Abstract
Context:
Surrogate decision makers experience significant amounts of anxiety, burden, and strain in their role as caregivers and decision makers for loved ones.
Objectives:
To investigate whether surrogate decision makers engaging in ACP together with their loved one longitudinally reduces perceived anxiety, burden, and strain felt by surrogate decision makers.
Methods:
Post-hoc analysis of a randomized controlled trial evaluating caregivers’ perceived self-efficacy to serve as surrogate decision makers. The trial employed a 2×2 study design of patient/caregiver dyads, who engaged in advance care planning (ACP) using a standard living will form vs “Making your wishes Known” (MYWK), and having the patient engage in ACP alone vs together with the family caregiver. Surrogates completed surveys longitudinally to compare levels of anxiety, burden, and strain.
Results:
246 of 285 dyads completed the validated instruments measuring anxiety, burden, or strain. No significant reductions in anxiety, burden, or strain were found longitudinally in surrogate decision makers using MYWK together with loved one’s vs other control groups. Increases in strain and anxiety were seen across all study groups and increases in burden across 2/4 study groups. Strain and burden increased most in the MYWK Together arm (▲ = +2.22 and ▲ = +1.91 respectively).
Conclusion:
Family caregivers who engaged in ACP together with patients using the decision support tool MYWK did not experience less strain, burden, or anxiety longitudinally compared to other study arms. These results may help inform the design of future studies and interventions that promote caregivers’ involvement in ACP interventions.
Keywords: Advance care planning, Surrogate decision making, Caregiver burden, Caregiver anxiety, Caregiver strain, ACP
Introduction:
Advance Care Planning (ACP), the process of planning for future medical decisions, can help family members and clinicians honor a patient’s wishes. However, for those tasked with making medical decisions for others, the process can be both stressful and anxiety-provoking. Up to 80% of family members who share in end-of-life decision making for patients in the ICU develop moderate to severe Post Traumatic Stress Disorder. 1 In fact, caregivers for patients with chronic conditions may be more likely to suffer from depression and anxiety than patients themselves. 2 Caregivers who experience heightened mental and emotional strain also experience higher mortality rates. 3 With the growing number of Americans assuming the role of caregiver, there is a need to better understand how to support these individuals in their role as surrogate decision-makers. 2, 4, 5
One factor that may influence caregiver strain, burden, and anxiety is a caregiver’s uncertainty as to what medical care a loved one actually wants to receive. For this reason, some researchers have suggested that robust advance care planning may reduce stress and burden among caregivers who serve as surrogate decision makers. 2, 6, 7 In prior work, we reported the benefit of one such decision support tool, Making Your Wishes Known (MYWK), in clarifying patients’ values, goals and preferences for medical treatments. The tool has also been shown to benefit clinicians, whose predictions about patients’ wishes improved after using the tool. 8 With this in mind, we set out to examine whether the use of MYWK would also help family caregivers to be better prepared for the role of surrogate decision-makers by reducing perceived anxiety, burden, and strain during the time period immediately prior to surrogate decision making.
In this paper, we address the question of whether the use of MYWK by family caregivers together with patients is associated with reductions in caregiver strain, burden, and anxiety compared to ACP done without caregivers and/or when caregivers do not use MYWK. We hypothesized that completing ACP together (patients and caregivers) using the MYWK decision tool would be associated with lower levels of perceived anxiety, burden, and strain for caregivers compared to the other groups.
Methods:
The parent study was a randomized, controlled trial designed to evaluate the effect of MYWK on caregivers’ perceived self-efficacy to serve as surrogate decision makers. The trial employed a 2×2 study design of patient/caregiver dyads, randomizing participants to engage in advance care planning (ACP) using a standard living will form vs MYWK, and having the patient engage in ACP alone vs doing so together with the family caregiver.
Data Collection:
After obtaining IRB approval, patients and caregivers were recruited as dyads, starting with the patients who then identified a family caregiver. We sought to enroll patients who were chronically ill and were expected to lose decision-making capacity in the following 18–24 months, thus requiring a surrogate to make a medical decision on their behalf. To diversify the sample, we recruited patients from three hospitals in two geographic areas: the Milton S. Hershey Medical center in Hershey, PA; Pinnacle Hospital (now UPMC) in Harrisburg, PA; or Brigham and Women’s Hospital in Boston, MA. Patients were eligible if they were: ≥18 yrs old; diagnosed with end stage renal disease, Stage IV cancer, Class III or Class IV heart failure, or Stage III or Stage IV lung disease by modified GOLD Spirometric Classification; able to read and understand English at an 8th grade level; neuro-cognitively able to engage in ACP (Mini Mental State Exam (MMSE) score >23); and lacking suicidal ideations. Caregivers were eligible if they were ≥18 years of age and identified by the patient as a family caregiver who would be the patient’s surrogate decision-maker if the patient lost decision-making capacity (if non-cohabitant, must have in-person interaction with patient at least once weekly). All caregivers were required to meet the same minimum cognitive standards as patients.
Dyads were divided into 4 study arms based on the type of ACP tool used by the patient (Standard or MYWK), and whether the caregiver was involved in the completion of the of ACP tool. The study arms were as follows: 1) patient completed Standard ACP alone (STA) (using a standard living will form + educational pamphlet) without the caregiver present; 2) patient and caregiver completed Standard ACP together (STT); 3) patient completed ACP alone using MYWK (MYWKA); 4) patient and caregiver completed MYWK together (MYWKT).
Data was collected using validated instruments to measure levels of anxiety, caregiver burden, and caregiver strain over time. 9–11 At baseline, caregivers completed the following: 1) the PROMIS emotional distress anxiety short form; 2) Zarit burden interview short form; and 3) Caregiver Strain Index.
The PROMIS emotional distress anxiety short form consists of 8 questions asking individuals how often they have felt various forms of anxiety over the last 7 days. It is scored based on a 5-point Likert scale from 1–5 (never, rarely, sometimes, often, or always). Numerical values are then summed ranging from 5–40. Higher/increasing scores are indicative of greater/increasing levels of anxiety.
The Zarit Burden interview short form consists of 12 questions aimed at understanding how much burden an informal caregiver is experiencing while caring for their loved one. The Zarit is scored from 12–60 based on a Likert scale where responses are recorded from 1–5 (never, rarely, sometimes, quite frequently, or always). Corresponding numerical values are summed, with higher/increasing scores indicating a greater/increasing level of caregiver burden.
The Caregiver Strain Index is a 13 question Yes/No survey that measures common stressors experienced by caregivers (1 point awarded for each response of “Yes”). Scores are summed and scored from 0–13, with higher scores indicating higher caregiver strain.
After completing the instruments during the first visit (Visit 1), caregivers and patients participated in one of the 4 ACP intervention arms described above. Four weeks later (Visit 2), patients and caregivers were asked to respond to case vignettes regarding patients’ wishes for treatment under various scenarios. To understand sources of stress, caregivers were interviewed by telephone every 3–4 months following the intervention. At each of these interviews they again completed the 3 validated instruments. This continued until the patient had a serious medical event potentially requiring the surrogate to make a medical decision. When this occurred, the caregiver was interviewed for a final time. Because of the unpredictability of serious medical events taking place, the study design dictated that some caregivers would complete surveys more times than others, since dyads progressed into different phases of the study at varying times.
Data Analysis:
All analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC), and all variables were summarized prior to analysis. Study groups were compared by demographic variables using analysis of variance or Chi-square tests (Table 1).
Table 1.
Caregiver Demographics
| Characteristic | MYWKT | MYWKA | STT | STA | P-Value |
|---|---|---|---|---|---|
| Caregiver | Caregiver | Caregiver | Caregiver | ||
|
| |||||
| Count (N) | N=67 | N=−56 | N=62 | N=61 | |
|
| |||||
| Age (years), Mean ± SD | 53.52 ± 14.83 | 57.75 ± 12.57 | 56.87 ± 12.68 | 56.21 ± 15.32 | 0.036 |
|
| |||||
| Gender, N (%) | 0.73 | ||||
| Male | 17 (25.37%) | 11 (19.64%) | 17 (27.42%) | 17 (27.87%) | |
| Female | 50 (74.63%) | 45 (80.36%) | 45 (72.58%) | 44 (72.13%) | |
|
| |||||
| Ethnicity, N (%) | 0.16 | ||||
| Hispanic | 2 (3.08%) | 1 (1.89%) | 2 (3.33%) | 6 (9.84%) | |
| Non-Hispanic | 63 (96.92%) | 52 (98.11%) | 58 (96.67%) | 55 (90.16%) | |
|
| |||||
| Race, N (%) | 0.78 | ||||
| African American | 15 (22.39%) | 10 (17.86%) | 11 (17.74%) | 12 (20.00%) | |
| Caucasian | 50 (74.63%) | 44 (78.57%) | 46 (74.19%) | 43 (71.67%) | |
| Other | 2 (2.99%) | 2 (3.57%) | 5 (8.06%) | 5 (8.33%) | |
|
| |||||
| Religion, N (%) | 0.49 | ||||
| Catholic | 12 (17.91%)) | 12 (21.82%) | 13 (20.97%) | 13 (21.31%) | |
| Protestant | 28 (41.79%) | 21 (38.18%) | 29 (46.77%) | 23 (37.70%0 | |
| Other Christian | 16 (23.88%) | 6 (10.91%) | 8 (12.90%) | 13 (21.31%) | |
| Other Religion | 7 (10.45%) | 6 (10.91%) | 8 (12.90%) | 7 (11.48%) | |
| No Formal Religion | 4 (5.97%) | 10 (18.18%) | 4 (6.45%) | 5 (8.20%) | |
|
| |||||
| Education, N (%) | 0.02 | ||||
| High school or less | 22 (32.84%) | 21 (37.50%) | 11 (17.74%) | 12 (19.67%) | |
| Some college | 25 (37.31%) | 14 (25.00%) | 26 (41.94%) | 16 (26.23%) | |
| College graduate | 20 (29.85%) | 21 (37.50%) | 25 (40.32%) | 33 (54.10%) | |
|
| |||||
| Previous AD, N (%) | 15 (22.39%) | 22 (39.29%) | 17 (27.42%) | 20 (32.79%) | 0.21 |
|
| |||||
| Relationship to patient, N (%) | 0.49 | ||||
| Spouse/domestic partner | 40 (59.70%) | 36 (64.29%) | 33 (53.23%) | 32 (52.46%) | |
| Parent | 19 (28.36%) | 10 (17.86%) | 16 (25.81%) | 15 (24.59%) | |
| Sibling | 1 (1.49%) | 2 (3.57%) | 4 (6.45%) | 1 (1.64%) | |
| Son/daughter | 2 (2.99%) | 2 (3.57%) | 1 (1.61%) | 5 (8.20%) | |
| Other relative | 1 (1.49%) | 0 (0.00%) | 3 (4.84%) | 2 (3.28%) | |
| Friend | 1 (1.49%) | 5 (8.93%) | 4 (6.45%) | 5 (8.20%) | |
| Other | 3 (4.48%) | 1 (1.79%) | 1 (1.61%) | 1 (1.64%) | |
|
| |||||
| Study site, N (%) | 0.87 | ||||
| Hershey | 51 (76.12%) | 45 (80.36%) | 47 (75.81%) | 45 (73.77%) | |
| Boston | 16 (23.88%) | 11 (19.64%) | 15 (24.19%) | 16 (26.23%) | |
The distributions of continuous variables were assessed using histograms, box plots, and tests for normality. Visit 1 of the study was considered as baseline, and the time from each interview to visit 1 in months was calculated. Only data from the first 24 months of phone interviews were used for analysis due to the attrition rate and the need to maintain an adequate sample size. The median scores were compared between study groups at visit 1 using Kruskal Wallis tests and Wilcoxon Rank Sum tests, but no differences were found. Because the original scores were not normally distributed, the change from visit 1 to each interview was calculated, and a repeated measures model including factors for the study group, interview month, and the interaction between the two factors was employed to determine any differences between study groups over time in the slope of the change from baseline. The slope of the change over time was compared between study groups rather than comparing the mean change between study groups at specific points in time because the interview months were not consistent for individual subjects.
Results:
Of the 285 dyads enrolled in the study, 246 completed both Visits 1 and 2, and completed all of the validated instruments. The mean age of these 246 caregivers was 56 years (range 19 to 89 years), 75% (n=184) were female, and 69.9% (n=172) had not completed an advance directive or living will prior to enrollment in the study. The characterization of the caregiver’s relationship to the patient included spouse/domestic partner (57.3%), parent (24.4%), sibling (3.3%), son/daughter (4.0%), other relative (2.4%), friend (6.1%), or other (2.4%). Of the patients enrolled in the study, 28.1% (n=69) had underlying cardiac disease, 22.3% (n=55) pulmonary disease, 31.3% (n=77) cancer, and 18.3% (n=45) renal disease. Demographic characteristics of participants in each of the four study groups are found in Table 1.
Longitudinal measures:
Over the course of the study, strain (as measured by the Caregiver Strain Index) increased from baseline in all study arms (Table 2). While this was expected given the trajectory of the patients’ illnesses, we were surprised to find that participation in the MYWK (T) arm was associated with the highest increase in caregiver strain from baseline among the study arms (▲ = +2.22). The difference in the slope of change in the MYWK (A) and MYWK (T) groups were both positive with the MYWK (T) group being the largest showing a greater change from visit 1 over time while both Standard groups had negative slopes. However, there were no significant differences between any of the groups.
Table 2.
Perceived Strain, Burden, and Anxiety Measured in Caregivers
| Group | Strain | Burden | Anxiety | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Initial | Final | Change | Slope | Initial | Final | Change | Slope | Initial | Final | Change | Slope | |
| MYWK (T) | 3.97 | 6.18 | 2.22 | 0.051 | 21.80 | 23.70 | 1.91 | 0.044 | 52.53 | 54.51 | 1.98 | 0.030 |
| MYWK (A) | 4.47 | 5.86 | 1.39 | 0.0296 | 22.93 | 23.95 | 1.02 | 0.102 | 51.20 | 52.72 | 1.52 | 0.098 |
| Standard (T) | 3.55 | 4.87 | 1.33 | −0.019 | 22.79 | 22.55 | −0.24 | −0.026 | 51.19 | 51.71 | 0.52 | −0.039 |
| Standard (A) | 3.51 | 4.86 | 1.35 | −0.002 | 21.41 | 21.33 | −0.08 | −0.014 | 49.63 | 51.80 | 2.16 | 0.036 |
**Reminder: Higher scores relate to increases in strain, burden, and stress.**
Caregiver burden (as measured by the Zarit Burden Index) increased from baseline in 2 of the 4 groups and participation in the MYWK (T) group was associated with the highest increase (▲ = +1.91). (Table 2) However, there were no significant differences in the slope of change when comparing MYWK (T) to the other groups.
Anxiety (as measured by the PROMIS emotional distress anxiety short form) increased from baseline in all study groups. Participation in the Standard (A) group was associated with the highest increase (▲ = 2.16). (Table 2) There were no significant differences in the slope of change when comparing MYWK (T) to the other groups.
Discussion:
In this post-hoc analysis of a randomized controlled trial that used a 2×2 factorial design, family caregivers who engaged in advance care planning (ACP) together with patients using the decision support tool MYWK did not experience less strain, burden, or anxiety over time–as compared to caregivers who either did not participate in ACP together (2 groups) or who did participate in ACP together but did so using a Standard ACP tool (1 group).
These findings were counterintuitive insofar as we had expected that by better preparing family caregivers for surrogate decision making (though a rigorous ACP planning tool and working closely with the patient), would lead to relative reductions in the adverse outcomes over time (compared to the other groups). From prior research, we know that MYWK increases both patients’ knowledge about ACP 12 and caregivers’ knowledge about their loved one’s wishes regarding end-of-life medical care. 13 As such, the research team anticipated that having patients and their family caregivers complete MYWK together would (if anything) ease caregivers’ distress.
There are several possible explanations for these unexpected results. First, the findings may be related to the exploratory nature of the analysis. While the endpoints reported here were pre-specified, and the respective instruments were built into the parent study, the overall study design was powered to detect changes in caregiver self-efficacy, not differences over time in caregiver distress. Although this limits the generalizability of the present results, insofar as (to our knowledge) this is the first study to systematically measure these endpoints over time with family caregivers of patients with severe chronic illness.
It is also possible that inviting caregivers to participate in ACP with patients using a decision support tool compelled them to confront uncomfortable realities regarding their role as surrogate decision-makers. Surrogate decision-making is stressful, particularly when surrogates disagree with patients about what treatments to provide or withhold. It is not unreasonable to surmise that such disagreements could account for the increase in strain, burden, or anxiety. In fact, findings from the parent RCT showed that caregiver/loved one disagreement (measured via hypothetical vignettes) was dramatically higher (~30–50%) when patients indicated they would defer/decline life-sustaining interventions.14 It is possible that knowing that one disagrees with their loved one’s wishes for end-of-life care in various circumstances could make the task of caregiving more strenuous, burdensome, or anxiety-provoking –especially if caregivers feel obliged to support those end-of-life decisions. Indeed, we find some evidence of this in the fact that mean changes in strain and burden over time were numerically highest in the MYWK (T) group. (Table 2) Finally, and perhaps most provocatively, it is possible that in the context of family caregivers for patients with severe chronic illness, standard quantitative instruments and scales do not accurately measure complex psychologic constructs like strain, burden, or anxiety. As discussed by Bastawrous, researchers may conceptulize and capture burden and stress in ways that are different from what caregivers are experiencing. For this reason, itemized measures may fall short in capturing the complexity of strain, burden, or anxiety experienced by patients and caregivers. 15
In conclusion, we did not find that family caregivers who engaged in ACP together with patients using the decision support tool MYWK experienced less strain, burden, or anxiety over time compared to caregivers in the other study groups. Knowledge of these negative findings may help clinicians and researchers who wish to advance the science of ACP, particularly in terms of caregivers’ experience. Specifically, these results may help inform the design of future studies and interventions that promote caregivers’ involvement in ACP interventions targeting caregivers and should not be relegated to the “file-drawer.” 16
Funding:
Supported by grant NCT02429479 (M.J.G. and B.H.L.) and National Institute of Nursing Research award 5R01NR012757 from the National Institutes of Health.
Footnotes
Conflict of Interest Statement:
Two of the authors (M.J. Green and B.H. Levi) have intellectual property and copyright interests for the decision aid Making Your Wishes Known (MYWK). To date, MYWK has been made available free of charge for use in research and education. None of the other authors have any conflicts of interest.
References:
- 1.Azoulay E, Pochard F, Kentish-Barnes N, et al. Risk of post-traumatic stress symptoms in family members of intensive care unit patients. Am J Respir Crit Care Med. May 1 2005;171(9):987–94. doi: 10.1164/rccm.200409-1295OC [DOI] [PubMed] [Google Scholar]
- 2.Grunfeld E, Coyle D, Whelan T, et al. Family caregiver burden: results of a longitudinal study of breast cancer patients and their principal caregivers. Cmaj. Jun 8 2004;170(12):1795–801. doi: 10.1503/cmaj.1031205 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Schulz R, Beach SR. Caregiving as a Risk Factor for Mortality The Caregiver Health Effects Study. JAMA. 1999;282(23):2215–2219. doi: 10.1001/jama.282.23.2215 [DOI] [PubMed] [Google Scholar]
- 4.Karmilovich SE. Burden and stress associated with spousal caregiving for individuals with heart failure. Prog Cardiovasc Nurs. Winter 1994;9(1):33–8. [PubMed] [Google Scholar]
- 5.Clark AM, Reid ME, Morrison CE, Capewell S, Murdoch DL, McMurray JJ. The complex nature of informal care in home-based heart failure management. Journal of Advanced Nursing. 2008;61(4):373–383. doi: 10.1111/j.1365-2648.2007.04527.x [DOI] [PubMed] [Google Scholar]
- 6.Song MK, Ward SE, Fine JP, et al. Advance care planning and end-of-life decision making in dialysis: a randomized controlled trial targeting patients and their surrogates. Am J Kidney Dis. Nov 2015;66(5):813–22. doi: 10.1053/j.ajkd.2015.05.018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Sautter JM, Tulsky JA, Johnson KS, et al. Caregiver Experience During Advanced Chronic Illness and Last Year of Life. Journal of the American Geriatrics Society. 2014;62(6):1082–1090. doi: 10.1111/jgs.12841 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Levi BH, Simmons Z, Hanna C, et al. Advance care planning for patients with amyotrophic lateral sclerosis. Amyotroph Lateral Scler Frontotemporal Degener. Aug 2017;18(5–6):388–396. doi: 10.1080/21678421.2017.1285317 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sullivan MT. Caregiver Strain Index (CSI). Urol Nurs. Jun 2007;27(3):251–2. [PubMed] [Google Scholar]
- 10.Salsman JM, Schalet BD, Merluzzi TV, et al. Calibration and initial validation of a general self-efficacy item bank and short form for the NIH PROMIS(®). Qual Life Res. Sep 2019;28(9):2513–2523. doi: 10.1007/s11136-019-02198-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Gratão ACM, Brigola AG, Ottaviani AC, et al. Brief version of Zarit Burden Interview (ZBI) for burden assessment in older caregivers. Dement Neuropsychol. Jan-Mar 2019;13(1):122–129. doi: 10.1590/1980-57642018dn13-010015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Green MJ, Schubart JR, Whitehead MM, Farace E, Lehman E, Levi BH. Advance Care Planning Does Not Adversely Affect Hope or Anxiety Among Patients With Advanced Cancer. J Pain Symptom Manage. Jun 2015;49(6):1088–96. doi: 10.1016/j.jpainsymman.2014.11.293 [DOI] [PubMed] [Google Scholar]
- 13.Green MJ, Van Scoy LJ, Foy AJ, et al. A Randomized Controlled Trial of Strategies to Improve Family Members’ Preparedness for Surrogate Decision-Making. Am J Hosp Palliat Care. Jun 2018;35(6):866–874. doi: 10.1177/1049909117744554 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Foy AJ, Levi BH, Van Scoy LJ, Bucher A, Dimmock A, Green MJ. Patient Preference to Accept Medical Treatment Is Associated with Spokesperson Agreement. Ann Am Thorac Soc. Apr 2019;16(4):518–521. doi: 10.1513/AnnalsATS.201806-428RL [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Bastawrous M. Caregiver burden--a critical discussion. Int J Nurs Stud. Mar 2013;50(3):431–41. doi: 10.1016/j.ijnurstu.2012.10.005 [DOI] [PubMed] [Google Scholar]
- 16.Scargle JD. Publication Bias: The “File-Drawer” Problem in Scientific Inference. Society for Scientific Exploration: Society for Scientific Exploration; 2000. p. 99–106. [Google Scholar]