INTRODUCTION
The end of life (EoL) is the phase when an individual's health conditions have deteriorated to the point where there is no option for curative treatment. In this last stage of life, individuals may need to make decisions about whether they would want or not want specific life-sustaining treatments and whether they would share their preferences with their family members and healthcare professionals (Sizoo et al., Reference Sizoo, Pasman and Buttolo2012). EoL communication between a patient and family is a difficult yet crucial process in advance care planning (ACP) that promotes the quality of the patient's last days (McMahan et al., Reference McMahan, Knight and Fried2013). Conventional notions of ACP focus on how patients explore available options for EoL care and select their preferences (Sudore & Fried, Reference Sudore and Fried2010). These preferences are often expressed in an advance directive (AD), the legal document (i.e., living will) that allows a patient to indicate his/her preferred treatments and designate surrogate decision makers (i.e., durable power of attorney). Newer frameworks put more emphasis on the importance of the “process” by which the patient identifies his/her values and goals regarding EoL care and communicates those attributes to his/her family and healthcare professionals (Sudore & Fried, Reference Sudore and Fried2010; Tulsky, Reference Tulsky2005). Within this framework, the family's understanding of their loved one's values and goals is especially important because they often play the role of surrogate, representing the patient's wishes in the event he/she has become cognitively incapacitated (Braun et al., Reference Braun, Beyth and Ford2008; Meeker, Reference Meeker2004).
The positive impacts of patient–family EoL communication are profound—for example, it enhances the family's satisfaction with the patient's care and reduces psychological distress and burden among surviving relatives related to making EoL decisions (Abbott et al, Reference Abbott, Katherine and Sago2001; Detering et al., Reference Detering, Hancock and Reade2010; Fried & O'Leary, Reference Fried and O'Leary2008; Wenrich et al., Reference Wenrich, Curtis and Shannon2001). Conversely, poor communication yields a variety of adverse effects, such as psychological distress, depression, anxiety, poor personal relationships, and caregiving burden among caregivers (Braun et al., Reference Braun, Beyth and Ford2008; Conner & Chase, Reference Conner and Chase2015). Despite such considerable benefits, EoL communication has been reported at only 55–63% among patients and community-dwelling older adults in the United States (U.S.) (Abbott et al., Reference Abbott, Katherine and Sago2001; Black et al., Reference Black, Fogarty and Phillips2009; Hopp, Reference Hopp2000; Phipps et al., Reference Phipps, True and Harris2003). The challenges for EoL communication are multifaceted, encompassing individual and social as well as cultural aspects. These challenges include an individual's reluctance to face his/her mortality, discomforts surrounding the topic of death and dying, relationship concerns, and cultural taboos regarding EoL discussions (Kehl et al., Reference Kehl, Kirchhoff and Kramer2009; Kreling et al., Reference Kreling, Selsky and Perret-Gentil2010; Schickedanz et al., Reference Schickedanz, Schillinger and Landefeld2009).
It has been well-documented that cultural differences frame notable differences when they come to EoL care. Carr (Reference Carr2012) found that Hispanics/Latinos are less engaged in both formal and informal ACP as compared to whites—they are less likely to complete an AD and less likely to discuss EoL care with others. Hispanics/Latinos typically value collectivism (Cruz-Oliver et al., Reference Cruz-Oliver, Talamantes and Sanchez-Reilly2014; Kelley et al., Reference Kelley, Wenger and Sarkisian2010). In line with this, Kelley and colleagues (Reference Kelley, Wenger and Sarkisian2010) found that older Latinos prefer family-centered EoL decision making. Hispanics/Latinos who have resided closer to their country of origin may have held more traditional cultural beliefs and values, which in turn may have shaped their unique beliefs and attitudes toward EoL care. Yet, little is known about this regionally and culturally distinct group of people with regard to patient–family EoL communication. Thus, our study aims to explore EoL communication and its contributing factors among family caregivers of Latino home healthcare recipients in the rural U.S.–Mexico border region.
METHOD
Study Design and Sample
This study is a secondary analysis of existing data that were collected for a hospice care needs assessment by a healthcare agency in the southern Californian U.S.–Mexico border region. The procedure and methodology utilized in the collection of original data are described in an earlier work (Ko et al., Reference Ko, Lee and Martinez2017). In brief, a convenience sampling was used to recruit participants. The agency staff contacted the family caregivers and asked about their interest in participating in a survey during a home visit. Of those 210 participants who completed the survey, only caregivers of Latino patients were included for analysis, which resulted in excluding 21 caregivers of non-Latino patients (final N = 189). This study was approved by the San Diego State University Institutional Review Board.
Data Collection
Data were obtained from in-person interviews and our electronic database between November of 2014 and August of 2015. Face-to-face interviews with family caregivers were conducted by bilingual staff using a structured questionnaire, in either English or Spanish according to the participant's preference. The patients' sociodemographic and health-related information were derived from the electronic patient database operated by the agency.
Measures
The dependent variable for this study was EoL communication. The independent variables were: preference for life-sustaining treatment (LST), comfort with EoL discussion, distrust toward physicians, knowledge of ADs, attitudes toward EoL care, and family caregivers' and patients' sociodemographic and health-related variables.
EoL communication was measured by asking the caregivers if they had ever discussed with the patient his/her wishes for extending life with life supports as he/she got sicker or was unable to take care of him/herself (Ko & Lee, Reference Ko and Lee2009). The response categories were “yes” and “no.”
Preference for LST was assessed by asking the caregivers whether or not they wanted their loved one to be alive as long as possible even if he/she would be dependent on life supports. The response categories ranged from 1 = “definitely no” to 4 = “definitely yes,” and the first two categories were collapsed into “no” and the latter two into “yes.”
Comfort with EoL discussion was assessed by asking the caregivers “How comfortable are you discussing the types of treatments your loved one wants during the EoL stage (e.g., whether he/she wants to prolong his/her life with life supports)?” The response categories ranged from 1= “definitely no” to 4 = “definitely yes,” and they were dichotomized into “yes” and “no.”
Distrust toward physicians was assessed by asking the caregivers “Do you worry that the doctor might want to stop treatments or ‘pull the plug’ too soon if your loved one were very sick?” This item was adapted from Ludke and Smucker (Reference Ludke and Smucker2007). The modified response categories ranged from 1 = definitely no to 4 = definitely yes, which were then dichotomized into “yes” and “no.”
Knowledge of ADs was assessed by asking the caregivers if they had ever heard about ADs. The definition of ADs was subsequently given and their response again verified. The response categories were “yes” and “no.”
Attitudes toward EoL care were assessed by asking the caregivers to what extent they agreed with the following statements: (1) talking about death can make people lose hope, and (2) losing hope makes people die sooner (Manu et al., Reference Manu, Mack-Biggs and Vitale2013). The response categories ranged from 1 = “strongly disagree” to 5 = “strongly agree.”
Caregivers' sociodemographic information was obtained by asking the caregivers about their age; gender (0 = female, 1 = male); education (1 = less than high school to 4 = college graduate); income (1 = $0–14,999 to 4 = $45,000 and above); marital status (0 = not married, 1 = married); and religion (2 = Catholic, 1 = non-Catholic-related religion, 0 = atheist). Patients' sociodemographic information included age, gender, and marital status; their health-related information included the number of chronic diseases and activities of daily living (ADL). Medical records were reviewed to identify chronic diseases—both physical (e.g., heart disease, lung disease, diabetes) and mental issues (e.g., anxiety, depression). Patients' medical information was provided by their physician when they were referred for the agency's home care services. ADL was measured by home care staff for six functional areas: dressing upper, dressing lower, bathing, toilet transferring, transferring, and ambulation. The total ADL score could range from 0 to 27, with a lower score indicating fewer impairments in overall ADL functioning.
Data Analysis
Descriptive statistics were used to report sociodemographic and health-related characteristics of the family caregivers and the patients for whom they cared. Bivariate tests (t and chi-square) were performed to compare the means and distributions of the study variables by the occurrence (yes or no) of EoL communication. Then, logistic regression analysis was conducted to identify significant predictors of EoL communication among sociodemographic and health-related variables, preference for LST, comfort with EoL discussion, distrust toward physicians, knowledge of ADs, and attitudes about EoL care.
RESULTS
The sociodemographic and health-related characteristics of the family caregivers and patients are reported in Table 1. The mean age of the caregivers was 56.5 years. A majority (78.3%) of the caregivers were female, and nearly two-thirds (63%) were married or living with a spouse/partner. The majority had less than high school (42.9%) or high school education (27%). More than half of the caregivers (58.2%) reported an annual income less than $15,000, followed by 23.6% between $15,000 and 29,000. Almost all (96.8%) had a religion, with the majority (75.1%) identifying as Catholic. In regard to their relationship to the patient, 50.3 and 25.9% were a child or spouse of the patient, respectively.
Table 1. Sociodemographic and health-related characteristics (N = 189)
* Patients could report more than one chronic disease.
ADL = activities of daily living.
On average, the patients were 76.7 years of age. The majority (56.1%) of the patients were female, and almost half (49.7%) were married or living with a spouse/partner. The patients reported on average two chronic issues: 59.3, 46, and 21.7% had diabetes, heart problems, and kidney disease, respectively; and 29.1, 22, and 17% had cognitive impairment, depression, and anxiety, respectively. The patients experienced moderate difficulties with ADL (M = 15.92, SD = 5.63). Bivariate analysis examined the associations between EoL communication (yes/no) and the caregivers' and patients' sociodemographic and health-related variables. The results of t and chi-square tests indicated no significant associations.
EoL Communication by LST Preference, Knowledge of AD, Comfort for EoL Discussion, Distrust toward Physicians, Attitudes toward EoL Care
Overall, about half (50.8%) of the family caregivers reported having ever had an EoL discussion with the patient. Table 2 presents the bivariate test results of comparing EoL communication (yes/no) with regard to EoL care-related variables. An EoL communication was more likely for the caregivers who did not want their loved one prolonging life with life supports (χ2(1) = 7.78, p < 0.01), and also for those who had heard about ADs (χ2(1) = 23.10, p < 0.001). In addition, the caregivers who reported feeling comfortable with EoL discussions were more likely to actually engage in EoL discussions (χ2(1) = 4.12, p < 0.05). Conversely, the caregivers who worried that physicians might want to stop treatments (i.e., “pull the plug”) too soon were less likely to engage in EoL discussions (χ2(1) = 12.70, p < 0.001). Other EoL-related variables were not significantly associated with EoL communication.
Table 2. EoL communication by EoL care-related variables (N = 189)
AD = advance directive, EoL = end of life, MD = medical doctor.
Factors Associated with EoL Communication
A logistic regression analysis was conducted to identify significant predictors of EoL communication among the family caregivers of our Latino patients. Table 3 presents the variables included in the model and the regression results, including the estimated odds ratio and its 95% confidence interval (CI 95%). The caregivers' preference for extending life with LST, knowledge about ADs, and distrust toward physicians significantly predicted EoL communication with the patient. That is to say, the caregivers who knew about ADs were almost six times more likely to engage in EoL communication than those who did not (OR = 5.50, CI 95% = [2.55, 11.87]). On the contrary, the caregivers who preferred that their loved one extend life even with LST were less than half as likely to engage in EoL communication (OR = 0.44, CI 95% = [0.20, 0.98]); and those who had a concern that doctors might want to stop treatments too soon were less than 30% as likely to talk about EoL care with the patient (OR = 0.29, CI 95% = [0.14, 0.60]).
Table 3. Logistic regression results (DV = EoL communication)
ADL = activities of daily living; CI 95% = 95% confidence interval; MD = medical doctor; SE = standard error.
*p < 0.05, **p < 0.01.
Model fit: –2 × log-likelihood = 207.40 χ2(14) = 46.29; p < 0.01. Nelgelkerke R 2 = 0.30.
DISCUSSION
The current study explored EoL communication and its contributing factors among the family caregivers of Latino patients who were home care recipients in a rural U.S.–Mexico border region. About half (50.8%) of the caregivers reported having an EoL discussion with their loved one (the patient). It is noteworthy that, of those 55 caregivers whose loved one was cognitively impaired (e.g., dementia, Alzheimer's disease, memory loss), only less than half (44%, n = 24) had an EoL discussion before the patient had the impairment, and only about 31% (n = 17) completed an AD. In the previous study by Black and colleagues (Reference Black, Fogarty and Phillips2009), more than half of the dementia patients had discussed EoL care preferences (56%) and completed an AD (59%). Our finding raises a concern about patients being incapacitated without an adequate ACP in place. Home care recipients often suffer from multiple chronic health conditions and confront ill health and impending mortality. In order to facilitate timely ACP, healthcare professionals need to assess patients' interest in ACP and encourage communication with their family and healthcare providers in advance.
The results of logistic regression indicated that family caregivers' knowledge about ADs, LST preference, and distrust toward physicians are significant predictors of EoL discussions with their loved one. Caregivers' knowledge about ADs increasing the likelihood of having an EoL communication is understandable, as ADs might have been used as a tool with which to dialogue about EoL care. In our study, more than half (58%) of the caregivers reported that they knew about ADs. Given that the patients for whom they cared also received regular home visits by nurses and that many of the patients were chronically ill, this finding calls for early education programs that target family caregivers and promote their involvement in EoL discussions. Furthermore, culturally appropriate practice can perhaps be optimized among healthcare professionals if their educational interventions are tailored to the unique individual and cultural characteristics of both patients and caregivers. Lacking health literacy widens health disparities and is a growing concern for Latino immigrants (Elder et al., Reference Elder, Ayala and Parra-Medina2009), particularly underserved rural Latinos. Hence, educational materials need to be language-concordant and take into consideration the low level of literacy. Both oral and written educational materials need to be available in Spanish and must be easy to comprehend (Sudore et al., Reference Sudore, Schickedanz and Landefeld2008). Volandes and colleagues (Reference Volandes, Ariza and Abbo2008) found that educational videos were an effective medium for Latinos with low levels of education and limited literacy. AD education needs to extend beyond its concepts and procedures by exploring beliefs and concerns about ACP. Distrust toward the healthcare system and misconceptions about ADs were found among racial/ethnic groups including Latinos (Perkins et al., Reference Perkins, Geppert and Gonzalez2002). Healthcare professionals' assessment about their beliefs and clarifying misinformation needs to be provided on a regular basis.
Another significant predictor of EoL communication was preference for LST. Family caregivers who preferred that their loved one prolong life even if he/she was to rely on LST were less likely to engage in EoL communication. Although our study did not assess patients' EoL care preferences, approximately 28% (n = 52) of the caregivers reported that they preferred LST, and only a third of them (n = 18, 34.6%) had discussed EoL care with their patients. There are mixed findings in the literature about EoL care preference. While Hines and colleagues (Reference Hines, Glover and Babrow2001) found that family caregivers as compared to patients are less likely to prefer prolonging the patient's life, other researchers (Tang et al., Reference Tang, Liu and Lai2005) presented contradictory findings. It has been found that a surrogate's decisions do not always adequately represent the patient's wishes. Oftentimes, decisions reflect the surrogate's values and preferences as well (Vig et al., Reference Vig, Taylor and Starks2006). In such cases, unclear or a lack of patient–family communication can complicate the decision-making process, leading to poor EoL care and conflicts among surviving relatives. Given that preference for family-centered decision making is more prevalent among Latinos (see Kelley et al., Reference Kelley, Wenger and Sarkisian2010), assessing the family's interests and values related to EoL care and facilitating dialogues with the patient can optimize the EoL decision-making process for patients and surrogates.
Lastly, family caregivers who were concerned that physicians might decide to forgo life-sustaining care (i.e., “pull the plug”) too soon were less likely to engage in EoL communication. We found that trust and confidence in physicians providing the best care to patients is one of the most important elements in EoL care (Heyland et al., Reference Heyland, Dodek and Rocker2006; Smith et al., Reference Smith, Sudore and Pérez-Stable2009). Hence, establishing a trustful relationship is a vital step in facilitating effective EoL communication among patients, family, and healthcare professionals. The efforts to address one's concerns about healthcare professionals' EoL decisions and care might begin with exploring the triad relationships among physicians, patient, and family, and assessing previous experiences of discrimination and each individual's expectations relevant to EoL care (Smith et al., Reference Smith, Sudore and Pérez-Stable2009)
LIMITATIONS OF THE STUDY
The current study has several limitations that need to mention. The data analyzed in this study were obtained from a convenience sample and the sample size was relatively small, which may limit the generalizability of our current findings. The patients were all Medicare beneficiaries recruited from a single site. The health conditions of the patients and EoL communication could have varied if this study included more and a wider range of patients and caregivers from multiple sites. Another limitation is relevant to the outcome variable: EoL communication was measured by participants' self-report only. Given that not all participants were primary caregivers, the patients' EoL discussions with other family members are unknown. It is recommended for future studies to evaluate EoL communication from both patients and family caregivers, which would offer a valuable insight on dyadic reciprocal understanding in EoL care.
There are also some limitations related to the measures used in our study. For example, the wording of the item on the distrust toward physicians question that refers to “pull the plug” may have biased respondents so that fewer respondents might endorse that item. In addition, it is important for future studies to explore the relationship between acculturation and EoL communication. One's attitude toward EoL care may vary according to culture, as his or her extent of acculturation experience is associated with EoL decision making (Kwak & Haley, Reference Kwak and Haley2005). Future research that includes acculturation measures, caregiver's place of residence (e.g., U.S., Mexico), and so forth will be beneficial.
CONCLUSIONS
Despite these limitations, this study contributes important findings that call for further research and attention during practice. Our finding on the lack of patient–caregiver EoL communication among the undeserved population highlights the importance of active involvement in ACP practice. For many Latinos, EoL decision making is more than an individual matter. Rather, it is an intricate and complex process that actively involves family members. This suggests that facilitating EoL communication and dialogue within a cultural framework is essential.
ACKNOWLEDGMENTS
We would like to thank the caregivers for their participation and the staff members at AccentCare Home Health for their support.
