Relationships between Advance Care Planning Engagement, Patients’ Religious Practices, and Spirituality

Abstract

Background:

Research has shown that people who consider themselves religious have a lower advance directive (AD) completion rate. However, advance care planning (ACP) includes a broader range of activities that are important in determining medical treatments. Little has been published on the relationship between spirituality, religion, and engagement in ACP activities.

Objective:

We examined whether religiosity and spirituality are associated with ACP engagement.

Methods:

This is a secondary analysis of interview data collected at baseline during a randomized controlled trial of a nurse-led ACP intervention. Measures included the 5-item Duke University Religious Index (DUREL), a measure of spirituality, and the 15-question ACP Engagement Survey, which examines engagement in multiple aspects of ACP.

Results:

We found a significant association between higher scores on each domain of the DUREL and higher ACP Engagement scores. In separate multivariable models adjusting for demographic factors, the relationship between greater religiosity (higher total DUREL scores) and greater ACP Engagement scores was significant (p<.0001 for each model). Identifying as spiritual was significantly associated with greater ACP Engagement scores (p=.0038) in adjusted multivariate models.

Conclusions:

In contrast to prior findings of lower AD completion in more religious individuals, this study found that higher religiosity and spirituality are associated with higher ACP engagement. These results suggest ACP opportunities offered for religious congregations or spirituality focused communities may be especially successful.

The study is registered at ClinicalTrials.gov with identifier NCT04070183 (https://clinicaltrials.gov/study/NCT04070183).

Introduction

Since the development of advance directives (AD) and the Patient Self-Determination Act of 1990, there have been a variety of initiatives to empower adults to complete them. ADs are legal documents, such as a Living Will, Power of Attorney for Health Care, or Health Care Representative form, which take effect when an adult loses decisional capacity. These forms are often regulated by individual states’ legislative bodies. There have been extensive publications on the completion rate of such documents and studies of program to assist and motivate completion of an AD.¹ Despite initiatives to encourage use, completion of ADs in adults remains at less than 50%.^2,3

Legal documents alone do not adequately prepare adults and surrogates for future complicated medical decisions.⁴ A more expansive view has evolved that considers other components of planning for future medical treatments. Advance care planning (ACP) is “a process that supports adults at any age or stage of health in understanding and sharing their personal values, life goals, and preferences regarding future medical care.”⁵ ACP is the process that empowers the patient to reflect and share why they have made the choices they have made and may also include preparing surrogate decision makers for future decisions.⁶ The ACP Engagement Survey was created to measure a person’s engagement in these ACP activities within the framework of established models of behavioral change. The goal was to develop a survey which examined a person’s readiness to engage in ACP, from having knowledge to being ready to act.⁷

Religion and spirituality have been found to be important factors in ACP^3,8. Values-based decisions and questions about who has the authority to make medical decisions may be influenced by religious and spiritual beliefs^9,10.

Prior studies that have examined the relationship of religion to AD completion have found mixed results. Studies have found that those who report religion is very important to them have a lower completion rate of AD, but those who attended religious service at least once a week had higher odds of completing an AD.^8,11 Other studies found that patients who used religion to help them cope¹² or believed that a higher power controls life length,⁹ were less likely to have an AD.

Little is known about whether religion influences readiness to engage in ACP. In a study of Muslim participants attending mosque seminars, the Duke University Religion Index (DUREL) and Psychological Measure of Islamic Religiousness (PMRI) were used to explore the relationship between Islam and ACP engagement.¹³ This study specifically focused on a single religious group. Other studies using the ACP Engagement Survey have asked participants about religion, religious behaviors, and spirituality but have not assessed a relationship between these variables and ACP Engagement scores.^4,14–18

To address this gap in the literature, this study examines the relationship between ACP engagement and religiosity and spirituality. Because ACP engagement is especially important for patients at increased mortality risk, we focused on older adults with chronic illness.

Methods

Participants & Setting

This is a secondary analysis of structured interview data collected during the Planning Ahead study, a randomized controlled trial of a nurse-led ACP intervention for community-dwelling older adults (age 65 or older) with serious illness.¹⁹ Eligible patients had an estimated mortality risk of 30% in the next year, according to the Gagne Index²⁰ or other measures of mortality based on specific chronic illnesses such as renal disease, heart disease, pulmonary disease, cancer, and dementia. Patients were ineligible if they were acutely ill, enrolled in hospice, or unable to communicate in English. Participants were recruited through primary care clinics in two large health systems in a Midwestern US city. All data for the present study were obtained during the enrollment interviews conducted prior to randomization with patients who had the capacity to make their own medical decisions. Additional details have been published elsewhere.¹⁹

The Planning Ahead study was approved by the Indiana University Institutional Review Board; research staff confirmed participants’ understanding and cognitive capacity prior to obtaining consent for participation.

Measures

All interview data were collected by telephone interview with a member of the research staff. Participants provided information about their religion (What is your religion?) and their spirituality (Do you consider yourself to be a spiritual person?). Participants who identified themselves as “Protestant” were invited to share their denomination.

Additional religion variables were collected using the Duke University Religion Index (DUREL).²¹ This 5-item index assesses 3 domains of religiosity: Organized Religious Activity (ORA), Non-Organized Religious Activity (NORA), and Intrinsic Religiosity (IR). The ORA item asks how often participants attend church or other religious meetings. The NORA item asks about private religious activities such as prayer. IR is assessed through 3 statements about religious beliefs where participants can respond how true or not true each statement is in their own life. Scores for both ORA and NORA range from 1–6, with higher scores indicating more frequent participation in religious activities. Scores for the 3-item IR domain range from 3–15, with higher scores signifying a greater degree of religiosity.

The 15-item validated ACP Engagement Survey was administered to measure participants’ self-efficacy (confidence) and readiness to engage in 4 different ACP topics: choosing a medical decision maker, discussing treatment they would or would not want if they were very sick, how much flexibility to give decision makers, and asking doctors questions.²² Participants responded with their level of confidence to discuss each of those domains on a 5-point Likert scale from “Not at all” confident (1) to “Extremely” confident (5). They also rated their readiness to engage in these discussions and activities through another 5-point scale (“I have never thought about it [1],” “I have thought about it, but I am not ready to do it [2],” “I am thinking about doing it in the next 6 months [3],” “I am definitely planning to do it in the next 30 days [4],” and “I have already done it [5]”). Each question on the ACP Engagement Survey has a possible score of 1–5, and the overall and subscale scores are obtained by calculating the mean of constituent items. Higher scores on the ACP Engagement Survey indicate greater engagement in ACP.

The health literacy of participants was assessed through a validated 3-item questionnaire.²³ Participants were given Likert response choices for questions. Higher scores indicate greater health literacy.

Statistical Analysis

Bivariable analyses examined the association between ACP Engagement Survey scales (overall and subscales) and religion variables, using Wilcoxon Rank Sum or Kruskal-Wallis tests for categorical religion variables and Spearman correlation analyses for continuous religion variables. Given that skewness of the overall ACP scale was mild, general linear multivariable models were used to examine the association of ACP Engagement scores with the spirituality item, the total DUREL scores, and other covariates of interest, with each religious variable in a separate model due to high collinearity. General linear bivariable models for the ACP overall scale were presented alongside the multivariable models to assist in interpreting the impact of covariates. All analytic assumptions were verified. All analyses were performed using SAS v9.4 (SAS Institute, Cary, NC).

Results

Study Participants

Data were collected from 364 participants. The mean age was 74.7 and 61% were female. Most interviewees were either Black (35.4%) or White (62.1%; Table 1). Participants were asked about their level of comfort with their household income with 12.1% saying they did not have enough to make ends meet, 44.8% saying they had just enough to make ends meet, and 43.1% saying they were comfortable. About one third (33%) of the cohort had a high school education or less, with the largest group (52.4%) having some college education, and 14.6% having a graduate degree.

Table 1:

Participant and Survey Characteristics

ACP engagement 15 item (n=363)	3.80 (0.85); 3.86 (1.31 – 5); 3.23 – 4.50
ACP engagement: Medical Decision Maker domain (n=363)	3.59 (1.30); 3.67 (1 – 5); 2.7 – 5
ACP engagement: What Matters Most in Life domain (n=363)	3.89 (0.93); 4 (1 – 5); 3.2 – 4.6
ACP engagement: Flexibility domain (n=361)	3.71 (0.96); 3.75 (1 – 5); 3 – 4.5
ACP engagement: Asking Questions domain (n=362)	4.10 (1.01); 4.5 (1 – 5); 3.5 – 5
Organized Religious Activity (ORA)a (n=361)	3.46 (1.84); 4 (1 – 6); 1 – 5
Non-Organized Religious Activity (NORA)a (n=359)	4.20 (1.62); 5 (1 – 6); 3 – 5
Intrinsic Religiosity (IR)a (n=362)	12.68 (3.18); 14 (3 – 15); 12 – 15
Religion (n=363)
Protestant	282 (77.7)
Catholic	52 (14.3)
Other	8 (2.2)
None	21 (5.8)
Protestant sects (n=281)
Liberal	16 (5.7)
Moderate	56 (19.9)
Conservative	25 (8.9)
Other	177 (63.0)
Don’t Know	7 (2.5)
Patient considers themselves spiritual (n=353)
Yes	317 (89.8)
No	36 (10.2)
Age (n=364)	74.71 (6.42); 73.59 (64.11 – 94.15); 69.34 – 78.68
Gender (n=364)
Male	142 (39.0)
Female	222 (61.0)
Race(n=364)
Black	129 (35.4)
White	226 (62.1)
Asian	1 (0.3)
Multi	4 (1.1)
Refused	4 (1.1)
Income (n=364)
Not enough to make ends meet	44 (12.1)
Enough to make ends meet	163 (44.8)
Comfortable	157 (43.1)
Education (n=364)
8th grade or less	8 (2.2)
Some high school (no diploma)	36 (9.9)
High school/GED	76 (20.9)
Some college (no degree/cert/trade)	99 (27.2)
Associate’s degree	42 (11.5)
Bachelor’s degree	50 (13.7)
Master’s degree	43 (11.8)
Doctorate degree	10 (2.8)
Health Literacyb (n=363)	9.13 (2.94); 10 (0 – 12); 7 – 12

Values are frequencies (percentages) for categorical variables and means (standard deviations); medians (ranges); IQRs for continuous variables.

^a)Domains from the Duke University Religion Index (DUREL)

^b)Chew Health Literacy scale

Religion and Spirituality

Most participants identified themselves as Protestant Christians (77.7%). Other participants identified as Catholic (14.3%), another religion (2.2%), or no religion (5.8%). When asked if they considered themself to be a spiritual person, most respondents said yes (89.8%).

Participants had a mean score of 3.46 (out of 5) on the DUREL ORA item about church or religious meeting attendance (standard deviation [SD] = 1.84), suggesting that most attended religious meetings between a few times per year and a few times per month. A mean score of 4.2 (out of 5) on the DUREL NORA item about personal prayer and meditation (SD = 1.62) suggested that most participants reported praying or meditating between twice a week and daily. For the 3-item intrinsic religiosity domain about how religious beliefs affect one’s life, respondents had a mean score of 12.68 (out of a possible 15) (SD = 3.18).

ACP Engagement

On the ACP Engagement Survey, participants had a mean score of 3.80 out of 5 (SD = 0.85). Respondents had similar scores when the 4 domains were calculated separately. The mean score for the Medical Decision Maker sub-scale was 3.59, the score on the What Matters Most in Life sub-scale was 3.89, the score for the Flexibility sub-scale was 3.71, and the highest level of engagement was shown for the Asking Questions sub-scale with a score of 4.1.

Religion and ACP Engagement

Bivariable analyses of the ACP Engagement Survey and each of the three DUREL domains showed significant associations between higher levels of ACP Engagement and each domain of religiosity (Table 2). Higher levels of ACP Engagement on the Flexibility domain and the Asking Questions domain were also significantly associated with greater religiosity on all three domains of the DUREL. The Medical Decision Maker domain and the What Matters Most in Life domain had a significant association with the organized religious activities (ORA) and intrinsic religiosity (IR) domains but not the non-organized religious activities domain (NORA).

Table 2:

Associations between domains of religiosity and ACP Engagement Survey domains

Scale	Spearman correlation	P-value
ACP Engagement Survey (15 item)
Organized Religious Activity (ORA)a	0.206	<.0001
Non-Organized Religious Activity (NORA)a	0.157	.0029
Intrinsic Religiosity (IR)a	0.239	<.0001
ACP (Medical Decision Maker sub-scale)
ORA	0.145	.0060
NORA	0.101	.0553
IR	0.142	.0069
ACP (What Matters Most in Life sub-scale)
ORA	0.145	.0060
NORA	0.098	.0640
IR	0.159	.0025
ACP (Flexibility sub-scale)
ORA	0.200	.0001
NORA	0.124	.0191
IR	0.260	<.0001
ACP (Asking Questions sub-scale)
ORA	0.173	.0010
NORA	0.223	<.0001
IR	0.278	<.0001

^a)Domains from the Duke University Religion Index (DUREL)

In separate multivariable models, the relationship between greater religiosity (i.e., higher total DUREL scores) and greater ACP Engagement scores remained significant (p<.0001) when controlling for age, race, sex, education, health literacy, comfort with income, and illness severity (Table 3).

Table 3:

Relationship between religiosity, spirituality and other covariates with the outcome of overall advance care planning engagement

	Bivariable analysis		Multivariable analysis Model 1 (includes spirituality)		Multivariable analysis Model 2 (includes religiosity)
Spirituality
Yes	3.86 (0.05)	P=.0064	3.80 (0.11)	P=.0035
No	3.45 (0.14)		3.36 (0.17)
Religiosity (DUREL total score)	0.04 (0.01)	P<.0001			0.04 (0.01)	P<.0001
Age	0.01 (0.01)	P=.3577	0.01 (0.01)	P=.3847	0.01 (0.01)	P=.4365
Race
Black	3.72 (0.07)	P=.3121	3.57 (0.11)	P=.4749	3.70 (0.09)	P=.1575
White	3.86 (0.06)		3.68 (0.09)		3.87 (0.07)
Other	3.66 (0.28)		3.48 (0.29)		3.57 (0.28)
Sex
Female	3.80 (0.06)	P=.9307	3.56 (0.13)	P=.6909	3.68 (0.11)	P=.4279
Male	3.81 (0.07)		3.60 (0.14)		3.75 (0.12)
Education
<=HS	3.66 (0.08)	P=.0775	3.50 (0.14)	P=.2181	3.63 (0.12)	P=.3374
Some coll	3.87 (0.06)		3.68 (0.13)		3.78 (0.11)
Post-grad	3.88 (0.12)		3.55 (0.18)		3.72 (0.16)
Health lit	0.02 (0.02)	P=.2449	0.02 (0.02)	P=.3342	0.01 (0.02)	P=.5104
Comfort with income
Not enough	3.79 (0.13)	P=.3128	3.62 (0.18)	P=.7770	3.71 (0.16)	P=.9985
Just enough	3.73 (0.07)		3.53 (0.13)		3.71 (0.11)
Comfortable	3.88 (0.07)		3.58 (0.14)		3.71 (0.12)
Illness
Severity (Gagne Index)	0.05 (0.02)	P=.0040	0.04 (0.02)	P=.0572	0.05 (0.02)	P=.0141

Values are general linear model raw coefficients (standard errors) which represent means (for bivariable models) and adjusted means (for multivariable models) for categorical predictors and slope parameter estimates (standard errors) for continuous predictors. P-values are from general linear models (i.e., ANOVA/ANCOVA models, respectively, for bivariable and multivariable models). DUREL is the Duke University Religion Index.

There was no significant association between ACP Engagement scores and religious affiliation for any of the ACP Engagement domains or for the ACP Engagement Survey overall.

Spirituality and ACP Engagement

Identifying as a “spiritual person” was also significantly related to greater overall ACP Engagement (those identifying as spiritual had a mean ACP Engagement score of 3.93 [range= 1.31–5] and those who said they were not a spiritual person had a mean ACP Engagement score of 3.52 [range= 1.50–5]; p=.0041; Table 4). A significant association was also shown between spirituality and two of the ACPE domains: What Matters Most in Life and Flexibility.

Table 4:

Religion groups/sects and spirituality with ACP Engagement

	ACP Overall Scale (15 item)	ACP (Medical Decision Maker sub-scale)	ACP (What Matters Most in Life sub-scale)	ACP (Flexibility sub-scale)	ACP (Asking Questions sub-scale)
Religion
Protestant	3.86 (1.31 – 5)	3.67 (1 – 5)	4 (1 – 5)	3.75 (1 – 5)	4.5 (1 – 5)
Catholic	4.07 (2.36 – 5)	3.83 (1.33 – 5)	4.2 (2.4 – 5)	3.88 (2.25 – 5)	4.5 (2.5 – 5)
Other	3.64 (2.57 – 4.77)	3.67 (1.33 – 5)	4.1 (2 – 5)	3.25 (2.5 – 5)	4 (1.5 – 5)
None	3.43 (2.10 – 4.93)	3.33 (1 – 5)	3.6 (1.75 – 5)	3.5 (1 – 5)	4 (2 – 5)
	P=.1456	P=.2822	P=.0736	P=.7921	P=.4335
Protestant sects
Liberal	4.14 (2.57 – 5)	3.67 (1 – 5)	4.4 (2.6 – 5)	3.67 (2.75 – 5)	4.5 (2.5 – 5)
Moderate	3.86 (1.86 – 5)	3.67 (1 – 5)	3.8 (1.6 – 5)	3.5 (1.5 – 5)	4 (1.5 – 5)
Conservative	4.14 (1.31 – 5)	4 (1 – 5)	4 (1 – 5)	4 (1 – 5)	5 (1 – 5)
Other	3.79 (1.50 – 5)	3.67 (1 – 5)	4 (1 – 5)	3.88 (1 – 5)	4.5 (1 – 5)
Don’t Know	3.64 (2.85 – 5)	3 (1.67 – 5)	3.8 (3.2 – 5)	3.5 (3 – 5)	5 (2.5 – 5)
	P=.5621	P=.7706	P=.3797	P=.5014	P=.6345
Spiritual
Yes	3.93 (1.31 – 5)	3.67 (1 – 5)	4.2 (1 – 5)	3.88 (1 – 5)	4.5 (1 – 5)
No	3.52 (1.50 – 5)	3.67 (1 – 5)	3.5 (1.2 – 5)	3.25 (1 – 5)	4 (1.5 – 5)
	P=.0041	P=.1708	P=.0110	P=.0060	P=.0665

Values are medians (ranges) for ACP across the religion/spirituality groups, with p-values from Wilcoxon-Kruskal-Wallis tests.

Identifying as spiritual was significantly associated with greater ACP Engagement scores (p=.0038) even when controlling for age, race, sex, education, health literacy, comfort with income, and illness severity (Table 3).

Discussion

This study of predominantly Christian community-dwelling older adults at high risk of mortality found positive associations between overall ACP Engagement and three dimensions of religiosity. This association persisted when controlling for demographic factors and illness severity. There were also associations between ACP Engagement subscales, including Choosing a Medical Decision Maker, What Matters Most, Flexibility, and Asking Questions and most domains of religiosity. Previous research addressing various dimensions of religion has found mixed results. Studies have found lower levels of ACP participation among those with a stronger belief in God’s control over the length of life⁹ or the role God plays at the end of life,¹⁰ and another found no association between AD completion and frequency of religious attendance.² A study among Muslims that, similar to ours, compared the DUREL and ACP Engagement Survey, found there was no correlation.¹³ In contrast, several studies found that those who attended religious services at least once a week had higher odds of completing an AD. ^3,8,11 Notably, we found that religious affiliations, including religious faith and denomination among Christian participants, were not associated with ACP Engagement scores.

One possible explanation for the mixed findings in the literature is that there is wide variability in how ACP is defined as well as how it is measured in research.²⁴ There is even inconsistency in what constitutes completing an AD. We see similar inconsistency in how a participant’s religious or spiritual identification is evaluated. Colenda, et al.’s review of religious variables in ACP research reiterates the challenges of inconsistency in ACP and religion research.²⁵ In the same way that a narrow view of ACP does not examine how much a person has engaged in the ACP process, a narrow view of religion and religiosity does not examine practices or spirituality. For example, many studies have shown that those who are religious have a low AD completion rate and often want more aggressive medical interventions.^10,25,26 It is important to consider how specific dimensions of religious experience align with specific ACP behaviors and medical decisions about life sustaining treatments.

An important question is why individuals with high religiosity or spirituality are more engaged in ACP. One possibility is that religion encourages values reflection, and ACP conversations are essentially values-based conversations. People who are spiritual may also be more willing to consider life-limiting illness or death because their spirituality helps them to cope with these topics. While future research needs to explore these hypotheses directly, existing literature does suggest that religion and spirituality are important coping mechanisms during illness and death.^{27 28}

One implication of these findings is that faith communities are a promising avenue for promoting ACP activities. There are a variety of ways to encourage community-based education in ACP, from games^29,30 to community engagement events.^31,32 There have been studies focused on promoting ACP in faith communities,³³ some tailoring the education to the particular faith or cultural group, such as African American,^34,35 Asian American,³⁶ Korean American,³⁷ or Muslim.³⁸ Each study emphasized how important it was to discuss ACP in the context of the religious tradition. Considering these results, it is important to continue ACP initiatives in religious communities.

Religious leaders and those who attend religious services are open and receptive to participating in ACP training opportunities. Two separate qualitative studies examined religious leaders’ perceptions of ACP, and one also included church attendees. Pereira-Salgado et al. interviewed faith leaders in the Buddhist, Christian, Hindu, Islamic, Jewish, Sikh, and Bahá’í traditions.³⁹ Sloan et al. focused on African American Christian pastors and parishioners.⁴⁰ Religious leaders are receptive to helping the people in their religious community engage with ACP activities if they receive more training. Parishioners also agreed that the church could be a place for ACP training, education, and act as a resource. Parishioners stated the importance of having someone who could advocate for them to help with understanding how to interact with the medical system. The healthcare system can be challenging to navigate; thus, someone with some healthcare experience can be helpful in the ACP training. Heirali et al.’s model involves education that combines family members with ACP stories, religious leaders, and health professionals.³⁸ Park’s model uses faith community nurses.³⁷ Faith leaders can, and should, be an integral part of ACP conversations in the faith community.

Some research identifies “religion/spirituality” as a dual term, but religion and spirituality are becoming more distinct constructs.^41–43 A novel finding of our study is that we found self-described spirituality was associated with ACP engagement. We are unaware of prior studies that have examined this. In most studies comparing religion and engagement with ACP, spirituality is not delineated. Participants are asked their religious affiliation,^2,9,12,14 their religiosity,^8,11 how often they attend services,^2,3 religious practices,¹⁰ religious coping,¹² or spiritual values.⁹ Our findings suggest it is important to differentiate between religiosity as practiced through organized religious activity and intrinsic spirituality.

Religious and spiritual inquiry has traditionally been an integral part of models of ACP conversations. One of the facilitator’s questions in the Respecting Choices model explores how spiritual beliefs might affect care choices.⁴⁴ PREPARE For Your Care’s Heath Care Directive has a place under “What Matters Most in Life” to indicate if religion or spirituality is an important part of life.⁴⁵ The National Institute on Aging, in their conversation guide, inquires about spiritual or religious beliefs.⁴⁶ There are also various websites that provide resources to religious communities. The Conversation Project has a section of their website dedicated to resources for faith communities to facilitate ACP conversations; ⁴⁷ the Shomer Collective provides a website for Jewish resources on ACP.⁴⁸ Religion and spirituality have routinely been an important part of ACP exploration, and this study confirms their importance.

Limitations

A key limitation is that almost all of our participants identified as either Black or White, making it difficult to compare our results to people in other racial categories. Additionally, 92% of our participants identified as either Protestant Christian or Catholic. Only about 2% identified with any other religion, and only 5.8% said they were not religious. These results differ from the United States as a whole where recent data show around 29% of adults consider themselves to be religiously unaffiliated.⁴² Future research should include a sample with greater religious diversity.

Another limitation is that we did not ask our participants more specific questions about how their religious beliefs relate to their healthcare. For example, some studies have shown that saying their religious beliefs influence their medical decisions or having a belief that God controls the length of a person’s life may be related to participation in ACP.⁹ Others have examined how various forms of religious coping¹² or views of God’s role at the end of life¹⁰ may influence ACP decisions. Asking these types of questions in conjunction with the ACP Engagement Survey could provide additional insight into what aspects of religious belief are most related to ACP Engagement.

Conclusion

Our findings that people with greater religiosity, including personal and organizational participation and intrinsic religiosity, and people who describe themselves as spiritual are more likely to engage in ACP suggest that religious and spiritual organizations have the potential for achieving high rates of ACP. Effective programs should consistently address the importance of religion and spirituality in the ACP process.