Digestive Diseases and Sciences (2023) 68:3625–3633

https://doi.org/10.1007/s10620-023-08042-9

ORIGINAL ARTICLE

CAREGIVER Randomized Trial of Two Mindfulness Methods to Improve

the Burden and Distress of Caring for Persons with Cirrhosis
Elliot B. Tapper1,3 · Zachary M. Saleh1 · Sofia Lizza2 · Xi Chen1 · Samantha Nikirk1 · Marina Serper2

Received: 10 February 2023 / Accepted: 12 July 2023 / Published online: 27 July 2023
© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2023

Abstract
Background Patient caregivers experience burden and distress that negatively impacts health-related quality of life (HRQOL).
Mindfulness may alleviate caregiver burden but randomized trials of mindfulness activities on caregiver burden and distress
are lacking.
Methods Caregivers for patients with advanced liver disease were recruited from the Universities of Michigan and Penn-
sylvania (12/2019–12/2021) and followed for 8 weeks. Participants were randomized 1:1:1 to: written emotional disclosure
for 4 weeks, resilience training for 4 weeks, or control (no active intervention). All completed assessments at baseline, week
4 and 8. The primary outcome was change in the Zarit Caregiver Burden Index-12 (ZBI) at week 8. Secondary outcomes
included changes at week 4 and 8 in the ZBI, distress thermometer (DT), HRQOL visual analog scale (VAS), and caregiver
captivity index (CCI).
Results Eighty seven caregivers were enrolled, 59(72%) completed the study. In unadjusted analyses at week 4, the burden
measured by ZBI was not significantly different between arms, increasing by 0.6 ± 5.7 and 2.5 ± 5.2 points, for the written
emotional disclosure and resilience training study arms, respectively, and by 2.9 ± 6.1 points in the control arm. At week 8,
the non-significant ZBI change was − 1.0 ± 8.9, 2.8 ± 6.1, and 1.5 ± 7.4 for written emotional disclosure, resilience train-
ing, at week 8, respectively. The DT and VAS worsened in all arms, however, it worsened the least in the written emotional
disclosure arm. In analyses adjusted for differences in baseline characteristics, the ZBI declined by − 4.21 ± 2.03(p = 0.04)
in the emotional disclosure arm at week 4. This decrease was attenuated week 8, − 1.13 ± 2.6(p = 0.67). There were no sig-
nificant differences in secondary outcomes save for resilience training reducing the CCI at week 4 by 1.36 ± 0.67(p = 0.04).
Conclusion Written emotional disclosure may reduce caregiver burden in the short term among caregivers for patients with
cirrhosis.
Registration: NCT04205396.

Keywords Hepatic encephalopathy · Ascites · Liver disease · Resilience

* Elliot B. Tapper Marina Serper

[email protected] [email protected]
Zachary M. Saleh 1
Division of Gastroenterology and Hepatology, University
[email protected]
of Michigan, Ann Arbor, MI, USA
Sofia Lizza 2
Division of Gastroenterology and Hepatology, University
[email protected]
of Pennsylvania Perelman School of Medicine, Philadelphia,
Xi Chen PA, USA
[email protected] 3
3912 Taubman, SPC 5362, 1500 E. Medical Center Dr.,
Samantha Nikirk Ann Arbor, MI 48109, USA
[email protected]

13
Vol.:(0123456789)
3626 Digestive Diseases and Sciences (2023) 68:3625–3633

Introduction disclosure or resilience training interventions administered

to caregivers of patients with cirrhosis will improve their
Cirrhosis often results in disability, frequent hospitaliza- self-reported burden, distress, and quality of life at 4 and
tion, and death [1]. Complications such as hepatic enceph- 8 weeks. Participants were enrolled at two sites between
alopathy (HE) and ascites render patients frail, unable to December 2019 and December 2021. All participants were
drive, and dependent on medication regimens that require offered a $25 honorarium. This study was approved sepa-
close monitoring and active dose titration [2–5] Informal rately by the Institutional Review Boards at the University
caregivers, therefore, play a major role in the manage- of Michigan Medical School and the University of Penn-
ment of cirrhosis [4, 6]. Cirrhosis is independently asso- sylvania. The study was registered as NCT04205396. All
ciated with more caregiving requirements and caregivers authors had access to the data and reviewed and approved
of patients with cirrhosis experience a substantial burden the final manuscript.
that impacts their own physical and mental health [2–5,
7, 8]. Interventions to reduce the burden on caregivers of Population
patients with cirrhosis are limited.
Mindfulness-based methods may reduce stress among We enrolled adults ≥ 18 years old who were self- or patient-
caregivers for older adults [9–11]. but randomized trials identified primary caregivers for patients with decompen-
of such methods for caregivers of patients with cirrhosis sated cirrhosis or advanced hepatocellular carcinoma (HCC).
are lacking. In a caregiver-focused intervention, Bajaj et al. Decompensated cirrhosis was defined as: (1) cirrhosis with
performed a 1-month long nonrandomized, in-person mind- any hospitalization within the previous 30 days; (2) at least
fulness-based intervention for 20 patient/caregiver dyads one of the following within the prior 6 months from the date
[6]. The study included weekly group therapy sessions and of consent: ascites (requiring paracentesis or diuretics), overt
materials with mindfulness exercises. Caregivers reported HE (requiring lactulose or rifaximin, metronidazole, or neo-
substantial improvements in their Zarit Burden Interview mycin), spontaneous bacterial peritonitis, hepatic hydrotho-
scores and depression. However, studies of home-based, low rax (requiring diuretics or thoracentesis), or variceal bleed.
touch, and low-cost interventions to support caregivers are Advanced HCC was defined as having Barcelona Clinic Liver
lacking in cirrhosis. We developed the CAREGIVER trial Cancer Stage C or D. Caregivers were recruited by screen-
to remotely examine two methods of delivering mindfulness ing outpatient and inpatient schedules, by clinician referral,
interventions to address this gap in knowledge. and by directly emailing patients. Caregivers were defined by
Recognizing that putting personal experiences into a the patient as the primary relative, spouse, partner, friend, or
story—whether in written or spoken form—is associated neighbor who has a significant personal relationship with and
with both physical and mental benefits across diverse popu- provides a broad range of assistance for a person with decom-
lations [12], we tested two journaling-based mindfulness pensated cirrhosis, including consistent involvement in their
interventions. One group was asked to perform written emo- medical care. Exclusion criteria were non-English speaking,
tional disclosure, writing about important personal experi- unable/unwilling to provide consent, severe cognitive impair-
ences in an emotional way. This has been associated with ment, concurrent participation in another interventional study,
improved psychological outcomes in other settings [13–16]. or use of a personal diary in the last 12 months (due to similar-
The second group was asked to attempt to re-frame their ity to the written response intervention).
experiences through the tools of positive psychology (resil-
ience training). Using the ‘few good things’ exercise of daily Randomization
reflection previously shown to reduce healthcare provider
burnout, participants were asked to write down three things After informed consent procedures were completed, the study
that went well each day and why [17]. Herein, we report the coordinator randomized the subject using the TATUM soft-
final results of the CAREGIVER trial. ware (at Michigan) or randomization feature on REDCap soft-
ware (at Pennsylvania) in equal allocation to each of the three
parallel arms 1:1:1: written emotional disclosure, resilience
training, or control, stratified by baseline caregiver burden
Methods severity (ZBI > 12 vs. ≤ 12) in blocks of 4.

Design Overview

We conducted a randomized, controlled, three arm paral-

lel group efficacy trial to determine if written emotional

13
Digestive Diseases and Sciences (2023) 68:3625–3633 3627

Interventions The study coordinator contacted participants in the inter-

vention arms using their preferred self-identified method
Written Emotional Disclosure (e.g., phone, text or email) to remind them to complete
each prompt. Intervention participants were provided pre-
While at home, participants were asked to write a story addressed and stamped envelopes to return their prompts.
in response to five prompts regarding being a caregiver of
someone with decompensated cirrhosis for at least 15 min Study Timeline
every 3 days for 2 weeks. The rules provided were that: the
content of the writing relates to the prompt, participants All three arms followed the same study timelines (Fig. 1).
write for at least 15 min and do so without stopping without Study Day 1 was the day of the baseline assessments. Par-
regard to spelling, grammar or sentence structure. The five ticipants randomized to either one of the treatment arms
separate prompts were: “(1) Tell me about yourself; (2) Do completed their five written prompts on study days 4, 7, 10,
you ever experience guilt as a caregiver? Why; (3) What 13 and 16 (± 1 day) because spaced, rather than consecutive,
makes you angry as a caregiver? Why; (4) What are you writing sessions have been associated with larger effect sizes
afraid of as a caregiver? Why; (5) What frustrates you as a in the oncology setting [18].
caregiver? Why?”.
Outcomes
Resilience Training
The primary outcome was change in caregiver burden on the
While at home, participants were asked to write three things Zari Burden Interview-12 (ZBI-12) from baseline to 4- and
that went well, and what was their role in making them hap- 8-weeks post baseline. The short form of the ZBI-12 is a
pen every 3 days for 2 weeks. Participants were asked to validated short 12-item questionnaire measuring perceived
complete this exercise at the end of their day before going burden associated with caregiving [19]. Each item is rated
to sleep, writing anything positive, small or large in impor- from 0 (never) to 4 (nearly always) with a range of 0–48.
tance. The exact prompt was: “Write down three things that Scores < 10 indicate no-to-mild burden, 10–20 indicate mild-
went well today. Give each event a title (e.g., “friend took to-moderate burden, and scores > 20 indicate high burden.
me out to dinner”). Write down exactly what happened in Secondary outcomes were change in distress thermometer
as much detail as possible. Include how this event made (DT), health related quality of life (HRQOL) by one-item
you feel at the time and how this event made you feel later visual analog scale (VAS), and caregiver captivity index
(including now, as you remember it). Explain what made (CCI). These were collected from the subject via phone,
them good and what was your role in bringing them about. mail or email. The DT is single-item self-report measure
If you find yourself focusing on negative feelings, refocus of distress presented as a 10-point scale in a thermometer
your mind on the good event and the positive feelings that format and asks to rate how distressed one feels in the past
came with it.” week from 0 (not distressed) to 10 (extremely distressed).
It is a brief valid instrument originally used to assess the
Control severity of psychosocial distress in patients with serious ill-
nesses that [17] has been validated among family members
Control subjects did not receive prompts. Participants were [19]. The VAS is a validated measure of HRQOL from 0 to
asked to complete baseline and outcome assessments with 100 where 0 represents the worst health imaginable and 100
no intervention. represents the best health imaginable. The participant marks

Fig. 1  Trial design

13
3628 Digestive Diseases and Sciences (2023) 68:3625–3633

an ‘X’ over the number that matches where they think their Secondary Analysis
health is that day. The CCI is three-item self-report meas-
ure of how much a caregiver feels like a captive or being We conducted multivariable adjusted analyses of the change
an unwilling or involuntary incumbent of a caregiver role in each outcome using linear mixed models accounting for
[18]. It is scored 0–9 with higher scores being worse. We baseline values for each outcome and the factors felt to
collected outcome measures at 4 and 8 weeks to evaluate if impact caregiver burden a priori (age, sex, college educa-
distance from completion of intervention to data collection tion of caregiver, and patient paracentesis in prior 90 days).
has a different effect on caregiver burden and distress. In As this is a randomized trial, statistical testing of baseline
previous literature, participants experienced mild short-term differences was not performed. All analyses were conducted
increases in stress and depression immediately following the using adjusted linear mixed models as above.
writing intervention, but longer-term improvements in these
domains [14].
Results
Baseline Data Study Population
Baseline data collected about the patients with cirrhosis We initially approached 167 caregivers. A total of 22 met
for whom the caregivers provide care included cirrhosis exclusion criteria and 58 declined to participate for a study
etiology, complications, medications, recent procedures, cooperation rate of 60% among eligible participants. A total
hospitalizations and emergency department visits in the of 87 were randomized; 31 in the written emotional dis-
past 90 days, transplant status, and patient daily needs with closure arm (18 of whom completed the study), 28 in the
respect to activities of daily living (ADL). Baseline data resilience training arm (18 of whom completed the study),
collected about the caregivers included age, sex, race, educa- and 28 caregivers in the control arm (23 of whom completed
tion level, marital status, employment, caregiving duration the study). The most common reasons for dropout were loss
and caregiver duties, hospitalizations in the past 90 days or to follow-up and death/transplantation of the patient. The
within the past year, and any medical or physical problems study population flowchart is shown in Fig. 2. Baseline car-
that require assistance from others. egiver and patient characteristics are presented in Table 1
with observed baseline differences in age, sex, caregiver
Statistical Approach education, and patient receipt of paracentesis. The age of
the caregivers ranged from 45 to 67 years. There were more
Sample Size women, most often being the spouse to a patient. The base-
line measures for the caregivers indicated high levels of bur-
In a previous uncontrolled study of mindfulness for caregiv- den with a median ZBI-12 ≥ 14, DT ≥ 4, CCI ≥ 1.5, and fair
ers of patients with HE by Bajaj et al. [6], the response to HRQOL with median VAS scores of 75–85 across arms.
the ZBI-12 during mindfulness practice was a very large Most of the patients had alcohol and nonalcoholic fatty
7.4 ± 5.0 point improvement. Given this magnitude of dif- liver disease (NAFLD) related cirrhosis. Ascites and HE
ference, assuming an alpha of 0.05, using a t-test for power were common cirrhosis decompensations. Supplementary
analyses, we determine that to achieve 80% power that we Tables 1 and 2 provide the details for the cohorts enrolled
will need eight participants per arm to be able to reject the from each center. In Supplementary Table 3 we provide
null hypothesis that the mean improvements for the experi- details for the cohort stratified by the presence of hepato-
mental and control groups are equal. Assuming a risk of cellular carcinoma.
dropout and a lower magnitude of benefit in a controlled
trial, we aimed to enroll at least 20 participants per arm. Outcomes

In the written emotional disclosure, completion of the

Primary Analysis prompts was 90% and was 74% in the resilience training. The
unadjusted outcomes by study arm are presented in Table 2.
We sought to determine differences in the change for each No significant differences in primary outcomes were noted
outcome using a modified intention-to-treat design (all par- in the primary analyses. The written emotional disclosure
ticipants who completed outcome assessments). Patients study arm had the lowest increase in caregiver burden com-
completing week 4 but not week 8 were only analyzed for pared to resilience training and control arms. At week 4, the
week 4 outcomes. To account for variable baseline meas- caregiver burden increased in all arms; ZBI for the writ-
ures, we utilized ANCOVA testing. ten emotional disclosure, resilience training, and control

13
Digestive Diseases and Sciences (2023) 68:3625–3633 3629

Fig. 2  Study population (CONSORT) flow chart

increased by 0.6 ± 5.7, 2.5 ± 5.2, and 2.9 ± 6.1 respectively. significant changes for the DT and VAS measures were
This change was − 1.0 ± 8.9, 2.8 ± 6.1, and 1.5 ± 7.4 for each noted by study arm. At week 8, there were no significant
arm at week 8, respectively. Caregiver distress as measured differences in study outcomes by arm.
by DT increased in all arms, however by week 8 the small-
est increase was in the written emotional disclosure arm.
For the HRQOL VAS and the CCI, both intervention study Discussion
arms reported less of a decrease of burden compared to con-
trol and the resilience training arm reported a small average Cirrhosis and its complications require considerable car-
decrease in CCI by week 8. egiver involvement, which take a substantial toll on the
wellbeing of the caregivers [2, 4, 6, 18]. Few interventions
Adjusted Outcomes have been studied to reduce caregiver burden and improve
quality of life in the setting of cirrhosis. Given these knowl-
Analyses adjusted for factors unbalanced at baseline (age, edge gaps, we conducted the CAREGIVER trial to exam-
sex, caregiver education, and patient paracentesis in the ine two promising, low-cost and pragmatic approaches that
prior 90 days) are in Table 3. At week 4, compared to con- utilize journaling-based mindfulness activities that have
trol, the written emotional disclosure arm was associated been shown to reduce stress and burden in other settings.
with reduced ZBI scores (− 4.21 ± 2.03, p = 0.04) whereas We found that at 4 weeks written emotional disclosure was
resilience training was not (− 0.44 ± 1.88, p = 0.82). Resil- associated with reduced caregiver burden whereas resilience
ience training was associated with lower CCI (− 1.36 ± 0.67, training improved subjective feelings of captivity. We also
p = 0.04), but written emotional disclosure was not. No found that a 4-week intervention did not lead to sustained

13
3630 Digestive Diseases and Sciences (2023) 68:3625–3633

Table 1  Caregiver and patient Caregiver characteristics N = 87 Control (N = 28) Written emotional Resilience
characteristics disclosure (N = 31) training
(N = 28)

Age (median, IQR) 60 [55, 67] 63 [54, 66] 57.5[46, 66]

Male 13 (46.4%) 8 (25.8%) 5 (17.9%)
College graduate 13 (46.4%) 6 (19.4%) 9 (32.1%)
Married 20 (71.4%) 29 (93.5%) 20 (71.4%)
Employment fulltime, part-time 15 (53.6%) 11 (35.5%) 11 (39.6%)
Years as caregiver (median, IQR) 3.085 (2, 6) 3.2 (1, 8) 3.5 (1, 5)
Relationship to patient
Spouse 17 (60.7%) 23 (74.2%) 10 (35.7%)
Partner 2 (7.1%) 2 (6.5%) 5 (17.9%)
Parent 3 (10.7%) 2 (6.5%) 5 (17.9%)
Child 4 (14.3%) 2 (6.5%) 4 (14.3%)
Zarit Burden Interview-12 score (median, IQR) 14 [8, 19] 17 [10, 23] 14.5 [10, 25]
Distress thermometer (median, IQR) 4 [2, 5] 5 [2, 7] 5 [2, 7]
Visual analogue scale score of health-related 85 [70, 90] 75 [60, 80] 75 [56, 80]
quality of life (median, IQR)
Caregiver captivity index (median, IQR) 2 [1, 4] 2 [0, 5] 3 [2, 6]
Patient characteristics
Etiology of Cirrhosis
Alcohol 12 (42.9%) 11(35.5%) 14(50%)
NAFLD 13 (46.4%) 13(41.9%) 7(25%)
Hepatitis C 1 (3.6%) 4(12.9%) 3(10.7%)
Hepatic Encephalopathy 17 (60.7%) 24(77.4%) 17(60.7%)
Ascites 19 (67.9%) 22 (71%) 18 (64.3%)
Paracentesis in prior 90 days 7 (25%) 11 (35.5%) 13 (46.4%)
History of Hepatocellular Carcinoma 3 (10.7%) 5 (16.1%) 4 (14.3%)
Listed for Transplant 5 (17.9%) 7 (22.6%) 5 (17.9%)
Unable to operate a ­Vehicleb 9 (32.1%) 9 (29%) 11 (39.3%)
Requiring Help with ≥ 1 ADL 11 (39.3%) 14 (45.2%) 16 (57.1%)
Hospitalized, prior 90 d­ aysa 15 (53.6%) 16 (51.6%) 12 (42.9%)
Visit to ER, prior 90 ­daysa 15 (53.6%) 17 (54.8%) 14 (50%)

IQR interquartile range (listed as quartile 1–quartile 3); NAFLD nonalcoholic fatty liver disease; ADL
activities of daily living; ER emergency room
a
In the specified number days prior to enrollment
b
Newly unable to operate a motor vehicle within the 6 months prior to enrollment

improvements in caregiver burden, feelings of captivity, or intervention. This may mean that the intervention must be
improve quality of life after four additional weeks. continuous to be effective. The intervention was not sig-
nificantly associated with measures of HRQOL, distress,
Written Emotional Disclosure May Be Effective or captivity, however it is not clear as to whether a longer
intervention would be required or whether reductions
In this controlled study enrolling caregivers, we found that in ZBI do not correlate with the other measures. Fewer
after adjusting for imbalanced factors, written emotional caregivers dropped out of the emotional disclosure arm
disclosure was associated with a significant reduction in than resilience training, which may indicate a preference
caregiver burden. Bajaj et al. [6] studied a 4-week mind- for this approach. While resilience training was associ-
fulness intervention and we sought to extend this with ated with improved feelings of captivity at 4 weeks (but
4- and 8-week outcomes. The effect—a 4.2 point reduc- not 8 weeks), the associations with worsened caregiver
tion in the ZBI—approximated that which was observed burden by 8 weeks suggest meditative journal techniques
with meditation in Bajaj et al. [6]. However, this effect alone may not reduce feelings of caregiver burden and
was attenuated at week 8, 4 weeks after discontinuing the stress. While the improvement in the CCI at 4-weeks with

13
Digestive Diseases and Sciences (2023) 68:3625–3633 3631

Table 2  Unadjusted study outcomes by arm

Change from baseline to week 4 (mean ± standard deviation) Change from baseline to week 8 (mean ± standard devia-
tion)
Measure Control Written emotional Resilience p-value(s) Control Written emotional Resilience p-value(s)
(n = 28) disclosure (WED) training (RT) 1: Control (n = 23) disclosure (WED) training 1: Control
(n = 31) (n = 28) vs WED (n = 18) (RT) vs WED
2: Control (n = 18) 2: Control
vs RT vs RT

Zarit Caregiver 2.9 ± 6.1 0.6 ± 5.7 2.5 ± 5.2 1: 0.42 1.5 ± 7.4 − 1.0 ± 8.9 2.8 ± 6.1 1: 0.73
Burden Index 2: 0.99 2: 0.70
Higher values rep-
resent worsening
Distress Ther- 1.2 ± 2.7 0.7 ± 2.0 1.2 ± 2.6 1: 0.96 0.6 ± 2.7 0.7 ± 3.1 1.2 ± 2.5 1: 0.85
mometer 2: 0.95 2: 0.48
Higher values rep-
resent worsening
Visual Analog − 4.8 ± 16.8 − 4.05 ± 19.9 − 6.29 ± 11.9 1: 0.99 − 8.2 ± 23.1 − 1.3 ± 14.6 − 4.1 ± 18.9 1: 0.87
Scale of Health- 2: 0.85 2: 0.98
Related Quality
of Life
Lower values rep-
resent worsening
Caregiver captivity 1.13 ± 2.13 0.58 ± 1.07 − 0.11 ± 2.11 1: 0.65 0.9 ± 1.94 0.29 ± 2.3 − 0.28 ± 1.5 1: 0.71
index 2: 0.12 2: 0.22
Higher values rep-
resent worsening

The p-values are derived from ANCOVA testing

Table 3  Adjusted study outcomes by arm

Zarit Burden Index Distress thermometer Visual analog scale of Caregiver captivity index
health-related quality of life
Mean, SE Mean, SE Mean, SE Mean, SE

Change from baseline to week 4

Control Ref Ref Ref Ref
Written emotional disclosure − 4.21, 2.03 (p = 0.04) − 0.08, 0.89 (p = 0.93) 1.86, 5.99 (p = 0.76) − 0.71, 0.72 (p = 0.32)
Resilience training − 0.44, 1.88 (p = 0.82) 0.03, 0.82 (p = 0.97) − 3.05, 5.74 (p = 0.60) − 1.36, 0.67 (p = 0.04)
Change from baseline to week 8
Control Ref Ref Ref Ref
Written emotional disclosure − 1.13, 2.60 (p = 0.67) 0.89, 0.93 (p = 0.34) 4.15, 6.73 (p = 0.54) − 0.17, 0.63 (p = 0.79)
Resilience training 2.67, 2.54 (p = 0.30) 1.15, 0.91 (p = 0.21) 2.76, 6.58 (p = 0.68) − 0.83, 0.63 (p = 0.19)

Control is the reference group for each model. Each cell represents the results of a linear mixed regression model where the control subjects are
the reference value and the estimates for each intervention’s effect on the outcomes are adjusted for the baseline value of the outcome, age, sex,
college education of caregiver, and patient paracentesis in the prior 90 days. SE standard error

resilience training may hint at some short-term benefits, Role of Mindfulness Methods in Alleviating
the lack of a parallel improvement in the other indices and Caregiver Burden
the primary outcome renders this effect uncertain. Future
interventions should test multicomponent approaches Implicit in any meditative or resilience-based interven-
among which meditative journaling could be one of sev- tion is the concept that the perceived burden is modifiable
eral techniques [20]. Multicomponent interventions could through mental exercises. Our data demonstrates that writ-
include mindfulness, individual or group psychotherapy, ten emotional disclosure may be helpful, but not for all and
exercise programs, and nursing outreach [20]. only while the intervention is ongoing. As such, it may

13
3632 Digestive Diseases and Sciences (2023) 68:3625–3633

complement other interventions. As some caregivers had while worsening of the patient’s status resulted in dropout,
improved burden, it may be reasonable to offer mindfulness improvements were not measured and if they occurred could
methods in a time-limited trial followed by short-term reas- have improved caregiver burden. Fifth, in a modified inten-
sessment prior to initiating interventions with greater costs. tion-to-treat analysis where the outcome is patient reported,
The most effective intervention to alleviate caregiver bur- we can only include those with outcome data. Dropout prior
den is likely control of the patient’s condition. In a quali- to outcome assessment results in some baseline character-
tative assessment conducted utilizing the subjects’ written istics becoming unbalanced. Finally, fidelity to the assigned
responses we found that caregivers feel isolated and finan- interventions was very high but could be variable in larger
cially harmed while struggling to manage unpredictable studies permitting per-protocol analyses.
encephalopathy and the challenging nutritional restrictions
posed by ascites while simultaneously feeling poorly served
by their clinicians [18]. Our primary conclusion, therefore,
mindfulness methods may be helpful but cannot replace Conclusions
efforts to optimize disease control and efforts to reduce
financial impact of caregiving associated with cirrhosis [21]. CAREGIVER was a low-cost, home-based intervention
Support for caregivers in the home and resources to provide aimed to reduce the perceived stress of caregiving for
respite are promising areas for future investigation. Payor patients with cirrhosis. Mindfulness practices may be effec-
supported trials of financial supports, increased access to in tive in the short-term or for selected patients. Clinicians may
home care, and respite care are warranted. wish to suggest time-limited trials for meditative journaling
Psychoeducational and psychosocial support interven- using our emotional disclosure protocol for caregivers who
tions often result in mild or modest improvements in mood are interested. Additional attention is needed to address the
or burden but may not be sufficient to procedure dura- multifactorial unmet needs of caregivers for patients with
ble improvements [11]. Interventions which improve the cirrhosis.
patient’s condition are most likely to be effective. Just as Supplementary Information The online version contains supplemen-
donepezil for Alzheimer’s dementia improves caregiver bur- tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 10620-0​ 23-0​ 8042-9.
den relating to that condition [22], it is likely that efforts to
Author's contribution EBT is the guarantor of this article. EBT: con-
optimize and improve therapies for HE, ascites, or frailty
cept, analysis, data acquisition, and writing. SN, MS, SL, XC, and
may prove more effective than strictly caregiver-focused ZS: analysis, data acquisition, and critical revision of the manuscript.
interventions. Beyond that, rather than asking caregivers
to add interventions (including mindfulness methods) to Funding This study is dedicated to and funded in part by Gloria Wend-
ling and the research fund established in her memory for Elliot Tapper.
address burden, removing stressors may be more efficacious.
Elliot Tapper receives funding from the National Institutes of Health
through the NIDDK (1K23DK117055). The funders did not conduct,
Contextual Factors design, or evaluate the results of this trial.

These data must be interpreted in the context of study Declarations

design. First, caregivers who volunteered for these interven-
Conflict of interest Elliot Tapper has served as a consultant to Novo
tions may have been the most distressed or in need of help, Nordisk, Axcella, Norvartis, Kaleido, and Allergan, has served on ad-
resulting in burden resistant to change with a low touch, visory boards for Mallinckrodt, Rebiotix, and Bausch Health, and has
meditative intervention. Second, the COVID pandemic received unrestricted research grants from Gilead. Marina Serper has
served as a consultant to Gilead, Inc. and has received research funding
may have worsened caregiver burden resulting in relative
from Grifols, SA. No outside interests were involved with the study. No
ineffectiveness of interventions. The COVID-19 pandemic other author has conflicts of interest to disclose.
may have affected the ways in which patients, caregivers,
and clinicians interact. Third, although the estimated effect
of the intervention was confirmed, it was derived from an
uncontrolled study and with a more comprehensive interven- References
tion potentially resulting in an underestimate of the needed
1. Tapper EB, Parikh ND. Mortality due to cirrhosis and liver can-
sample size. As the average effect size is likely smaller
cer in the United States, 1999–2016: observational study. BMJ.
than anticipated given the heterogeneity of treatment effect, 2018;362:k2817.
large sample sizes are needed and future trials should be 2. Bajaj JS, Wade JB, Gibson DP, et al. The multi-dimensional bur-
conducted using a pragmatic trial framework, recruiting den of cirrhosis and hepatic encephalopathy on patients and car-
egivers. Am J Gastroenterol. 2011;106:1646–1653. https://​doi.​
at diverse institutions [23]. Fourth, we discontinued the
org/​10.​1038/​ajg.​2011.​157. (PubMed PMID: 21556040).
study if the patient was transplanted or died but future stud-
ies could continue beyond these landmarks. Conversely,

13
Digestive Diseases and Sciences (2023) 68:3625–3633 3633

3. Thomson MJ, Lok AS, Tapper EB. Optimizing medication man- modulates latent Epstein-Barr virus antibody titers. J Consult Clin
agement for patients with cirrhosis: evidence-based strategies and Psychol. 1994;62:130–140 (Epub 1994/02/01 PubMed PMID:
their outcomes. Liver Int. 2018;38:1882–1890. 8034815).
4. Ufere NN, Donlan J, Indriolo T, et al. Burdensome transitions of 16. Greenberg MA, Wortman CB, Stone AA. Emotional expression
care for patients with end-stage liver disease and their caregivers. and physical health: revising traumatic memories or fostering
Dig Dis Sci. 2020. https://​doi.​org/​10.​1007/​s10620-​020-​06617-4. self-regulation? J Pers Soc Psychol. 1996;71:588–602. https://​
(PubMed PMID: 32964286). doi.​org/​10.​1037//​0022-​3514.​71.3.​588. (Epub 1996/09/01 Pub-
5. Rakoski MO, McCammon RJ, Piette JD, et al. Burden of cirrhosis Med PMID: 8831163).
on older Americans and their families: analysis of the health and 17. Seligman ME, Steen TA, Park N, Peterson C. Positive psychology
retirement study. Hepatology. 2012;55:184–191. progress: empirical validation of interventions. Am Psychologist.
6. Bajaj JS, Ellwood M, Ainger T, et al. Mindfulness-based stress 2005;60:410.
reduction therapy improves patient and caregiver-reported out- 18. Saleh ZM, Salim NE, Nikirk S, Serper M, Tapper EB. The emo-
comes in cirrhosis. Clin Transl Gastroenterol. 2017;8:e108. tional burden of caregiving for patients with cirrhosis. Hepatol
7. Hareendran A, Devadas K, Sreesh S, et al. Quality of life, car- Commun. 2022;6:2827–2835.
egiver burden and mental health disorders in primary caregivers 19. Bédard M, Molloy DW, Squire L, Dubois S, Lever JA, O’Donnell
of patients with Cirrhosis. Liver Int. 2020;40:2939–2949. M. The Zarit Burden interview: a new short version and screening
8. Tapper EB, Kenney B, Nikirk S, Levine DA, Waljee AK. Animal version. Gerontologist. 2001;41:652–657.
naming test is associated with poor patient-reported outcomes and 20. Williams F, Moghaddam N, Ramsden S, De Boos D. Interventions
frailty in people with and without cirrhosis: a prospective cohort for reducing levels of burden amongst informal carers of persons
study. Clin Transl Gastroenterol. 2022;13:e00447. https://d​ oi.o​ rg/​ with dementia in the community: a systematic review and meta-
10.​14309/​ctg.​00000​00000​000447. analysis of randomised controlled trials. Aging Mental Health.
9. Hou RJ, Wong SYS, Yip BHK, et al. The effects of mindfulness- 2019;23:1629–1642.
based stress reduction program on the mental health of family car- 21. Ufere NN, Satapathy N, Philpotts L, Lai JC, Serper M. Financial
egivers: a randomized controlled trial. Psychother Psychosomat. burden in adults with chronic liver disease: a scoping review. Liver
2014;83:45–53. Transpl. 2022;28:1920–1935. https://​doi.​org/​10.​1002/​lt.​26514.
10. Whitebird RR, Kreitzer M, Crain AL, Lewis BA, Hanson 22. Lingler JH, Martire LM, Schulz R. Caregiver-specific outcomes
LR, Enstad CJ. Mindfulness-based stress reduction for fam- in antidementia clinical drug trials: a systematic review and meta-
ily caregivers: a randomized controlled trial. Gerontologist. analysis. J Am Geriatrics Soc. 2005;53:983–990.
2013;53:676–686. 23. Tapper EB, Serper M, Goldberg D. Implementing pragmatic clini-
11. Adelman RD, Tmanova LL, Delgado D, Dion S, Lachs MS. Car- cal trials in hepatology. Hepatology. 2023;10:1097.
egiver burden: a clinical review. JAMA. 2014;311:1052–1060.
12. Pennebaker JW. The effects of traumatic disclosure on physi- Publisher's Note Springer Nature remains neutral with regard to
cal and mental health: the values of writing and talking about jurisdictional claims in published maps and institutional affiliations.
upsetting events. Int J Emerg Ment Health. 1999;1:9–18 (Epub
2001/03/03 PubMed PMID: 11227757). Springer Nature or its licensor (e.g. a society or other partner) holds
13. Pennebaker JW, Seagal JD. Forming a story: the health benefits of exclusive rights to this article under a publishing agreement with the
narrative. J Clin Psychol. 1999;55:1243–1254 (PubMed PMID: author(s) or other rightsholder(s); author self-archiving of the accepted
11045774). manuscript version of this article is solely governed by the terms of
14. Petrie KJ, Booth RJ, Pennebaker JW. The immunological effects such publishing agreement and applicable law.
of thought suppression. J Pers Soc Psychol. 1998;75:1264–1272
(Epub 1998/12/29 PubMed PMID: 9866186).
15. Esterling BA, Antoni MH, Fletcher MA, Margulies S, Schnei-
derman N. Emotional disclosure through writing or speaking

13