happen to me but I know that’s not possible but people wouldn’t want to talk about that if something happens to me, I want this done, I want that done. So I still say it’s no right or wrong way. It’s something needs to be discussed. – 56-year-old Woman I don’t know [if I need to discuss ACP]. I say, when is the right time? I don’t know, because I don’t never want to talk about it. It’s a simple fact. If the mind conceive it, then you’re going to achieve it. So, don’t start talking about death. Then I might die. See, look, even though I know that’s addictive, we’re all going to keep. So, who should be in the room? I guess the doctor, the patient, and the significant others in his life. – 53-year-old Man [I might discuss ACP with] probably the head of my family, my sister, the oldest sister, maybe because she’s the one that is the head of the family. And she’s retired now but she used to be a healthcare provider. She used to be a nurse. She used to be an RN and worked with special doctors for all her life ever since she was young. Now she’s retired.[There is benefit to telling her my care preferences because] we close as peanut butter and butter. She’s very religious. She goes to church and she’s a very committed in the community and all of this stuff. And she’s all for me. She been knowing me all my life, I’m the baby. I’m the baby. She’s the oldest…She went through a lot with me, my whole family, always supported me, always wanted the best. – 59-year-old Woman
Caregiving-related coping and Social support	Support from Loved Ones (Including Care recipientes): [We’re mutually supportive]. Pretty much though, when we first started dialoguing with one another it was more like a one-way street and I didn’t mind. So like I’m doing all the work to get him on the right path on a one-way street [taking care of Care recipient]. But now it became a two-way street, which is good because even though he’s suffering with what he’s suffering with [HIV], I’m suffering from what I’m suffering with and— we talk, you know what I mean? – 53-year-old Man You know, when I first began helping others [as an HIV caregiver], it was always easy to give advice, it was always easy to say, “Don’t give up.” You know, “Create [ph?] everything that’s in you,” and, “You must fight,” and.. But when you have to put on those shoes, it’s not always that easy. So resorting to that information, resorting to those words of encouragement that I gave others and using it for myself has been a challenge sometimes, but certainly worthwhile. To have someone that I’ve helped reciprocate it to me, and remind me, “Look at me now,” W_____ has a Master’s degree, he went to _____ University, we went together to – I remember – to register for classes some 10 years ago, almost. And I never—- to be honest with you, I honestly didn’t think that he would continue; he got his Bachelor’s, and then to go back and get his Master’s, shows that, you know, if he can do it, I can. And now I’m in classes, I’ve registered for classes, because he returned those same inspiring words that I gave him, and the encouragement. And now I’m using that, what he gave me, what I gave him, you know? It works, it’s a two-way street, two-way street. – 45-year-old Man [Our family supports Care recipient and me a lot].. See back when it first came out it was a death sentence; and everybody didn’t want to be around people. They wasn’t real familiar with the disease or what it was all about or how you got infected or whatever. So-- but she was never shunned. You know?..We’re-- I can say we might be-- well financially wise we are not-- I don’t know what to say. But when it comes to family we’re a family. We’re there for each other. More support out there that way. – 74-year-old Woman Community support: Well I’ve been positive for 12 years. In my first year I started wanting to know more so I started taking the classes. So I got all my certificate from there where I can go anywhere that I choose to teach about H.I.V. I can go to colleges, high schools…I am able to go to any site to do H.I.V. testing and educate them, let them know there’s nothing to be ashamed about it… I have a support group that I do at my house.. our support group where you have we have it at each other’s houses…I’m doing good in educating and support because at my clinic it’s also support groups down there. They do a support group every day down at --- for their clients that’s H.I.V. positive… – 36-year-old Woman Yeah, [I go to] two different [support groups]… They said that we’re supposed to get [to know] each other and we’re supposed to help each other. Say like if I get up and I’m going through something and I’m just walking, not speaking, not doing my usual routine, they encourage us just to see what’s going on with the other person, if it’s no more than saying, “Good morning. How are you doing? Maybe I can help you with something. Do you want to pray?” Yeah, they encourage it, building back the relationships. Because when we was out there we was, what’s the word, reclusive…[and it helps me] because, you know, I had a lot of trust issues, a lot of people let me down that I depend on and every class is a open discussion and I look back on it now and I find myself, I would never say anything, open my mouth, I would just sit there and sit there. And I would say, “I’m tired of hearing this, I heard this yesterday and then day in and day in and day in and day.” And we have a pastor down there too…he was one of the founders… – 51-year-old Man Religious support/Spirituality: I got on my knees in front of the church and I was just coming from getting me some drugs and church was going and they were singing and the song they were singing was “Amazing Grace”. And I dropped on my knees and started crying and the drugs and stuff that I had just bought I threw it and I ain’t picked it back up since. So that was my biggest challenge meaning I was wanting to do that for ______. It took me a while to want to stop self-medicating myself and I felt that God did that for a reason, had took me put me in the direction of that church, and He had me hear them people singing that opened my eyes to let me know that is not [HIV is not a death sentence]… – 36-year-old Woman He’s doing it. He stopped getting high. Church helped [us both] the most as far as that was concerned. He goes to the doctor’s for substance abuse, that helps him out also. He’s been doing it so far, as far as I know he’s been doing it well. – 56-year-old Woman I pray a lot. I pray a lot because I know who saved my life, I know the reason I’m here is to help somebody else, even though I go through what I go through and I’ve been through what I’ve been through. I know my purpose now is to help somebody else, preferably, I can’t say that word right, preferably another drug addict. – 59-year-old Woman

Demographic characteristics.

The qualitative sample derived from the larger quantitative sample consisted of 11 participants, all of whom were African American. Most of the participants were women (n=8). The mean age was 52 years. Three participants disclosed being HIV-positive themselves in addition to caring for a care recipient living with HIV. Gender roles as related to caregiver identity are highlighted in Theme 5.

Theme 1: Knowledge related to ACP.

Interview participants varied in their ACP knowledge. This variation was unrelated to whether care recipients had recently sought medical care. Advance directives was most familiar to interviewees, relative to other ACP topics. One interview participant had completed an advance directive himself. Another mentioned that she learned about it while caring for her adopted son, not the care recipient for whom she provided HIV care.

Unlike advance directives, palliative care was the most unfamiliar to participants. When participants were unfamiliar with the term the following definition was read verbally to them: “palliative care is medical care that helps to improve quality of life both for patients and their families. It can also to assist patients and their families to make treatment decisions with their health providers.” After hearing this definition, participants discussed experiences with loved ones where their quality of life was reduced at end of life, including but not limited to AIDS-related conditions. One woman discussed the need for palliative care in the context of her own end-of-life preferences. She stated that because quality of care is poor in hospice facilities, she would only be willing to go to a nursing home at end of life because her ‘nobody can be able to take care of me at home.’ Even then, she stated need for family members to help care for her and maintain some quality of life.

…I’ve never heard of palliative care, but… I see what you’re saying. I think I would be okay like that just say for instance they want to put me in a nursing home or whatever because…nobody can be able to take care of me at home. As long as my family came to see me every day then I will go into a nursing home…I don’t trust the nursing home to give me as well or that place you talking about take care of me… – 59-year-old Woman

Theme 2: Knowledge of care recipient’s care preferences.

Many supporters knew care recipients had documented their care preferences, however, most did not know specifics about their preferences. Supporters generally described circumstances which led to them learning about care recipients’ preferences, and virtually none had directly asked the care recipient about their preferences. One supporter mentioned that while he knew ‘for a fact’ the care recipient has advance directives; the care recipient had mostly included family members in outlining his preferences. The supporter learned that about the care recipient’s advance directive by asking about his experiences with his own mother’s end-of-life care.

…[Care recipient’s] mom just recently died, and I was to go with him all the way to Boston, Massachusetts to bury her. And so, going with him, discussing … “How do you feel? What would you have done differently?” Or, “What would you like to do?” This, this, that. “Would you like to be cremated? [so]…I do know for a fact that he does have advance directives. I am not clear as to what they are…I was not present at that time… – 45-year-old Man

One woman mentioned that her care recipient (who had passed away) had volunteered his care preferences in the context of burial, rather than healthcare planning. Another woman, during the course of interview, stated that she felt prompted to ask the care recipient’s preferences even though she was unsure ‘if there would be an easy way to talk to him about it.’

Most caregivers were not named as their care recipient’s medical designee. One woman, who was their care recipient’s main decision-maker, saw no need to discuss ACP topics because it was not ‘in my story’ and she planned to take care of her family members.

… My husband [living with HIV] is 54 years old. He going to live with me. I take care of him…I don’t have to deal with nobody going to no nursing home…A lot of things with H.I.V. A lot of new things coming up. I don’t have no involvement in it because…I don’t have that problem for my husband. I don’t have that problem for my mom…Because I will take care of them before I put them in a home…. – 36-year-old Woman

Theme 3: Caregivers’ own care preferences.

Many supporters had clearly articulated preferences related to end-of-life care and had discussed them with their loved ones. This finding was similar among those who were living with HIV, as well as those who were HIV-negative. Among those who had clearly defined preferences, many stated their preferences was related to their quality of life. In general, these supporters said if their health was severely deteriorated, ‘let me go.’ One man had shared his preferences with his wife, and that if he didn’t know what was going on and is just alive, ‘you let me go and move on.’ He stated that he did not want funeral expenses to be incurred for his burial. One woman stated that if she had dementia for example, she would not want her family worrying every day about her health status and condition. In contrast, one woman was very adamant that she be kept alive as long as possible. She stated that she had discussed this with her daughter (her medical designee) and also her healthcare provider years ago, when she was hospitalized (unrelated to HIV). As she explained:

…This has been about five years ago. [They said] That if I wanted to sign it if I needed someone to take care of things for me if I wasn’t able to do it myself. [I listed] My daughter. Well I thought of I just decided myself. They [the doctors] didn’t help me decide on who to pick. I decided on it myself… I wanted for them to do every measure possible to revive me. Don’t give up on me – 56-year-old Woman

Theme 4: Perceived benefits of ACP discussion.

Caregivers generally saw benefit to discussing ACP preferences with loved ones, from their own perspective and also that of their care recipient. Supporters stated that there was need to discuss ACP with whomever would be the decision-maker for the individual. In this vein, several supporters identified family members as their intended decision-maker, who were usually their children or sibling. One woman stated that her older sister who was a registered nurse would be the person with whom she would discuss preferences, because she ‘always supported me.’

[I might discuss ACP with] probably the head of my family, my sister, the oldest sister, maybe because she’s the one that is the head of the family…She used to be an RN… [There is benefit to telling her my care preferences because] we close as peanut butter and butter. She’s very religious. She goes to church and she’s a very committed in the community and all of this stuff. And she’s all for me… – 59-year-old Woman

Another woman saw perceived benefit to discussing her ACP preferences with others because ‘at least they would know what you want’ although she perceived difficulty in discussing ACP with the care recipient for whom she cared. One man did not perceive need to discuss ACP because he ‘don’t never want to talk about it.’ He felt that discussing ACP issues might hasten the actual need for said type of planning. Many supporters, including these two individuals, felt there was no right or wrong time to discuss ACP. However, most discussed perceived benefit to discuss ACP in the context of knowing an individual’s wishes before something happens.

Theme 5: Caregiving-related coping and social support.

Caregivers commonly discussed three dimensions of support that they, as well as care recipients in their lives accessed. First, caregivers shared that they received social support from loved ones, predominantly from the care recipient they cared for. They described this reciprocity of support as mutual, and mostly in the form of emotional support. One man stated that at first, he helped to get the care recipient more comfortable with his HIV medical regimen. Over time, however, the care recipient began reciprocating emotional support and they currently help each other because they were both ‘suffering,’ albeit with different issues. Similarly, another man discussed having to humble himself to receive support, because he was more used to giving support as an HIV caregiver to many individuals. He stated that the education attainment that he saw the care recipient achieve inspired him to go back to school. These and other individuals all described the care recipient-supporter relationship as a ‘two-way street.’

Second, supporters discussed the availability of community support. Many supporters were either currently or previously enrolled in HIV education classes or support groups. Notably, all of the supporters who had completed educational courses were also HIV-positive. They discussed need to learn about managing their own illness as well as caring for others. Other supporters discussed being enrolled in support groups because they sometimes needed help with managing the stress of caregiving responsibilities along with other demands. One man discuss that support groups were helping him, because he did not trust people.

Third, religious support and spirituality were mentioned as frequent coping mechanisms for both care recipients and caregivers. The previously mentioned individual, who was reluctant to join a support group, discussed that the reason he was willing to go to the support group was because his pastor founded it, like the religious support that many supporters mentioned. A few caregivers discussed going to church as a crucial regimented activity which helped them with receiving emotional support related to HIV. However, many of the supporters (like care recipients) had a history of injection drug use. Therefore, many discussed religious ties as a means by which they ended their substance use and stopped ‘self-medicating’ to manage their issues. Several also prayed to cope with their own health issues (HIV or otherwise), and with the stress of caregiving. Many discussed its positive impact on them and their care recipient.

Gender-related caregiving identity and perceptions.

Caregivers were not asked specific questions related to their gender. However, gender identity became a salient part of their discussion about experiences as caregivers to individuals living with HIV. Gender roles related to being ‘a woman,’ were discussed in two main ways. Both men and women caregivers described that women often take a caretaker role in their lives even when it is to their own detriment. One woman explained that the trauma of coping with her HIV diagnosis kept her from being able to do what she needed to do to care for her family “as a woman.” One man explained that the care recipient he cared for often put her main partners need of both of theirs, even though they were both HIV-positive and the partner did not help with managing the household. The man explained that this as “the woman gives and the man takes.”

So, [Care recipient living with HIV] she is ill, but she is able to maintain her own house…she has a boyfriend…something ain’t right with him too, you know, because I know that--he’s got to be [HIV-positive too because he’s] very, very small…he’s taking her to too much change. And I told her, “With your disease, you don’t need that. You don’t need that.”… not paying her money to help her out, and he want to lay there and don’t want to do nothing. I told her, “You don’t need to go through that.” …You know what I’m saying? It’s not-- the woman gives [and the man takes]. – 49-year-old Man

‘Being a strong woman’ also arose when a caregiver discussed the death of her care recipient. She stated that his mother was ‘such a strong woman’ who never cried during her child’s end of life, which spanned several days. She also supported him as he struggled to overcome a drug addiction before he was diagnosed with HIV.

…we had to look at him for a couple of more days deteriorate. His mother I could just say she’s such a strong woman. That was her only child. I didn’t see a tear fall from her eye. I remember I kept staring at her because I was thinking about her…And she went through a lot with him before he even got sick and stuff… – 59-year-old Woman

Discussion of ‘being a man’ arose in relation to being a provider for one’s family. Both men and women discussed this provider identity related to men, in contrast to perceptions of women as caretakers. Notably, discussion of ‘the man serving as a provider’ was not directly linked to individuals’ experiences as caregivers. For example, one man discussed not being able to “be a man” and provide for himself and his family, as his father had expected him to do. He stated that struggling with his HIV status and drug use had estranged him from his father and other family members, and that he was still working through his shame from neglecting these responsibilities.

I feel ashamed…I’m going through this relationship class…to get back at it because they teach you not about where she going…They say it, “You have to imagine what we did to them, and what they’re going through.” It’s not always about us…And I feel ashamed because he-- because my father was a man. And all he wanted me to do was just be a man. He said…you need to get yourself together, and come in and help me with this.” And I never did… – 51-year-old Man

DISCUSSION

The present research explored advance care planning among African American informal caregivers, including understanding the role of gender in ACP outcomes. Our research fills a critical gap in knowledge about the views and experiences of African Americans living with HIV from the vantage point of their caregivers. Quantitatively, we found that just forty percent of individuals had discussed ACP previously with their care recipients (Table 1).

Results revealed several factors associated with ACP discussions in the larger quantitative sample on which descriptive and bivariate analyses were conducted, (N=311), the sample with which adjusted quantitative analyses were conducted (N=267), and in the smaller qualitative sample (n=11), several of which differed by gender. First, we found that nearly three-fourths of the overall quantitative sample were unfamiliar with all three ACP topics on which they were surveyed (advance directives, hospice, palliative care). Similarly, none of the qualitative sample was familiar with all three ACP topics discussed (advance directives, living wills, palliative care). Given that two-fifths of the surveyed population had less than a high school education, these findings are in consistent with previous studies which suggest that knowledge about ACP is low overall, and particularly among racial minorities who have limited educational attainment (Hong et al., 2018; Clark et al., 2018). Reasons might include limited health literacy and numeracy, which may be explored further in future research to promote ACP among older African Americans with HIV.

Qualitatively, participants were most familiar with advance directives, and least familiar with palliative care (Table 3). Knowledge of all three ACP topics was not quantitatively associated with ACP discussions with care recipients among men (Table 2). Surprisingly, among women, knowledge of all three topics was associated with a nearly 80 percent lower likelihood of ACP discussions (p<.01). It is possible that women had more experiences caring for individuals with chronic conditions, given that women are more likely than men to perform these duties (Taylor et al., 2018; Knowlton et al., 2015). As a result, they may have been familiar with these topics because of having provided care for more individuals and may have felt less need to discuss these topics with caregivers because they were already knowledgeable about them and ‘knew what to do’ themselves. Future research should explore the role of caregiver gender and willingness to engage in ACP with care recipients, given African Americans’ cultural preferences for life-sustaining treatments and that African American women may provide and be expected to provide care more than their male counterparts (Kamal et al., 2017; Maragh-Bass et al., 2017; Mitchell et al., 2018; CDC, 2015).

Caregivers who were their care recipient’s sibling were roughly half as likely to have discussed ACP compared to main partners or spouses. This finding was the same among men (p<.10) and women (p<.05). Although sibling caregivers are relatively understudied (Thomas, Liu, & Umberson, 2017), recent research found that sibling caregivers often feel less affected by the health needs of the care recipient as compared to main partners and spouses (Namkung, Greenberg, & Mailick, 2017). Post-hoc analyses revealed that siblings spent less time in a given week caring for care recipients as compared to caregivers and spouses (p<.001); it is possible that siblings were less involved in all aspects of the care recipient’s care and therefore less likely to have discussed ACP. Further research is needed to promote engagement of siblings who care for persons living with HIV in resource-deprived settings.

Next, quantitative findings showed men who agreed that others should be considered in one’s ACP had a 40 percent lower likelihood of discussing ACP with care recipients compared to men who disagreed (p<.05). Conversely, women who neither agreed nor disagreed had more than twice the likelihood of ACP discussions as women who disagreed (p<.01). It is possible that male supporters who were unclear on their own ACP wants were less willing to discuss these topics with care recipients; however, women who felt their wants and needs were unrelated to the care recipient’s ACP may have been more willing to initiate this discussion.

Caregiver’s age also had an inverse relationship with ACP planning among men and women. Increased age was inversely associated with lower likelihood of ACP discussion among men (p<.05); however, a positive association was found among women caregivers (p<.10). Given that previous research has identified age as a predictor of increased likelihood of ACP discussions (CDC, 2015; Hong et al., 2018), future research is needed to explore the needs of older men caregivers in facilitating ACP discussions with their care recipients.

Qualitatively, caregivers generally saw benefit to discussing ACP with care recipients regardless of gender. Much of the extant research on HIV caregiving and gender has been conducted in international settings; nonetheless our previous research identified female sex may be related to informal care preferences and greater presence of informal caregivers (Mitchell, Robinson, Nguyen, Smith, & Knowlton, 2015). Existing research has identified potential differences in communication needs among men and women caregivers of other chronic illnesses (Schwartz & McInnis-Dittrich, 2015;Mazanec, Reichlin, Gittleman, & Daly, 2018); therefore, future interventions should consider gender-specific communication skill building in HIV-related ACP.

Finally, among both men and women, more time spent with care recipient (>10 hours) and more caregiving support reciprocity from care recipients was associated with a higher likelihood of ACP discussions (p<.10). Qualitatively, we found that many caregivers discussed that their caregiving relationship with care recipients was a ‘two-way street’ where support was provided and received mutually. Our previous research has identified caregiving reciprocity as protective for both caregivers and care recipients which can prevent caregiving cessation (Knowlton et al., 2011). However, given that well under half of the overall sample had discussed ACP with care recipients, more research is needed to caregiving reciprocity as a pathway to encourage ACP among African American HIV caregiving relationships.

Limitations

Several limitations exist in consideration of our study findings. First, the outcome of interest, having ever discussed ACP with care recipients is very broad. While, nuances of how, reasons for, or frequency of discussions were not ascertained; nonetheless, ACP definitions vary widely and our analyses were intentionally broad in scope. Second, “male” and “men”, (e.g. gender and sex) were considered equivalent both theoretically and analytically, although sex is biological while gender is sociocultural and more subjective. Furthermore, gender is not binary, therefore findings do not reflect the views of other gender identities; nonetheless, nearly 90% of the study sample were cisgender and heterosexual. Further, many participants in qualitative data collection used sex and gender interchangeably themselves. Third, all study data were cross-sectional, which does not permit causal inferences over time in ACP-related discussion between care recipients and caregivers. Fourth, other variables of interest could explain additional variance in the outcome, such as controlling for caregivers’ perceived willingness to discuss ACP, and/or caregivers’ involvement in other aspects of participants’ life beyond HIV. Religiosity, healthcare satisfaction, caregivers’ health status, and caregivers’ substance use history were all assessed with the outcome of interest but were not associated with ACP discussion.

Next, restricting the analytic sample to African American main participants and stratifying by gender may have led to a loss of statistical power to detect significant findings (although this was a theoretically-informed decision); as a result, some marginally significant associations are shared but must be interpreted with caution. Finally, care recipient participants in the larger Affirm study were nearly all African American, middle-aged, mostly injection drug users, insured, and enrolled in medical care. Due to the homogeneity of this sample and its implications for limited variation among caregivers, we did not adjust for several socio-demographic variables (e.g. income, insurance). Thus, while middle-aged African American, injection drug users, that are insured, and enrolled in medical care, are often excluded in research, these population characteristics limit the generalizability of our study findings.

Study Implications

A systematic review by Weathers and colleagues suggests that existing caregiving interventions are generally lacking in rigor and underrecruit African Americans for participation (Weathers et al., 2016). Additional innovations are sorely needed; faith-based interventions may be a viable setting for future ACP planning to address the cumulative barriers of HIV stigma and medical mistrust in this community. Our recent research exploring faith-based approaches to health program interest among older African Americans qualitatively found that faith/spirituality were instrumental to coping with life challenges such as HIV and substance abuse, and that faith leaders were trusted sources of health information (Maragh-Bass, Sloan, Alghanim, & Knowlton, 2020). We also found that satisfaction with religion/spirituality was quantitatively associated with greater interest in a future health programs which may focus on quality of life (Maragh-Bass et al., 2020). Future ACP interventions should consider educational components delivered by faith leaders, who are often called upon to consult older African Americans for ACP-related decisions. Similarly, the FAmily-CEntered (FACE) Advance Care Planning trial currently being conducted is comparing physician-based to team-based delivery of an ACP intervention (Lyon et al., 2019; Lyon et al., 2020). Future iterations of this work, along with our findings might explore faith leaders as part of a team-based approach to increasing ACP among older African Americans with HIV and their caregivers.

Also, in a recent trial of the Sharing Patient’s Illness Representations to Increase Trust (SPIRIT) ACP intervention improved caregiving depressive symptoms, reduced decisional conflict, and improved ACP in African American caregiving relationships (Maragh-Bass et al., 2017; Knowlton et al., 2011; Song et al., 2016). Our findings suggest that future replication of the SPIRIT intervention should also consider skill-building to promote care reciprocity, and amount of weekly caregiving time as important facilitators of ACP discussions in HIV caregiving relationships among both men and women caregivers.

Conclusions

The present research is one of few studies to explore ACP planning among informal caregivers in African American populations who are the hardest hit by HIV (Maragh-Bass et al., 2017). Results suggest that, regardless of gender, caregiver time and reciprocity positively influence ACP discussion, whereas, caregiver sibling relationship with care recipients negatively influenced ACP discussion. Among women, care recipient older age and agreement with involving a loved one was positively associated with ACP discussion, whereas ACP knowledge was negatively associated with ACP discussion. Conversely, in men care recipient’s agreement with involving a loved one and older age was negatively associated with ACP discussion. This underscores the need for education to document preferences and foster communication skills to include caregivers in future decision-making. Moreover, recent findings by Lyon and colleagues (2020) suggest that caregivers are accurate sources of knowledge about their care recipients ACP planning wishes; therefore, education must be informed and culturally tailored to understanding the dyadic experiences/knowledge and dynamics of both parties (Lyon et al., 2019, Lyon et al., 2020). Future research and intervention should also consider the role of gender in caregiving needs, identity, and experiences; both qualitative and quantitative results suggest that men and women view and experience caregiving differently.

Older adults of color living with HIV are highly vulnerable to COVID-19 infection, and least likely of all racial groups to have documentation in line with the care needs and wishes (Poteat et al., 2020; Millet et al., 2020). Given the current issues in health inequities that COVID-19 has further illuminated in older African American communities, research to improve ACP and care planning outcomes in this community are more urgent than ever.

SUMMARY BOX.

‘What does this paper contribute to the wider global clinical community?’

• We have much to learn to promote advance care planning (ACP) for older African Americans with HIV, which refers to documentation such as advance directives and is associated with increased quality of life (QoL).

• For older African Americans living with HIV, unpaid caregivers provide a substantial amount of care and support. Therefore, education and skill-building with older African Americans living with HIV, and their caregivers is a critical pathway to increasing both ACP rates and QoL in this population.

• When engaging caregivers, gender-specific communication is important, given that African American women are providing, and expected to provide informal care more often than their male counterparts. Results underscore need for ACP education which includes healthcare providers, given African Americans’ strong cultural preference for life-sustaining treatments at end-of-life.