- Email: [email protected]
Correlates and Predictors of Conflict at the End of Life Among Families Enrolled in Hospice
Vol. 50 No. 2 August 2015
Journal of Pain and Symptom Management 155
Original Article
Correlates and Predictors of Conflict at the End of Life Among Families Enrolled in Hospice Betty J. Kramer, PhD, MSSW, and Amy Z. Boelk, PhD, MSSW School of Social Work (B.J.K.), University of Wisconsin-Madison, Madison, Wisconsin; and Department of Sociology and Social Work (A.Z.B.), University of Wisconsin-Stevens Point, Stevens Point, Wisconsin, USA
Abstract Context. Despite the palliative care mandate to view family as the unit of care, and the high prevalence and detrimental consequences of conflict at the end of life, little research has been conducted with hospice families to understand what contributes to family conflict. Objectives. Using a recently generated explanatory matrix of family conflict at the end of life, this study sought to identify the correlates and predictors of family conflict. Methods. As part of a larger mixed methods cross-sectional study, a 100-item survey was administered to 161 hospice family caregivers enrolled in a Medicare/Medicaid certified non-profit hospice organization located in the Midwest U.S. Results. Although overall levels of conflict were relatively low, 57% of hospice caregivers reported experiencing some family conflict at the end of life. Contextual variables associated with family conflict included a history of family conflict, female gender, younger caregiver age, presence of children in the home, and less advance care planning discussions. Significant main effects in the prediction of family conflict in the final hierarchical multiple regression model included prior family conflict, caregiver age, caregiver gender, advance care planning discussions, family ‘‘coming out of the woodwork,’’ communication constraints, and family members asserting control. The model explained 59% of the variance in family conflict. Conclusion. Results support the multidimensional theoretical model of family conflict specifying the importance of the family context, key conditions that set the stage for conflict, and essential contributing factors. Implications for routine assessment and screening to identify families at risk and recommendations for future research are highlighted. J Pain Symptom Manage 2015;50:155e162. Ó 2015 American Academy of Hospice and Palliative Medicine. Published by Elsevier Inc. All rights reserved. Key Words Family conflict, hospice, caregiver, end of life, palliative care
Introduction Family caregivers provide the bulk of care for hospice patients while facing challenges of making difficult decisions, providing near-acute care, and managing family and work role conflicts.1 Other challenges reported by hospice caregivers include bearing witness to suffering, communication difficulties, uncertainty, fulfilling myriad obligations, enduring
Address correspondence to: Betty J. Kramer, PhD, MSSW, School of Social Work, Office 210, University of Wisconsin-Madison, 1350 University Avenue, Madison, WI 53706, USA. E-mail: [email protected] Ó 2015 American Academy of Hospice and Palliative Medicine. Published by Elsevier Inc. All rights reserved.
substantial loss, and facing the reality of death.2 Patient and caregiver well-being varies tremendously by the extent to which family members support and care for one another. In one of the first studies to interview inpatient hospice enrollees about family relationships, patients described the distress, loneliness, and abandonment associated with family conflict.3 Family conflict adversely effects caregiver burden, health,4 depression, anger,5 and bereavement.6,7 In
Accepted for publication: February 18, 2015.
0885-3924/$ - see front matter http://dx.doi.org/10.1016/j.jpainsymman.2015.02.026
156
Kramer and Boelk
addition, family conflict has a negative effect on the interdisciplinary team’s ability to address patient and family needs8 and is a substantial barrier to providing quality end-of-life care.9 Despite the palliative care mandate to view family as the unit of care10 and the myriad negative consequences that result from conflict, relatively little research has been conducted with hospice families to determine what contributes to family conflict at the end of life. The purpose of this study was to examine the correlates and predictors of conflict reported by hospice family caregivers. The conceptual framework guiding this study was drawn from an explanatory matrix of family conflict originally generated from a case study of an innovative managed care organization providing end-of-life care to low-income elders.11 This matrix was subsequently validated in a study of lung cancer caregivers12 and then replicated and expanded in a study involving the hospice context.13 As shown in Figure 1, this model proposes that conflict is best understood by considering the ‘‘family context,’’ ‘‘conditions’’ that shape family member actions or interactions, and ‘‘contributing factors.’’11,13 Contextually, families entering hospice vary in relationship quality, resources and demands, and how they have approached advance care planning. Prior conflict in the family is a strong predictor of conflict at the end of life,12 and spousal caregivers tend to be less aware of conflict than adult children.14,15 Higher rates of violent forms of conflict are reported by younger, less educated, poorer, and female respondents.16 Although advance care planning may reduce conflict in families facing difficult treatment decisions arising from uncertainty of patient wishes,17 conflict is reportedly higher among lung cancer caregivers of patients with previously specified care preferences.12
Vol. 50 No. 2 August 2015
Not surprisingly, siblings report that conflict is higher when other siblings assert control and disregard the previously stated wishes of the parent.18 The patient’s clinical care needs and the phenomenon of ‘‘family coming out of the woodwork’’ at end of life are two primary conditions that underlie the processes of family conflict. Absent family members ‘‘coming out of the woodwork’’ is a frequently cooccurring condition, which refers to a previously uninvolved and/or distant family member arriving on the scene as the patient’s condition deteriorates.13 For example, it is not uncommon for adult children who were not actively involved in caring for a parent, who live in another state, to return home when they learn that mom or dad has been admitted to hospice. According to the explanatory matrix, family members of patients with more distressing symptoms and who have family members ‘‘come out of the woodwork’’ at the end of life will likely experience greater ‘‘difficulty integrating death awareness,’’ communication constraints and attempts at asserting control in the decision-making process, and ultimately higher family conflict.11,13 Difficulty integrating death awareness ‘‘involves situations in which the patient and/or family has difficulty accepting the patient’s condition and/or being around death and dying in general.’’ 13, p. 9 Emotionally accepting terminality is one of the most difficult tasks facing caregivers19 and trouble doing so may prompt family members to behave in ways that would stimulate conflict; for example, pressing for aggressive treatments irrespective of others’ wishes.11 Communication constraints and asserting control were highly correlated with conflict at the end of life among lung cancer caregivers.12 We hypothesize that: 1) the family context will be associated with family conflict. More specifically, we expect that higher levels of conflict will be found
Family Context Historical Relationship Patterns History of family conflict Relationship to patient
Caregiver Characteristics, Demands & Resources Age Gender Education Income Presence of children in home Duration of caregiving
Conditions
Contributing Factors
Patient Care Needs
Difficulty Integrating Death Awareness
ADLs IADLs
Communication Constraints
“Coming out of the “Woodwork”
Asserting Control
Advance Care Planning Advance planning discussions Advance directives
Fig. 1. Explanatory matrix of family conflict at the end of life.
Conflict Family Conflict
Vol. 50 No. 2 August 2015
Correlates and Predictors of Family Conflict
among family members with more demands (i.e., longer duration of caregiving and presence of children in the home), fewer resources (i.e., lower income, lower education, and younger age), and less advance care planning (i.e., absence of advance directives or conversations before illness onset); 2) certain conditions (i.e., patient care needs and family ‘‘coming out of woodwork’’) will be associated with contributing factors and family conflict; and 3) after controlling for family context and conditions, contributing factors (i.e., death anxiety, communication constraints, and family asserting control) will significantly predict family conflict.
Methods Study Design and Sampling As part of a larger mixed methods study,13 quantitative data were collected via a 100-item survey administered to 161 hospice family caregivers enrolled in a Medicare/Medicaid-certified non-profit hospice organization located in the Midwest U.S. Patients aged 65 years and older comprised approximately 90% of the patients served, with approximately 44% being male and 56% being female. Diagnoses of patients were distributed as follows: cancer (47%), cardiovascular disease (14%), pulmonary disease (10%), renal failure (5%), Alzheimer’s disease (6%), amyotrophic lateral sclerosis (2%), and other (17%). Although the hospice organization serves a somewhat older demographic and has fewer very short lengths of stay, it mirrors the characteristics of hospice care throughout the state and America.20 During the hospice admission process, the family caregiver most actively involved in patient care and decision making was identified by the patient and/or family. All family caregivers of patients admitted during the data collection period were eligible to participate. The initial plan was to recruit with every admission, unless a determination was made, based on the clinical judgment of the social worker, that the patient was actively dying or that introducing the study would result in undue caregiver distress. Given concern about the very low response rate of 22% during the initial two-month study period, the following recruitment barriers were uncovered that needed to be addressed: the admitting social workers were not consistently offering the survey, neglecting to followup when family member was not present at the time of admission, were screening potential candidates out for reasons other than noted in the protocol, and family caregivers would neglect to review consent forms left with them in stacks of other information provided at admission. Recruitment was temporarily put on hold, and the project researchers met with social work staff to review and discuss barriers to
157
protocol integrity and strategies for improving the response rate. A new protocol was then devised, which included social workers directly reviewing and obtaining caregiver consent, project researchers making follow-up calls to caregivers, and enhanced appearance of survey packets to help caregivers better distinguish them from other hospice admission materials. Once initial recruitment barriers were identified during the initial two-month study period, and protocols were modified, 58% of the survey packets administered by social work staff were completed by caregivers. Per institutional review board requirements, patient consent was sought, and if secured, the social worker completed the consent process with the family caregiver, who was given a packet containing an introductory letter, consent form, survey, and return envelope. Admitting social workers then mailed patient and caregiver consent forms, and caregivers completed and mailed surveys to the project researchers.
Measures Outcome-Family Conflict. Given the absence of available instruments to specifically measure family conflict at the end of life, an eight-item scale was developed for use in this study. The first author generated items for this measure from findings from a qualitative study of family conflict,11 a review of the literature, prior clinical experience, and consultation with clinicians with expertise working with families in palliative care. The construct validity of this scale was supported by significant correlations with a standardized measure of family functioning, and internal consistency was supported by a Cronbach’s alpha of 0.89. Respondents were asked to use a 5-point scale (1 ¼ not at all to 5 ¼ very much) to answer eight questions anchored in the end-of-life experience: ‘‘As you think about your family since your relative was diagnosed with a life-threatening illness: how much do any family members insult or yell at one another; or disagree or argue about health care decisions for your relative; your relative’s illness or physical condition; the way a member was treating your relative; certain family members not pulling their weight; what is meant by ‘‘a good death;’’ how money is being spent; or where your relative should live out his/her remaining days?’’ Higher scores on this measure indicated higher levels of family conflict at the end of life. Contextual Factors. We assessed history of family conflict before illness with a two-item scale developed by the first author: ‘‘Prior to your family member’s illness, how much did any family members have serious arguments with one another? Insult or yell at one another?’’ (0 ¼ not at all to 5 ¼ very much)
158
Kramer and Boelk
(alpha ¼ 0.83). Caregiver characteristics included sociodemographic variables such as age (continuous variable in years); gender (0 ¼ male; 1 ¼ female); education (0 ¼ less than high school; 1 ¼ high school grad, 3 ¼ post-high school); yearly income ($19,999 or less; $20,000e$49,999; $50,000e$99,999; $100,000þ); presence of children in the home (0 ¼ no; 1 ¼ yes); and duration of caregiving (continuous variable in months). Advance care planning was ascertained by asking if the patient had completed a durable power of attorney for health care and/or a living will before their illness (0 ¼ no; 1 ¼ yes), and the extent to which the family had discussed the patient’s care wishes before the illness using a 5-point scale (1 ¼ strongly disagree; 5 ¼ strongly agree). Conditions. Patient care needs were measured with the Activities of Daily Living (ADLs) scale, a six-item scale to determine patient needs relevant to bathing, dressing, toileting, transferring, incontinence, and feeding21 (alpha ¼ 0.89) and the Instrumental Activities of Daily Living (IADLs), a seven-item scale to determine patient needs relevant to shopping, food preparation, housekeeping, laundry, transportation, medication, and finances22 (alpha ¼ 0.93). A single question assessed the ‘‘coming out of the woodwork’’ phenomena. It simply stated: ‘‘Some families say that when a family member becomes seriously ill, other family members suddenly ‘come out of the woodwork.’ This means that family members who were not previously in regular contact suddenly become more involved. Has this occurred as a result of your family member’s illness?’’ (0 ¼ no; 1 ¼ yes). Contributing Factors. Difficulty integrating death awareness was assessed by a single item that asked respondents to use a 5-point scale (1 ¼ strongly disagree; 5 ¼ strongly agree) to indicate the degree to which ‘‘thinking about my family member’s death is very difficult for me.’’ Four items embedded in the Family Assessment Device23 were utilized to measure communication constraints. Respondents were asked to use a 5-point scale (1 ¼ strongly agree to 5 ¼ strongly disagree) to answer the following items: ‘‘We cannot talk to each other about the sadness we feel,’’ ‘‘We avoid discussing our fears and concerns,’’ ‘‘We can express feelings to each other,’’ and ‘‘We confide in each other.’’ Two of the items were reverse coded for directionality, and higher scores on this measure indicated higher levels of communication constraints (alpha ¼ 0.82). Family asserting control was measured through a single item, which asked respondents to use a 5-point scale (1 ¼ not at all to 5 ¼ very much) to indicate the degree to which any family member had asserted control in decision making
Vol. 50 No. 2 August 2015
regarding patient care. Higher scores on this measure indicated higher levels of asserting control.
Statistical Analysis All variables were examined for missing values and fit between their distributions and the analytic assumptions. The outcome variable was positively skewed and a logarithm transformation was necessary to improve its distribution.24 Analysis of survey responses involved standard univariate analysis of major variables (i.e., descriptive statistics) followed by a series of bivariate and mutlivariate analyses. Zero-order correlations, t-tests, and analyses of variance were conducted to check for presence and size of statistical relationship between each of the context and condition variables, the contributing factors and the outcome of interest, as per the hypothesized relationships. Regression analysis was conducted, controlling for context and condition variables to examine the effect of contributing factors (i.e., death anxiety, communication constraints, family asserting control) on family conflict.
Results Sample Description Sample characteristics of caregivers and patients are described in Table 1. Most caregivers were adult children or spouses, female, white, and had a high school or college education. The mean ages were 63 and 80 years for caregivers and patients, respectively. Approximately 14% of caregivers had children under the age of 18 years living at home. Nearly half of the patients were being cared for in the home setting and slightly over half were female. The majority of patients had completed a power of attorney for health care (96%) and/or a living will (81%) before the illness, but only 63% had engaged in advance care planning conversations with the family. Typical for hospice programs nationwide, the largest percentage of patients had a primary diagnosis of cancer (44%), followed by heart disease (17%), and dementia (11%).
Presence of Family Conflict and Family Context: Hypothesis 1 Fifty-seven percent of caregivers reported experiencing one or more types of family conflict at the end of life. Overall, levels of family conflict at the end of life were relatively low in this sample as evidenced by the mean score of 1.39 (SD ¼ 0.63; range ¼ 1e4.5). Additionally, 48% of caregivers reported some family conflict before their family member’s illness. Of these, 41% reported a history of serious arguments and 32% reported a history of insulting and/or yelling within the family.
Vol. 50 No. 2 August 2015
Correlates and Predictors of Family Conflict
Table 1 Sample Characteristics of Family Caregivers and Patients Caregiver Relationship to patient (%) Spouse/significant other/life partner Daughter/son/daughter or son-in-law Sibling/sister or brother-in-law Parent Other family and family like friends Age in yrs Mean (SD) Gender (% female) Race (% white) Education (%) Less than high school High school graduate Post-high school education Yearly income (%) $19,999 or less 20,000e49,999 50,000e99,999 100,000þ Missing Presence of children in caregiver’s home (%) Duration of caregiving in months Mean (SD) Patient Age in years Mean (SD) Gender (% female) Lives at home or with other family member (%) Patient primary diagnosis (%) Cancer Heart disease Lung disease Kidney disease Alzheimer’s or other dementia ALS Failure to thrive Other illness Engaged in advance care planning before illness (%) Completed a power of attorney for health care Completed living will Advance planning discussions with family
27.3 55.3 4.3 3.1 9.9 59 (13) 84 98.8 5.6 48.4 46.0 16.7 39.8 23.6 9.9 9.9 13.7 72 (107) 80 (14) 56.5 47.8 43.8 16.9 4.4 3.8 10.6 0.6 7.5 12.5 96.3 80.8 62.9
ALS ¼ amyotrophic lateral sclerosis.
We hypothesized that conflict at the end of life will be higher among caregivers with a history of family conflict, more family demands (i.e., longer duration of caregiving, presence of children in the home), fewer resources (i.e., lower income, lower education, younger age), and less advance care planning (i.e. absence of advance directives or conversation before illness onset). Bivariate relationships between the contextual variables and the outcome variable revealed that family conflict at the end of life was significantly associated with a history of family conflict (r ¼ 0.51, P # 0.001), female gender (t ¼ 2.11, P # 0.05), younger caregiver age (r ¼ 0.23, P # 0.01), presence of children in the home (t ¼ 2.13, P # 0.05), and less family discussions about patient wishes before illness (r ¼ 0.24, P # 0.01). No relationship was found between conflict and education, income, duration of caregiving, relationship to the patient and presence of advance directives. This hypothesis is partially supported.
159
The Role Played by Conditions: Hypothesis 2 We hypothesized that certain conditions (i.e., patient care needs and ‘‘family coming out of woodwork’’) will be associated with contributing factors (i.e., difficulty integrating death awareness, communication constraints, asserting control) and family conflict at the end of life. Bivariate correlations revealed that more ADL impairments were associated with less difficulty integrating death awareness (r ¼ 0.16, P # 0.05). Greater IADL impairments were associated with family members asserting control (r ¼ 0.15, P # 0.05) and higher levels of family conflict (r ¼ 0.17, P # 0.05). ‘‘Coming out of the woodwork’’ correlated with greater difficulty integrating death awareness (r ¼ 0.15, P # 0.05), and higher levels of family conflict at trend level (r ¼ 0.13, P # 0.10). This hypothesis is partially supported.
The Influence of Contributing Factors: Hypothesis 3 We hypothesized that after controlling for family context and conditions that contributing factors will significantly predict family conflict. To test this third hypothesis, hierarchical multiple regression analysis was conducted that controlled for the family context and conditions of the patient and family. Because of the number of independent variables and the limited sample size, only the contextual variables that were significantly associated with family conflict were entered into the regression equation. The first block of independent variables included five contextual variables (i.e., history of conflict, caregiver age and gender, presence of children in the home, and advance care planning discussions). The second block included the two clinical care need variables (i.e., ADLs and IADLs) and ‘‘coming out of the woodwork.’’ The third block included the three contributing factors (i.e., difficulty integrating death awareness, communication constraints, and asserting control). Table 2 presents the results of the regression analysis, which explained 59% of the total variance in family conflict. In the first step of the model, family contextual variables explained 30% of the variance in conflict. Prior family conflict held the strongest relationship to end-of-life conflict, and two contextual variables (i.e., younger caregiver age and less advance planning discussions) revealed trend level associations. In the second step, conditional variables did not make a significant contribution to the prediction of family conflict as a set, explaining only 2% of the total variance. In the third step, contributing factors uniquely explained 27% of the variance in conflict, the most important two being communication constraints and asserting control. Significant main effects in the prediction of family conflict in the final model included prior family conflict (b ¼ 0.22, P # 0.001),
160
Kramer and Boelk
Table 2 Hierarchical Regression Analysis for Family Conflict Betas Variables Family context Prior family conflict Caregiver age Caregiver gender Presence of children Advance planning discussions Conditions Patient ADLs Patient IADLs ‘‘Coming out of the woodwork’’ Contributing factors Difficulty integrating death awareness Communication constraints Asserting control Total R2 2 R change F for R2 change
Step 1
Step 2
Step 3
0.41a 0.15b 0.10 0.02 0.13b
0.40a 0.11 0.13b 0.01 0.15d
0.23a 0.16c 0.10b 0.04 0.11d
0.01 0.14b 0.09
0.01 0.05 0.11d 0.05
0.30 0.30 12.41a
0.32 0.02 1.58
0.22a 0.46a 0.59 0.27 31.94a
ADL ¼ Activities of Daily Living; IADL ¼ Instrumental Activities of Daily Living. a P # 0.001. b P # 0.10 (two-tailed). c P # 0.01. d P # 0.05.
caregiver age (b ¼ 0.17, P # 0.01), caregiver gender (b ¼ 0.10, P # 0.10), advance care planning discussions (b ¼ 0.12, P # 0.05), family ‘‘coming out of the woodwork’’ (b ¼ 0.10, P # 0.10), communication constraints (b ¼ 0.22, P # 0.001), and family members asserting control (b ¼ 0.46, P # 0.001). Hypothesis 3 is supported.
Discussion Over the past decade, there has been substantial progress in identifying typologies of family functioning25e27 and their relationship to psychosocial morbidity in families receiving palliative care.25 Although conflict is an important dimension of family functioning,27 little is known about the precipitants of conflict itself. To our knowledge, this is the first study to investigate the correlates and predictors of conflict reported by family caregivers concurrently enrolled in hospice. Results support the multidimensional theoretical model of family conflict specifying the importance of the family context, key conditions that set the stage for conflict, and essential contributing factors. Three contextual variables associated with conflict that also were reported in recent studies include prior family conflict,12 younger age,12,26 and female gender.26 Although there was no relationship between conflict and advance directives, conflict was associated with advance planning discussions. Recent evidence
Vol. 50 No. 2 August 2015
suggests that more highly functioning and emotionally supportive families are more likely to have such discussions.28 Given the cross-sectional limitation of this study, it is not possible to disentangle these effects, although the relationship persisted after controlling for prior conflict. Although those with children in the home had higher levels of conflict, this relationship disappeared when other contextual variables were taken into consideration, possibly because younger women are more likely to have children at home. Although it is not surprising that families with a history of conflict are likely to experience conflict at the end of life, it is noteworthy that a subset of caregivers experiencing conflict (9%) reported no history of prior conflict within the family. After controlling for family context, caregivers who reported that family members were ‘‘coming out of the woodwork,’’ asserting control and experiencing communication constraints, were likely to experience higher levels of family conflict at the end of life. These contributing factors were predictive of conflict in a lung cancer caregiver study.12 These findings provide additional insights regarding families who may be at greater risk of conflict and the potential interventions that may be employed to help. It is possible that additional factors, not identified in our explanatory matrix, might also contribute to family conflict and would be worthy of further investigation. For example, perceptions of support and cohesiveness that are associated with conflict6 and other key processes associated with family resilience12 may influence whether conflict will arise and/or how it is communicated or dealt with. There are several limitations of our study. Given the racially homogeneous sample (99% white), and low response rate (58% once barriers to recruitment were addressed), results need be interpreted with caution. We cannot generalize these findings in the absence of a representative and more diverse sample. Although more than half of the caregivers interviewed reported some conflict, the degree of conflict was relatively low, which is consistent with several other studies reporting low rates of high conflict overall.15,25e26 It is quite possible that respondents were less burdened and may have had fewer demands, more resources, and overall better family relationship quality than non-respondents. The extent to which the correlates and predictors of conflict would be different among a more diverse sample is an empirical question worthy of future investigation. As a cross-sectional study, we cannot adequately evaluate the temporal ordering of variables. Some of our measures were limited. For example, the income variable categories were not equal in size and a large number of caregivers chose not to answer this question (10% were missing)
Vol. 50 No. 2 August 2015
Correlates and Predictors of Family Conflict
161
making it difficult to detect differences, and two of the contributing factors were measured with single-item indicators developed for this study. In addition, the single-item ‘‘asserting control’’ indicator reflects the caregiver’s perception of the extent to which any family member had asserted control in decision making, without validation by other family members. Further measurement development and testing of these constructs is needed. In addition, future studies eliciting multiple family members’ perceptions regarding various dimensions of the explanatory matrix would further enrich understanding of the family conflict phenomenon.
References
Conclusion
6. Kissane DW, Bloch S, Onghena P, McKenzie DP, Snyder RD, Dowe DL. The Melbourne Family Grief Study, II: psychosocial morbidity and grief in bereaved families. Am J Psychiatry 1996;153:659e666.
Our results have important implications for palliative care and future research. First, the results suggest the importance of implementing routine assessment protocols and monitoring of historical and current family relationship quality, demands, resources, advance care planning discussions, conditions, and contributing factors. As with the benefits demonstrated by clinicians who recommend routine assessment of family functioning,29 such protocols and monitoring would allow service providers to identify and target appropriate interventions to families at most risk. Second, our findings suggest the importance of interventions for promoting and facilitating open communication, shared decision making, and advance care planning discussions as a potential means to prevent conflict at the end of life. Finally, while this study provides a preliminary understanding of the correlates and predictors of conflict at the end of life, there is much that remains to be understood. Prospective studies representing more diverse family forms are needed. Future research is needed to examine conflict that arises in families that have not previously experienced it and the extent to which conflict might vary by socioeconomic status and more diverse family forms. Translational questions that may be investigated include: What, if anything, might be done to prevent or to address the ‘‘coming out of the woodwork’’ phenomenon? What interventions might alleviate communication constraints and promote shared decision making rather than assertions of control?
Disclosures and Acknowledgments This work was supported in part by the University of Wisconsin-Stevens Point Personnel Development Committee. No competing financial interests exist. The authors express their gratitude to the family caregivers who participated in this study.
1. Waldrop DP, Kramer BJ, Skrentny JA, Milch RA, Finn W. Final transitions: family caregiving at the end of life. J Palliat Med 2005;8:623e638. 2. Kehl K, Kirchoff K, Kramer BJ, Hovland-Scafe C. Challenges facing families at the end-of-life in three settings. J Soc Work End Life Palliat Care 2010;5:144e168. 3. Broom K, Kirby E. The end of life and the family: hospice patients’ views on dying as relational. Sociol Health Illn 2013;35:499e513. 4. Strawbridge WJ, Wallhagen MI. Impact of family conflict on adult child caregivers. Gerontologist 1991;31:770e777. 5. Semple SJ. Conflict in Alzheimer’s caregiving families: its dimensions and consequences. Gerontologist 1992;32: 648e655.
7. Kramer BJ, Kavanaugh M, Trentham-Dietz A, Walsh M, Yonker JA. Complicated grief in caregivers of persons with lung cancer: the role of family conflict, intrapsychic strains and hospice utilization. Omega 2010e2011;62:201e220. 8. Kramer BJ, Yonker JA. Perceived success in addressing end-of-life care needs of low-income elders and their families: what’s family conflict got to do with it? J Pain Symptom Manage 2011;41:35e48. 9. Kayashima R. Barriers to good end-of-life care: a physician survey. Hawaii Med J 2001;60:40e44. 10. National Consensus Project for Quality Palliative Care. Clinical practice guidelines for quality palliative Care. 2nd ed. 2009. Available at: http://www.nationalconsensusproject. org. Accessed June 6, 2011. 11. Kramer BJ, Boelk A, Auer C. Family conflict at the endof-life: lessons learned in a model program for vulnerable older adults. J Palliat Med 2006;9:791e801. 12. Kramer BJ, Kavanaugh M, Trentham-Dietz A, Walsh M, Yonker JA. Predictors of family conflict at the end of life: the experience of spouses and adult children of persons with lung cancer. Gerontologist 2010;50:215e225. 13. Boelk AZ, Kramer BJ. Advancing theory of family conflict at the end of life: a hospice case study. J Pain Symptom Manage 2012;44:655e670. 14. Clark PC, Shields CG, Ayccock D, Wolf SL. Preliminary reliability and validity of a family caregiver conflict scale for stroke. Prog Cardiovasc Nurs 2003;18:77e82. 92. 15. Kissane DW, Bloch S, Burns WI, McKenzie DP, Posterino M. Psychological morbidity in the families of patients with cancer. Psychooncology 1994;3:47e56. 16. Kessler RC, Molnar BE, Feurer ID, Appelbaum M. Patterns and mental health predictors of domestic violence in the United States: results from the National Comorbidity Survey. Int J Law Psychiatry 2001;24:487e508. 17. Tilden VP, Tolle SW, Garland MJ, Nelson CA. Decisions about life-sustaining treatments. Arch Intern Med 1995;155: 633e638. 18. Su CT, McMahan BS, Williams BA, Sharma RK, Sudore RL. Family matters: effects of birth order, culture,
162
Kramer and Boelk
and family dynamics on surrogate decision-making. J Am Geriatr Soc 2014;62:175e182. 19. Chan CW, Chang AM. Managing caregiver tasks among family caregivers of cancer patients in Hong Kong. J Adv Nurs 1999;29:484e489. 20. NHPCO Facts and Figures: Hospice Care in America. Alexandria, VA: National Hospice and Palliative Care Organization, 2012. 21. Katz S, Down TD, Cash HR, Grotz RC. Progress in the development of the index of ADL. Gerontologist 1970;10:20e30. 22. Lawton MP, Brody EM. Assessment of older people: selfmaintaining and instrumental activities of daily living. Geron 1969;9:179e186. 23. Miller IW, Epstein NB, Bishop DS, Keitner GI. The McMaster Family Assessment Device: reliability and validity. J Marital Fam Ther 1985;11:345e356. 24. Tabachnick BG, Fidell LS. Using multivariate statistics, 4th ed. Boston, MA: Allyn & Bacon, 2001.
Vol. 50 No. 2 August 2015
25. Kissane DW, McKenzie M, McKenzie DP, O’Neill AF, Bloch S. Psychosocial morbidity associated with patterns of family functioning in palliative care: baseline data from the family focused grief therapy controlled trial. Palliat Med 2003;17:527e537. 26. Schuler TA, Zaider TI, Li Y, Hichenberg S, Masterson M, Kissane DW. Typology of perceived family functioning in an American sample of patients with advanced cancer. J Pain Symptom Manage 2014;48:281e288. 27. Kissane D, Bloch S. Family focused grief therapy. Berkshire, UK: Open University Press, 2010. 28. Boerner K, Carr D, Moorman S. Family relationships and advance care planning: do supportive and critical relations encourage or hinder planning. J Gerontol B Psychol Sci Soc Sci 2013;68:246e256. 29. King DA, Quill T. Working with families in palliative care: one size does not fit all. J Palliat Med 2006;9: 704e715.
Journal of Pain and Symptom Management 155
Original Article
Correlates and Predictors of Conflict at the End of Life Among Families Enrolled in Hospice Betty J. Kramer, PhD, MSSW, and Amy Z. Boelk, PhD, MSSW School of Social Work (B.J.K.), University of Wisconsin-Madison, Madison, Wisconsin; and Department of Sociology and Social Work (A.Z.B.), University of Wisconsin-Stevens Point, Stevens Point, Wisconsin, USA
Abstract Context. Despite the palliative care mandate to view family as the unit of care, and the high prevalence and detrimental consequences of conflict at the end of life, little research has been conducted with hospice families to understand what contributes to family conflict. Objectives. Using a recently generated explanatory matrix of family conflict at the end of life, this study sought to identify the correlates and predictors of family conflict. Methods. As part of a larger mixed methods cross-sectional study, a 100-item survey was administered to 161 hospice family caregivers enrolled in a Medicare/Medicaid certified non-profit hospice organization located in the Midwest U.S. Results. Although overall levels of conflict were relatively low, 57% of hospice caregivers reported experiencing some family conflict at the end of life. Contextual variables associated with family conflict included a history of family conflict, female gender, younger caregiver age, presence of children in the home, and less advance care planning discussions. Significant main effects in the prediction of family conflict in the final hierarchical multiple regression model included prior family conflict, caregiver age, caregiver gender, advance care planning discussions, family ‘‘coming out of the woodwork,’’ communication constraints, and family members asserting control. The model explained 59% of the variance in family conflict. Conclusion. Results support the multidimensional theoretical model of family conflict specifying the importance of the family context, key conditions that set the stage for conflict, and essential contributing factors. Implications for routine assessment and screening to identify families at risk and recommendations for future research are highlighted. J Pain Symptom Manage 2015;50:155e162. Ó 2015 American Academy of Hospice and Palliative Medicine. Published by Elsevier Inc. All rights reserved. Key Words Family conflict, hospice, caregiver, end of life, palliative care
Introduction Family caregivers provide the bulk of care for hospice patients while facing challenges of making difficult decisions, providing near-acute care, and managing family and work role conflicts.1 Other challenges reported by hospice caregivers include bearing witness to suffering, communication difficulties, uncertainty, fulfilling myriad obligations, enduring
Address correspondence to: Betty J. Kramer, PhD, MSSW, School of Social Work, Office 210, University of Wisconsin-Madison, 1350 University Avenue, Madison, WI 53706, USA. E-mail: [email protected] Ó 2015 American Academy of Hospice and Palliative Medicine. Published by Elsevier Inc. All rights reserved.
substantial loss, and facing the reality of death.2 Patient and caregiver well-being varies tremendously by the extent to which family members support and care for one another. In one of the first studies to interview inpatient hospice enrollees about family relationships, patients described the distress, loneliness, and abandonment associated with family conflict.3 Family conflict adversely effects caregiver burden, health,4 depression, anger,5 and bereavement.6,7 In
Accepted for publication: February 18, 2015.
0885-3924/$ - see front matter http://dx.doi.org/10.1016/j.jpainsymman.2015.02.026
156
Kramer and Boelk
addition, family conflict has a negative effect on the interdisciplinary team’s ability to address patient and family needs8 and is a substantial barrier to providing quality end-of-life care.9 Despite the palliative care mandate to view family as the unit of care10 and the myriad negative consequences that result from conflict, relatively little research has been conducted with hospice families to determine what contributes to family conflict at the end of life. The purpose of this study was to examine the correlates and predictors of conflict reported by hospice family caregivers. The conceptual framework guiding this study was drawn from an explanatory matrix of family conflict originally generated from a case study of an innovative managed care organization providing end-of-life care to low-income elders.11 This matrix was subsequently validated in a study of lung cancer caregivers12 and then replicated and expanded in a study involving the hospice context.13 As shown in Figure 1, this model proposes that conflict is best understood by considering the ‘‘family context,’’ ‘‘conditions’’ that shape family member actions or interactions, and ‘‘contributing factors.’’11,13 Contextually, families entering hospice vary in relationship quality, resources and demands, and how they have approached advance care planning. Prior conflict in the family is a strong predictor of conflict at the end of life,12 and spousal caregivers tend to be less aware of conflict than adult children.14,15 Higher rates of violent forms of conflict are reported by younger, less educated, poorer, and female respondents.16 Although advance care planning may reduce conflict in families facing difficult treatment decisions arising from uncertainty of patient wishes,17 conflict is reportedly higher among lung cancer caregivers of patients with previously specified care preferences.12
Vol. 50 No. 2 August 2015
Not surprisingly, siblings report that conflict is higher when other siblings assert control and disregard the previously stated wishes of the parent.18 The patient’s clinical care needs and the phenomenon of ‘‘family coming out of the woodwork’’ at end of life are two primary conditions that underlie the processes of family conflict. Absent family members ‘‘coming out of the woodwork’’ is a frequently cooccurring condition, which refers to a previously uninvolved and/or distant family member arriving on the scene as the patient’s condition deteriorates.13 For example, it is not uncommon for adult children who were not actively involved in caring for a parent, who live in another state, to return home when they learn that mom or dad has been admitted to hospice. According to the explanatory matrix, family members of patients with more distressing symptoms and who have family members ‘‘come out of the woodwork’’ at the end of life will likely experience greater ‘‘difficulty integrating death awareness,’’ communication constraints and attempts at asserting control in the decision-making process, and ultimately higher family conflict.11,13 Difficulty integrating death awareness ‘‘involves situations in which the patient and/or family has difficulty accepting the patient’s condition and/or being around death and dying in general.’’ 13, p. 9 Emotionally accepting terminality is one of the most difficult tasks facing caregivers19 and trouble doing so may prompt family members to behave in ways that would stimulate conflict; for example, pressing for aggressive treatments irrespective of others’ wishes.11 Communication constraints and asserting control were highly correlated with conflict at the end of life among lung cancer caregivers.12 We hypothesize that: 1) the family context will be associated with family conflict. More specifically, we expect that higher levels of conflict will be found
Family Context Historical Relationship Patterns History of family conflict Relationship to patient
Caregiver Characteristics, Demands & Resources Age Gender Education Income Presence of children in home Duration of caregiving
Conditions
Contributing Factors
Patient Care Needs
Difficulty Integrating Death Awareness
ADLs IADLs
Communication Constraints
“Coming out of the “Woodwork”
Asserting Control
Advance Care Planning Advance planning discussions Advance directives
Fig. 1. Explanatory matrix of family conflict at the end of life.
Conflict Family Conflict
Vol. 50 No. 2 August 2015
Correlates and Predictors of Family Conflict
among family members with more demands (i.e., longer duration of caregiving and presence of children in the home), fewer resources (i.e., lower income, lower education, and younger age), and less advance care planning (i.e., absence of advance directives or conversations before illness onset); 2) certain conditions (i.e., patient care needs and family ‘‘coming out of woodwork’’) will be associated with contributing factors and family conflict; and 3) after controlling for family context and conditions, contributing factors (i.e., death anxiety, communication constraints, and family asserting control) will significantly predict family conflict.
Methods Study Design and Sampling As part of a larger mixed methods study,13 quantitative data were collected via a 100-item survey administered to 161 hospice family caregivers enrolled in a Medicare/Medicaid-certified non-profit hospice organization located in the Midwest U.S. Patients aged 65 years and older comprised approximately 90% of the patients served, with approximately 44% being male and 56% being female. Diagnoses of patients were distributed as follows: cancer (47%), cardiovascular disease (14%), pulmonary disease (10%), renal failure (5%), Alzheimer’s disease (6%), amyotrophic lateral sclerosis (2%), and other (17%). Although the hospice organization serves a somewhat older demographic and has fewer very short lengths of stay, it mirrors the characteristics of hospice care throughout the state and America.20 During the hospice admission process, the family caregiver most actively involved in patient care and decision making was identified by the patient and/or family. All family caregivers of patients admitted during the data collection period were eligible to participate. The initial plan was to recruit with every admission, unless a determination was made, based on the clinical judgment of the social worker, that the patient was actively dying or that introducing the study would result in undue caregiver distress. Given concern about the very low response rate of 22% during the initial two-month study period, the following recruitment barriers were uncovered that needed to be addressed: the admitting social workers were not consistently offering the survey, neglecting to followup when family member was not present at the time of admission, were screening potential candidates out for reasons other than noted in the protocol, and family caregivers would neglect to review consent forms left with them in stacks of other information provided at admission. Recruitment was temporarily put on hold, and the project researchers met with social work staff to review and discuss barriers to
157
protocol integrity and strategies for improving the response rate. A new protocol was then devised, which included social workers directly reviewing and obtaining caregiver consent, project researchers making follow-up calls to caregivers, and enhanced appearance of survey packets to help caregivers better distinguish them from other hospice admission materials. Once initial recruitment barriers were identified during the initial two-month study period, and protocols were modified, 58% of the survey packets administered by social work staff were completed by caregivers. Per institutional review board requirements, patient consent was sought, and if secured, the social worker completed the consent process with the family caregiver, who was given a packet containing an introductory letter, consent form, survey, and return envelope. Admitting social workers then mailed patient and caregiver consent forms, and caregivers completed and mailed surveys to the project researchers.
Measures Outcome-Family Conflict. Given the absence of available instruments to specifically measure family conflict at the end of life, an eight-item scale was developed for use in this study. The first author generated items for this measure from findings from a qualitative study of family conflict,11 a review of the literature, prior clinical experience, and consultation with clinicians with expertise working with families in palliative care. The construct validity of this scale was supported by significant correlations with a standardized measure of family functioning, and internal consistency was supported by a Cronbach’s alpha of 0.89. Respondents were asked to use a 5-point scale (1 ¼ not at all to 5 ¼ very much) to answer eight questions anchored in the end-of-life experience: ‘‘As you think about your family since your relative was diagnosed with a life-threatening illness: how much do any family members insult or yell at one another; or disagree or argue about health care decisions for your relative; your relative’s illness or physical condition; the way a member was treating your relative; certain family members not pulling their weight; what is meant by ‘‘a good death;’’ how money is being spent; or where your relative should live out his/her remaining days?’’ Higher scores on this measure indicated higher levels of family conflict at the end of life. Contextual Factors. We assessed history of family conflict before illness with a two-item scale developed by the first author: ‘‘Prior to your family member’s illness, how much did any family members have serious arguments with one another? Insult or yell at one another?’’ (0 ¼ not at all to 5 ¼ very much)
158
Kramer and Boelk
(alpha ¼ 0.83). Caregiver characteristics included sociodemographic variables such as age (continuous variable in years); gender (0 ¼ male; 1 ¼ female); education (0 ¼ less than high school; 1 ¼ high school grad, 3 ¼ post-high school); yearly income ($19,999 or less; $20,000e$49,999; $50,000e$99,999; $100,000þ); presence of children in the home (0 ¼ no; 1 ¼ yes); and duration of caregiving (continuous variable in months). Advance care planning was ascertained by asking if the patient had completed a durable power of attorney for health care and/or a living will before their illness (0 ¼ no; 1 ¼ yes), and the extent to which the family had discussed the patient’s care wishes before the illness using a 5-point scale (1 ¼ strongly disagree; 5 ¼ strongly agree). Conditions. Patient care needs were measured with the Activities of Daily Living (ADLs) scale, a six-item scale to determine patient needs relevant to bathing, dressing, toileting, transferring, incontinence, and feeding21 (alpha ¼ 0.89) and the Instrumental Activities of Daily Living (IADLs), a seven-item scale to determine patient needs relevant to shopping, food preparation, housekeeping, laundry, transportation, medication, and finances22 (alpha ¼ 0.93). A single question assessed the ‘‘coming out of the woodwork’’ phenomena. It simply stated: ‘‘Some families say that when a family member becomes seriously ill, other family members suddenly ‘come out of the woodwork.’ This means that family members who were not previously in regular contact suddenly become more involved. Has this occurred as a result of your family member’s illness?’’ (0 ¼ no; 1 ¼ yes). Contributing Factors. Difficulty integrating death awareness was assessed by a single item that asked respondents to use a 5-point scale (1 ¼ strongly disagree; 5 ¼ strongly agree) to indicate the degree to which ‘‘thinking about my family member’s death is very difficult for me.’’ Four items embedded in the Family Assessment Device23 were utilized to measure communication constraints. Respondents were asked to use a 5-point scale (1 ¼ strongly agree to 5 ¼ strongly disagree) to answer the following items: ‘‘We cannot talk to each other about the sadness we feel,’’ ‘‘We avoid discussing our fears and concerns,’’ ‘‘We can express feelings to each other,’’ and ‘‘We confide in each other.’’ Two of the items were reverse coded for directionality, and higher scores on this measure indicated higher levels of communication constraints (alpha ¼ 0.82). Family asserting control was measured through a single item, which asked respondents to use a 5-point scale (1 ¼ not at all to 5 ¼ very much) to indicate the degree to which any family member had asserted control in decision making
Vol. 50 No. 2 August 2015
regarding patient care. Higher scores on this measure indicated higher levels of asserting control.
Statistical Analysis All variables were examined for missing values and fit between their distributions and the analytic assumptions. The outcome variable was positively skewed and a logarithm transformation was necessary to improve its distribution.24 Analysis of survey responses involved standard univariate analysis of major variables (i.e., descriptive statistics) followed by a series of bivariate and mutlivariate analyses. Zero-order correlations, t-tests, and analyses of variance were conducted to check for presence and size of statistical relationship between each of the context and condition variables, the contributing factors and the outcome of interest, as per the hypothesized relationships. Regression analysis was conducted, controlling for context and condition variables to examine the effect of contributing factors (i.e., death anxiety, communication constraints, family asserting control) on family conflict.
Results Sample Description Sample characteristics of caregivers and patients are described in Table 1. Most caregivers were adult children or spouses, female, white, and had a high school or college education. The mean ages were 63 and 80 years for caregivers and patients, respectively. Approximately 14% of caregivers had children under the age of 18 years living at home. Nearly half of the patients were being cared for in the home setting and slightly over half were female. The majority of patients had completed a power of attorney for health care (96%) and/or a living will (81%) before the illness, but only 63% had engaged in advance care planning conversations with the family. Typical for hospice programs nationwide, the largest percentage of patients had a primary diagnosis of cancer (44%), followed by heart disease (17%), and dementia (11%).
Presence of Family Conflict and Family Context: Hypothesis 1 Fifty-seven percent of caregivers reported experiencing one or more types of family conflict at the end of life. Overall, levels of family conflict at the end of life were relatively low in this sample as evidenced by the mean score of 1.39 (SD ¼ 0.63; range ¼ 1e4.5). Additionally, 48% of caregivers reported some family conflict before their family member’s illness. Of these, 41% reported a history of serious arguments and 32% reported a history of insulting and/or yelling within the family.
Vol. 50 No. 2 August 2015
Correlates and Predictors of Family Conflict
Table 1 Sample Characteristics of Family Caregivers and Patients Caregiver Relationship to patient (%) Spouse/significant other/life partner Daughter/son/daughter or son-in-law Sibling/sister or brother-in-law Parent Other family and family like friends Age in yrs Mean (SD) Gender (% female) Race (% white) Education (%) Less than high school High school graduate Post-high school education Yearly income (%) $19,999 or less 20,000e49,999 50,000e99,999 100,000þ Missing Presence of children in caregiver’s home (%) Duration of caregiving in months Mean (SD) Patient Age in years Mean (SD) Gender (% female) Lives at home or with other family member (%) Patient primary diagnosis (%) Cancer Heart disease Lung disease Kidney disease Alzheimer’s or other dementia ALS Failure to thrive Other illness Engaged in advance care planning before illness (%) Completed a power of attorney for health care Completed living will Advance planning discussions with family
27.3 55.3 4.3 3.1 9.9 59 (13) 84 98.8 5.6 48.4 46.0 16.7 39.8 23.6 9.9 9.9 13.7 72 (107) 80 (14) 56.5 47.8 43.8 16.9 4.4 3.8 10.6 0.6 7.5 12.5 96.3 80.8 62.9
ALS ¼ amyotrophic lateral sclerosis.
We hypothesized that conflict at the end of life will be higher among caregivers with a history of family conflict, more family demands (i.e., longer duration of caregiving, presence of children in the home), fewer resources (i.e., lower income, lower education, younger age), and less advance care planning (i.e. absence of advance directives or conversation before illness onset). Bivariate relationships between the contextual variables and the outcome variable revealed that family conflict at the end of life was significantly associated with a history of family conflict (r ¼ 0.51, P # 0.001), female gender (t ¼ 2.11, P # 0.05), younger caregiver age (r ¼ 0.23, P # 0.01), presence of children in the home (t ¼ 2.13, P # 0.05), and less family discussions about patient wishes before illness (r ¼ 0.24, P # 0.01). No relationship was found between conflict and education, income, duration of caregiving, relationship to the patient and presence of advance directives. This hypothesis is partially supported.
159
The Role Played by Conditions: Hypothesis 2 We hypothesized that certain conditions (i.e., patient care needs and ‘‘family coming out of woodwork’’) will be associated with contributing factors (i.e., difficulty integrating death awareness, communication constraints, asserting control) and family conflict at the end of life. Bivariate correlations revealed that more ADL impairments were associated with less difficulty integrating death awareness (r ¼ 0.16, P # 0.05). Greater IADL impairments were associated with family members asserting control (r ¼ 0.15, P # 0.05) and higher levels of family conflict (r ¼ 0.17, P # 0.05). ‘‘Coming out of the woodwork’’ correlated with greater difficulty integrating death awareness (r ¼ 0.15, P # 0.05), and higher levels of family conflict at trend level (r ¼ 0.13, P # 0.10). This hypothesis is partially supported.
The Influence of Contributing Factors: Hypothesis 3 We hypothesized that after controlling for family context and conditions that contributing factors will significantly predict family conflict. To test this third hypothesis, hierarchical multiple regression analysis was conducted that controlled for the family context and conditions of the patient and family. Because of the number of independent variables and the limited sample size, only the contextual variables that were significantly associated with family conflict were entered into the regression equation. The first block of independent variables included five contextual variables (i.e., history of conflict, caregiver age and gender, presence of children in the home, and advance care planning discussions). The second block included the two clinical care need variables (i.e., ADLs and IADLs) and ‘‘coming out of the woodwork.’’ The third block included the three contributing factors (i.e., difficulty integrating death awareness, communication constraints, and asserting control). Table 2 presents the results of the regression analysis, which explained 59% of the total variance in family conflict. In the first step of the model, family contextual variables explained 30% of the variance in conflict. Prior family conflict held the strongest relationship to end-of-life conflict, and two contextual variables (i.e., younger caregiver age and less advance planning discussions) revealed trend level associations. In the second step, conditional variables did not make a significant contribution to the prediction of family conflict as a set, explaining only 2% of the total variance. In the third step, contributing factors uniquely explained 27% of the variance in conflict, the most important two being communication constraints and asserting control. Significant main effects in the prediction of family conflict in the final model included prior family conflict (b ¼ 0.22, P # 0.001),
160
Kramer and Boelk
Table 2 Hierarchical Regression Analysis for Family Conflict Betas Variables Family context Prior family conflict Caregiver age Caregiver gender Presence of children Advance planning discussions Conditions Patient ADLs Patient IADLs ‘‘Coming out of the woodwork’’ Contributing factors Difficulty integrating death awareness Communication constraints Asserting control Total R2 2 R change F for R2 change
Step 1
Step 2
Step 3
0.41a 0.15b 0.10 0.02 0.13b
0.40a 0.11 0.13b 0.01 0.15d
0.23a 0.16c 0.10b 0.04 0.11d
0.01 0.14b 0.09
0.01 0.05 0.11d 0.05
0.30 0.30 12.41a
0.32 0.02 1.58
0.22a 0.46a 0.59 0.27 31.94a
ADL ¼ Activities of Daily Living; IADL ¼ Instrumental Activities of Daily Living. a P # 0.001. b P # 0.10 (two-tailed). c P # 0.01. d P # 0.05.
caregiver age (b ¼ 0.17, P # 0.01), caregiver gender (b ¼ 0.10, P # 0.10), advance care planning discussions (b ¼ 0.12, P # 0.05), family ‘‘coming out of the woodwork’’ (b ¼ 0.10, P # 0.10), communication constraints (b ¼ 0.22, P # 0.001), and family members asserting control (b ¼ 0.46, P # 0.001). Hypothesis 3 is supported.
Discussion Over the past decade, there has been substantial progress in identifying typologies of family functioning25e27 and their relationship to psychosocial morbidity in families receiving palliative care.25 Although conflict is an important dimension of family functioning,27 little is known about the precipitants of conflict itself. To our knowledge, this is the first study to investigate the correlates and predictors of conflict reported by family caregivers concurrently enrolled in hospice. Results support the multidimensional theoretical model of family conflict specifying the importance of the family context, key conditions that set the stage for conflict, and essential contributing factors. Three contextual variables associated with conflict that also were reported in recent studies include prior family conflict,12 younger age,12,26 and female gender.26 Although there was no relationship between conflict and advance directives, conflict was associated with advance planning discussions. Recent evidence
Vol. 50 No. 2 August 2015
suggests that more highly functioning and emotionally supportive families are more likely to have such discussions.28 Given the cross-sectional limitation of this study, it is not possible to disentangle these effects, although the relationship persisted after controlling for prior conflict. Although those with children in the home had higher levels of conflict, this relationship disappeared when other contextual variables were taken into consideration, possibly because younger women are more likely to have children at home. Although it is not surprising that families with a history of conflict are likely to experience conflict at the end of life, it is noteworthy that a subset of caregivers experiencing conflict (9%) reported no history of prior conflict within the family. After controlling for family context, caregivers who reported that family members were ‘‘coming out of the woodwork,’’ asserting control and experiencing communication constraints, were likely to experience higher levels of family conflict at the end of life. These contributing factors were predictive of conflict in a lung cancer caregiver study.12 These findings provide additional insights regarding families who may be at greater risk of conflict and the potential interventions that may be employed to help. It is possible that additional factors, not identified in our explanatory matrix, might also contribute to family conflict and would be worthy of further investigation. For example, perceptions of support and cohesiveness that are associated with conflict6 and other key processes associated with family resilience12 may influence whether conflict will arise and/or how it is communicated or dealt with. There are several limitations of our study. Given the racially homogeneous sample (99% white), and low response rate (58% once barriers to recruitment were addressed), results need be interpreted with caution. We cannot generalize these findings in the absence of a representative and more diverse sample. Although more than half of the caregivers interviewed reported some conflict, the degree of conflict was relatively low, which is consistent with several other studies reporting low rates of high conflict overall.15,25e26 It is quite possible that respondents were less burdened and may have had fewer demands, more resources, and overall better family relationship quality than non-respondents. The extent to which the correlates and predictors of conflict would be different among a more diverse sample is an empirical question worthy of future investigation. As a cross-sectional study, we cannot adequately evaluate the temporal ordering of variables. Some of our measures were limited. For example, the income variable categories were not equal in size and a large number of caregivers chose not to answer this question (10% were missing)
Vol. 50 No. 2 August 2015
Correlates and Predictors of Family Conflict
161
making it difficult to detect differences, and two of the contributing factors were measured with single-item indicators developed for this study. In addition, the single-item ‘‘asserting control’’ indicator reflects the caregiver’s perception of the extent to which any family member had asserted control in decision making, without validation by other family members. Further measurement development and testing of these constructs is needed. In addition, future studies eliciting multiple family members’ perceptions regarding various dimensions of the explanatory matrix would further enrich understanding of the family conflict phenomenon.
References
Conclusion
6. Kissane DW, Bloch S, Onghena P, McKenzie DP, Snyder RD, Dowe DL. The Melbourne Family Grief Study, II: psychosocial morbidity and grief in bereaved families. Am J Psychiatry 1996;153:659e666.
Our results have important implications for palliative care and future research. First, the results suggest the importance of implementing routine assessment protocols and monitoring of historical and current family relationship quality, demands, resources, advance care planning discussions, conditions, and contributing factors. As with the benefits demonstrated by clinicians who recommend routine assessment of family functioning,29 such protocols and monitoring would allow service providers to identify and target appropriate interventions to families at most risk. Second, our findings suggest the importance of interventions for promoting and facilitating open communication, shared decision making, and advance care planning discussions as a potential means to prevent conflict at the end of life. Finally, while this study provides a preliminary understanding of the correlates and predictors of conflict at the end of life, there is much that remains to be understood. Prospective studies representing more diverse family forms are needed. Future research is needed to examine conflict that arises in families that have not previously experienced it and the extent to which conflict might vary by socioeconomic status and more diverse family forms. Translational questions that may be investigated include: What, if anything, might be done to prevent or to address the ‘‘coming out of the woodwork’’ phenomenon? What interventions might alleviate communication constraints and promote shared decision making rather than assertions of control?
Disclosures and Acknowledgments This work was supported in part by the University of Wisconsin-Stevens Point Personnel Development Committee. No competing financial interests exist. The authors express their gratitude to the family caregivers who participated in this study.
1. Waldrop DP, Kramer BJ, Skrentny JA, Milch RA, Finn W. Final transitions: family caregiving at the end of life. J Palliat Med 2005;8:623e638. 2. Kehl K, Kirchoff K, Kramer BJ, Hovland-Scafe C. Challenges facing families at the end-of-life in three settings. J Soc Work End Life Palliat Care 2010;5:144e168. 3. Broom K, Kirby E. The end of life and the family: hospice patients’ views on dying as relational. Sociol Health Illn 2013;35:499e513. 4. Strawbridge WJ, Wallhagen MI. Impact of family conflict on adult child caregivers. Gerontologist 1991;31:770e777. 5. Semple SJ. Conflict in Alzheimer’s caregiving families: its dimensions and consequences. Gerontologist 1992;32: 648e655.
7. Kramer BJ, Kavanaugh M, Trentham-Dietz A, Walsh M, Yonker JA. Complicated grief in caregivers of persons with lung cancer: the role of family conflict, intrapsychic strains and hospice utilization. Omega 2010e2011;62:201e220. 8. Kramer BJ, Yonker JA. Perceived success in addressing end-of-life care needs of low-income elders and their families: what’s family conflict got to do with it? J Pain Symptom Manage 2011;41:35e48. 9. Kayashima R. Barriers to good end-of-life care: a physician survey. Hawaii Med J 2001;60:40e44. 10. National Consensus Project for Quality Palliative Care. Clinical practice guidelines for quality palliative Care. 2nd ed. 2009. Available at: http://www.nationalconsensusproject. org. Accessed June 6, 2011. 11. Kramer BJ, Boelk A, Auer C. Family conflict at the endof-life: lessons learned in a model program for vulnerable older adults. J Palliat Med 2006;9:791e801. 12. Kramer BJ, Kavanaugh M, Trentham-Dietz A, Walsh M, Yonker JA. Predictors of family conflict at the end of life: the experience of spouses and adult children of persons with lung cancer. Gerontologist 2010;50:215e225. 13. Boelk AZ, Kramer BJ. Advancing theory of family conflict at the end of life: a hospice case study. J Pain Symptom Manage 2012;44:655e670. 14. Clark PC, Shields CG, Ayccock D, Wolf SL. Preliminary reliability and validity of a family caregiver conflict scale for stroke. Prog Cardiovasc Nurs 2003;18:77e82. 92. 15. Kissane DW, Bloch S, Burns WI, McKenzie DP, Posterino M. Psychological morbidity in the families of patients with cancer. Psychooncology 1994;3:47e56. 16. Kessler RC, Molnar BE, Feurer ID, Appelbaum M. Patterns and mental health predictors of domestic violence in the United States: results from the National Comorbidity Survey. Int J Law Psychiatry 2001;24:487e508. 17. Tilden VP, Tolle SW, Garland MJ, Nelson CA. Decisions about life-sustaining treatments. Arch Intern Med 1995;155: 633e638. 18. Su CT, McMahan BS, Williams BA, Sharma RK, Sudore RL. Family matters: effects of birth order, culture,
162
Kramer and Boelk
and family dynamics on surrogate decision-making. J Am Geriatr Soc 2014;62:175e182. 19. Chan CW, Chang AM. Managing caregiver tasks among family caregivers of cancer patients in Hong Kong. J Adv Nurs 1999;29:484e489. 20. NHPCO Facts and Figures: Hospice Care in America. Alexandria, VA: National Hospice and Palliative Care Organization, 2012. 21. Katz S, Down TD, Cash HR, Grotz RC. Progress in the development of the index of ADL. Gerontologist 1970;10:20e30. 22. Lawton MP, Brody EM. Assessment of older people: selfmaintaining and instrumental activities of daily living. Geron 1969;9:179e186. 23. Miller IW, Epstein NB, Bishop DS, Keitner GI. The McMaster Family Assessment Device: reliability and validity. J Marital Fam Ther 1985;11:345e356. 24. Tabachnick BG, Fidell LS. Using multivariate statistics, 4th ed. Boston, MA: Allyn & Bacon, 2001.
Vol. 50 No. 2 August 2015
25. Kissane DW, McKenzie M, McKenzie DP, O’Neill AF, Bloch S. Psychosocial morbidity associated with patterns of family functioning in palliative care: baseline data from the family focused grief therapy controlled trial. Palliat Med 2003;17:527e537. 26. Schuler TA, Zaider TI, Li Y, Hichenberg S, Masterson M, Kissane DW. Typology of perceived family functioning in an American sample of patients with advanced cancer. J Pain Symptom Manage 2014;48:281e288. 27. Kissane D, Bloch S. Family focused grief therapy. Berkshire, UK: Open University Press, 2010. 28. Boerner K, Carr D, Moorman S. Family relationships and advance care planning: do supportive and critical relations encourage or hinder planning. J Gerontol B Psychol Sci Soc Sci 2013;68:246e256. 29. King DA, Quill T. Working with families in palliative care: one size does not fit all. J Palliat Med 2006;9: 704e715.