Journal of Consulting and Clinical Psychology © 2015 American Psychological Association

2015, Vol. 83, No. 4, 760 –772 0022-006X/15/$12.00 http://dx.doi.org/10.1037/ccp0000028

Cognitive–Behavioral Therapy (CBT) Versus Acceptance and Commitment

Therapy (ACT) for Dementia Family Caregivers With Significant
Depressive Symptoms: Results of a Randomized Clinical Trial
Andrés Losada María Márquez-González
Universidad Rey Juan Carlos Universidad Autónoma de Madrid

Rosa Romero-Moreno Brent T. Mausbach

Universidad Rey Juan Carlos University of California, San Diego
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Javier López Virginia Fernández-Fernández and

This document is copyrighted by the American Psychological Association or one of its allied publishers.

Universidad CEU San Pablo Celia Nogales-González

Universidad Rey Juan Carlos
Objective: The differential efficacy of acceptance and commitment therapy (ACT) and cognitive– behavioral
therapy (CBT) for dementia family caregivers’ is analyzed through a randomized controlled trial. Method:
Participants were 135 caregivers with high depressive symptomatology who were randomly allocated to the
intervention conditions or a control group (CG). Pre-, postintervention, and follow-up measurements assessed
depressive symptomatology, anxiety, leisure, dysfunctional thoughts, and experiential avoidance. Results:
Depression: Significant effects of interventions compared with CG were found for CBT (p ⬍ .001, d ⫽ 0.98,
number needed to treat [NNT] ⫽ 3.61) and ACT (p ⬍ .001, d ⫽ 1.17, NNT ⫽ 3.53) at postintervention, but
were maintained only at follow-up for CBT (p ⫽ .02, d ⫽ 0.74, NNT ⫽ 9.71). Clinically significant change
was observed in 26.7% participants in CBT, 24.2% in ACT, and 0% in CG. At follow-up, 10.53% in CBT
and 4% in ACT were recovered (0% CG). Anxiety: At postintervention, ACT participants showed lower
anxiety than CBT participants (p ⬍ .05, d ⫽ 0.50) and CG participants (p ⬍ .01, d ⫽ 0.79, NNT ⫽ 3.86),
with no effects at follow-up. At postintervention, 23.33% in CBT, 36.36% in ACT, and 6.45% in CG showed
clinically significant change. At follow-up, 26.32% in CBT, 36% in ACT, and 13.64% in CG were recovered.
Significant changes at postintervention were found in leisure and dysfunctional thoughts in both ACT and
CBT, with changes in experiential avoidance only for ACT. Conclusion: Similar results were obtained for
ACT and CBT. ACT seems to be a viable and effective treatment for dementia caregivers.

What is the public health significance of this article?

One of every four dementia caregivers treated with acceptance and commitment therapy (ACT) or
cognitive– behavioral therapy (CBT) shows a clinically significant reduction in depressive and
anxiety symptoms at postintervention as compared with caregivers in the control group.
The findings suggest that adding booster sessions may consolidate the outcomes of brief ACT and
CBT interventions for dementia caregivers.
ACT may be an appropriate alternative to CBT for helping dementia caregivers with high depressive
symptoms.

Keywords: acceptance and commitment therapy, caregivers, cognitive– behavioral therapy, dementia,
intervention

This article was published Online First June 15, 2015. also the following centers for collaborating with us in the project: Fun-
Andrés Losada, Psychology Department, Universidad Rey Juan Carlos; dación Cien, Fundación María Wolff, Centros de día Vitalia, Centro de
María Márquez-González, Department of Biological and Clinical Psychology, Salud General Ricardos, Centro de Salud García Noblejas, Centro de Salud
Universidad Autónoma de Madrid; Rosa Romero-Moreno, Psychology De- Benita de Ávila, Centro de Salud Vicente Muzas, Centro Reina Sofía de
partment, Universidad Rey Juan Carlos; Brent T. Mausbach, Department of Cruz Roja, Unidad de Memoria de Cantoblanco, Servicio de Neurología
Psychiatry, University of California, San Diego; Javier López, Department of del Hospital La Paz, Asociación de Familiares de Alzheimer de Alcorcón,
Psychology, Universidad CEU San Pablo; Virginia Fernández-Fernández and Instituto de Familia de la Universidad CEU San Pablo, Centro de Psi-
Celia Nogales-González, Psychology Department, Universidad Rey Juan Car- cología Aplicada de la Universidad Autónoma de Madrid, and Servicios
los. Sociales de Getafe.
The preparation of this article was supported in part by Grant PSI2009- Correspondence concerning this article should be addressed to Andrés
08132 from the Spanish Ministry of Science and Innovation and Grant Losada Baltar, Facultad de Ciencias de la Salud, Universidad Rey Juan
PSI2012-31293 from the Spanish Ministry of Economy and Competitive- Carlos, Departamental II. Avda. de Atenas s/n 28922, Alcorcón, Spain.
ness. We thank all the caregivers for their participation in the study and E-mail: [email protected]

760
 CBT VS. ACT FOR CAREGIVERS 761

The analysis of the efficacy of psychological interventions with caregivers adapt to many adverse external and internal (e.g., neg-
dementia family caregivers presents a great interest considering ative emotions) experiences.
the chronic stressful nature of their situation. Caring for relatives Acceptance has been defined as the active tendency to remain in
with dementia is associated with significant negative mental and contact with private experiences (emotions, thoughts, memories, or
physical health consequences for caregivers (Pinquart & Sörensen, behavioral predispositions), without altering their form, frequency,
2003). For example, Cuijpers (2005) found, through a systematic or the contexts that cause them (Hayes et al., 2004). The opposite
review of the literature, a 22.3% prevalence of depressive disor- side of acceptance, experiential avoidance, is “implicated in a
ders in dementia family caregivers. The proportion of caregivers’ wide range of clinical problems and disorders” (Hayes et al., 2004,
identified as at risk for depression using the Center for Epidemi- p. 254). Acceptance and commitment therapy (ACT; Hayes, Stro-
ological Studies Depression scale (CES–D) ranges from 27.9% to sahl, & Wilson, 1999, 2011) has been developed as an intervention
55.0% (Schulz, O=Brien, Bookwala, & Fleissner, 1995). In this for addressing psychological problems in which experiential
line, Cooper, Balamurali, Selwood, and Livingston (2007) con- avoidance plays a central role. ACT appears to be particularly
ducted a systematic review of the prevalence of anxiety; they useful to situations that involve aspects that may be unchangeable,
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

found that 25% of dementia caregivers experienced clinically including chronic stress situations such as chronic pain (Wetherell,
This document is copyrighted by the American Psychological Association or one of its allied publishers.

significant anxiety. These data are especially relevant considering Afari, et al., 2011), parenting of children with autism (Blackledge
the number of people who are taking care of a relative with & Hayes, 2006), or coping with life-threatening illness (Burke et
dementia (Alzheimer’s Association, 2013). Furthermore, it is an- al., 2014). Research also suggests that ACT is effective for reduc-
ticipated that there will be a dramatic increase in the number of ing depressive, anxious, and pain symptomatology in older adults
caregivers in the near future given the aging of the population and (Karlin et al., 2013; McCracken & Jones, 2012; Wetherell, Liu, et
the increase in life expectancy (Alzheimer’s Association, 2013; al., 2011).
Gallagher-Thompson et al., 2012). In the past decades, research in clinical psychology has shown
Various psychological therapies for caregivers are considered a great interest in the analysis of the differential efficacy of
empirically validated, and depression is the most common out- psychological therapies. In recent years, some studies have been
come variable of these interventions (Gallagher-Thompson & conducted to analyze the differential efficacy of CBT and ACT
Coon, 2007). Specifically, the effect of cognitive– behavioral ther- (Öst, 2008), finding empirical support for the efficacy of both
apies (CBTs) for depression seems to be stronger than other therapies. However, the number of studies comparing both thera-
interventions with this population (Pinquart & Sörensen, 2006), pies is still sparse, especially with regard to their differential
with individual intervention showing a stronger effect size com- long-term effects. Some studies have shown results that suggest a
pared with group-based interventions (Knight, Lutzky, & similar efficacy of both interventions in follow-up outcomes in
Macofsky-Urban, 1993), especially with caregivers with signifi- persons with mixed anxiety disorders (Arch et al., 2012) and
cant levels of depression (Gallagher-Thompson & Coon, 2007). tinnitus (Hesser et al., 2012). There are also studies supporting a
CBT interventions with caregivers are usually aimed at training better maintenance of treatment gains for CBT when applied to
caregivers in strategies for changing dysfunctional thoughts or depressed and anxious patients (Forman, Shaw, et al., 2012); still
behaviors associated with a maladaptive coping style, replacing others have found the opposite results, that is, better long-term
them with thoughts that promote pleasant activities, seeking help effects of ACT with patients with addictions (González-Menéndez,
from others, or modifying antecedents and consequences of be- Fernández, Rodríguez, & Villagrá, 2014), depression (Zettle &
havioral and psychological symptoms of dementia (Gallagher- Rains, 1989), or mood, anxiety, and interpersonal problems (Lap-
Thompson et al., 2003; Losada, Márquez-González, & Romero- palainen et al., 2007).
Moreno, 2011). Consistently, the main mechanisms of action of To our knowledge, no study has been done so far to test the
CBT for attenuating negative emotions associated with caregiving efficacy of ACT for distressed dementia caregivers, except for a
involve changing maladaptive cognitions and increasing pleasant pilot study done by our research team that used an ACT interven-
activities (Losada et al., 2011). tion in group format (Márquez-González, Romero-Moreno, &
Regarding anxiety, a recent review and meta-analysis of inter- Losada, 2010). The aim of the present study was to analyze,
ventions in dementia caregiving have highlighted the need for through a randomized controlled trial, the differential efficacy of
randomized controlled trial studies analyzing intervention effects two psychological therapies for reducing depression and anxiety
on anxiety as the dependent outcome for interventions with care- symptoms and for changing specific coping variables (frequency
givers (Li, Cooper, Austin, & Livingston, 2013). In recent years, of pleasant events, dysfunctional thoughts, and experiential avoid-
research has shown the relevance of accepting adverse internal ance) in dementia family caregivers. The first therapy, CBT, was
experiences (thoughts and emotions) in the process of adaptation designed to change maladaptive beliefs and increase pleasant ac-
to the caregiving role (Losada, Márquez-González, Romero- tivities, and was adapted from our previous work (Márquez-
Moreno, & López, 2014; Spira et al., 2007). Caring for a relative González, Losada, Izal, Pérez-Rojo, & Montorio, 2007) to be
with dementia is associated with negative emotions, difficulties, delivered in an individual (face-to-face) format. The second ther-
demands, and losses that are, to a large degree, unchangeable. apy, ACT, was aimed at promoting the acceptance of aversive
Taking care of a relative with dementia implies witnessing his or internal events related to caregiving and caregivers= commitment
her increasing limitations, as well as his or her intimate needs and and action toward their values. Our hypotheses were as follows:
weaknesses. In addition, providing care requires time and physical
and mental resources, usually to the detriment of time for oneself, Hypothesis 1: Both interventions will be superior to a minimal
for self-care, and for personal needs. In sum, these characteristics support control group in terms of reducing caregivers’ depres-
make the caregiving situation a vital context that demands that sive and anxious symptomatology.
 762 LOSADA ET AL.

Hypothesis 2: Compared with participants randomized to the a CES–D score lower than 16), and an additional 42 chose not to
control condition, those in the CBT intervention will show a participate. The remaining 135 participants were randomized to
significant decrease in dysfunctional thoughts and significant CBT (n ⫽ 42), ACT (n ⫽ 45), or minimal support control group
increases in pleasant activities. (CG; n ⫽ 48) using computer-generated random numbers. All
interventions took place in the collaborative centers. Participants
Hypothesis 3: Compared with participants in the control con- were assessed at pretreatment, posttreatment, and 6 months after
dition, those in the ACT intervention will show significant posttreatment. All the assessments were done by psychologists
reduction in experiential avoidance. trained in the assessment protocol who were blind to treatment
conditions and to the main hypotheses of the study. Caregivers
Method gave their informed consent to participate in the study, which was
approved by the Spanish Ministry of Science and Innovation and
Participants the Ethics Committee at the Universidad Rey Juan Carlos (Ma-
drid). This study has followed Journal Article Reporting Standards
Participants were 135 dementia family caregivers recruited through
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

(APA Publications and Communications Board Working Group on

Social and Health Care Centers as well as through Internet advertise-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Journal Article Reporting Standards, 2008) and CONSORT guide-

ment in Madrid (Spain). Inclusion criteria were as follows: (a) self- lines (Moher, Schulz, & Altman, 2001) for reporting randomized
identifying as the principal person taking care of a relative diagnosed clinical trials.
with dementia, (b) devoting at least 1 hr daily to the care of the
relative, (c) having cared for at least 3 months, (d) having not partic- Measures
ipated previously in any psychotherapeutic intervention aimed at
helping the caregiver cope with caregiving, and (e) scoring at least 16 In addition to sociodemographic information (caregivers’ age,
on the CES–D scale (Radloff, 1977). Sample characteristics are gender, education, time since being a caregiver, and number of
shown in Table 1 and the flow of participants is shown in Figure 1. daily hours devoted to caregiving), the following variables were
measured.
Cognitive status. The cognitive status of the care recipient
Procedure
was measured using the Global Deterioration Scale for Assessment
The research team established contact with several Social and of Primary Degenerative Dementia (Reisberg, Ferris, de Leon, &
Health Care Centers in Madrid, described the main objectives of Crook, 1982).
the research project, and requested their collaboration in contacting Care recipient functional capacity. The Barthel Index (Ma-
potential dementia family caregivers. Contact information (name honey & Barthel, 1965) was used to measure care recipients’
and phone number) from potential participants was provided by functional status. Internal consistency (Cronbach’s alpha) for this
the collaborative centers or obtained through direct contact with study was .91.
potential participants through e-mail or phone calls. As shown in Frequency of behavioral problems. Care recipients’ behav-
Figure 1, potential participants (N ⫽ 377) were first interviewed by ioral and psychological symptoms were measured with the Dis-
phone to screen for fulfillment of the inclusion criteria. If the ruptive Behaviors subscale of the Revised Memory and Behavior
inclusion criteria were met, potential participants were cited for a Problems Checklist (Teri et al., 1992). Internal consistency (Cron-
pretreatment assessment in the collaborative center through which bach’s alpha) was .67 in this study.
contact was made, and the time of the assessment was agreed with Primary outcomes.
the caregiver. Of the potential participants, 242 did not meet Depressive symptomatology. The CES–D (Radloff, 1977) was
inclusion criteria (e.g., not being the primary caregiver or having used. It is the most frequently used measure for assessing depres-

Table 1
Sample Characteristics by Treatment Condition

Variable CG ACT CBT F(2, 132) ␹2

Mean (SD) caregiver age (years) 62.28 (12.92) 61.69 (15.31) 61.48 (12.40) 0.04
Sex (female), n (%) 39 (81.02) 37 (82.2) 38 (90.5) 1.71
Mean (SD) time caring (years) 5.05 (3.56) 3.30 (2.02) 3.67 (2.58) 5.04ⴱ
Mean (SD) daily hours caring 15.57 (8.42) 16.24 (7.70) 13.07 (7.79) 1.89
Mean (SD) formal education (years) 15.98 (6.60) 16.36 (6.96) 17.67 (7.06) 0.73
Care recipient, n (%) 3.54
Spouse 20 (41.7) 22 (48.9) 13 (31.0)
Parent 25 (52.1) 19 (42.2) 24 (57.1)
Other relative 3 (6.2) 4 (8.9) 5 (11.9)
Care recipient dementia (Alzheimer’s), n (%) 35 (72.9) 34 (75.6) 33 (78.6) 0.39
Mean (SD) care recipient functional status 53.85 (28.36) 61.89 (21.51) 62.74 (26.37) 1.40
Mean (SD) care recipient cognitive status 5.04 (1.22) 4.95 (0.96) 4.83 (0.96) 0.43
Mean (SD) frequency of disruptive behaviors 15.27 (8.04) 17.27 (8.85) 14.86 (8.60) 1.02
Note. CG ⫽ minimal support control group; ACT ⫽ acceptance and commitment therapy; CBT ⫽ cognitive– behavioral therapy.
ⴱ
p ⬍ .01.
 CBT VS. ACT FOR CAREGIVERS 763
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Figure 1. Flow of the study. CES–D ⫽ Center for Epidemiological Studies Depression scale; CBT ⫽
cognitive– behavioral therapy; ACT ⫽ acceptance and commitment therapy.

sive symptomatology in intervention studies with caregivers (Pin- has been computed following the McGlinchey, Atkins, and Jacob-
quart & Sörensen, 2006). The scale has 20 items assessing depres- son (2002) criteria and using POMS normative data provided by
sive symptoms that caregivers might have felt over the past week the Nyenhuis, Yamamoto, Luchetta, Terrien, and Parmentier
(e.g., “I felt sad”). Item responses range from 0 (rarely or none of (1999) study. Scores equal to or higher than 13 might identify
the time) to 3 (most or all of the time). Scores equal to or higher people with clinical anxiety. Cronbach’s alpha for the scale in this
than 16 might identify people with clinical depression (Kohout, study was .92.
Berkman, Evans, & Cornoni-Huntley, 1993). Internal consistency Secondary outcomes.
(Cronbach’s alpha) for this scale in the present study was .88. Leisure. Frequency of leisure activities was measured by an
Anxiety. The Tension-Anxiety subscale from the Profile of adaptation of the Leisure Time Satisfaction scale developed by
Mood States (POMS; McNair, Lorr, & Droppleman, 1971) was Stevens et al. (2004). The scale includes six items that measure the
used. Caregivers were asked to rate the degree to which each of the extent to which caregivers had participated in leisure activities
nine symptoms (e.g., “tense”) described them during the past (e.g., quiet time by yourself, taking part in hobbies, etc.). Re-
week. Responses were provided on a Likert-type scale ranging sponses are rated on a 5-point Likert-type scale (0 ⫽ not at all; 4 ⫽
from 0 (not at all) to 4 (extremely). The cutoff score for the POMS a lot). Cronbach’s alpha for this scale in this study was .64.
 764 LOSADA ET AL.

Experiential avoidance. The Experiential Avoidance in Care- actions consistent with those values, and to analyze the barriers to
giving Questionnaire (Losada et al., 2014) was used. It is com- those actions and ways of overcoming them. A detailed description
posed of 15 items (e.g., “I have never felt bad in relation to caring of this intervention may be found in Márquez-González et al.
for my relative”) with a 5-point Likert-type scale ranging from 0 (2010).
(not at all) to 5 (a lot). Cronbach’s alpha for this scale in this study Minimal support control group (CG). Caregivers in the
was .71. minimal support CG participated in a 2-hr workshop during which
Dysfunctional thoughts about caregiving. The Dysfunctional they were provided a booklet and psychoeducation on dementia
Thoughts About Caregiving Questionnaire (Losada, Montorio, (e.g., definition of the dementia and its common symptoms, de-
Izal, & Márquez-González, 2006) was used. The questionnaire has scription of frequent behavioral and psychological symptoms of
16 items (e.g., “It is selfish for a caregiver to dedicate time to dementia). No specific therapeutic techniques were applied in this
himself/herself when a relative is frail/sick and needs care”) with group.
Likert-type answers ranging from 0 (totally disagree) to 4 (totally
agree). Cronbach’s alpha for this scale in this study was .89.
Treatment Implementation Measures
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Caregivers were asked to rate their satisfaction with the inter-

Treatments
vention, the therapists, and the content and exercises of each
CBT and ACT interventions consisted of eight weekly individ- program. Responses options ranged from 0 (not satisfied at all) to
ual sessions (total duration ⫽ 2 months) lasting about 90 min each. 10 (totally satisfied). Caregivers also were asked whether they
Therapists were six clinical psychologists trained in CBT and ACT would recommend the intervention to other caregivers.
principles and techniques. All of them received specific training in Following the recommendations of Burgio et al. (2001), we used
the intervention conditions and participated in weekly supervision measures of treatment receipt and treatment enactment. To assess
meetings with members from the research team and other thera- treatment receipt, therapists asked each participant to complete (at
pists. All therapists had master’s or doctoral training in clinical pre- and posttreatment) an 18-item questionnaire assessing knowl-
psychology, including training in CBT and ACT techniques. There edge and skills related to the content and exercises developed in
were no differences between therapists in the allocation to the each program. The response options ranged from 0 (no knowledge/
intervention conditions (␹2 ⫽ 4.73, p ⫽ .45). not skilled at all) to 10 (total knowledge/very skilled). For the
CBT. The CBT intervention was based on previous interven- assessment of treatment enactment, we followed a similar proce-
tions by Gallagher-Thompson et al. (2003) and developed accord- dure to the one used in Márquez-González et al. (2007). The
ing to a cognitive– behavioral model of caregiver distress (Losada quantity and quality of the homework assignments done by the
et al., 2006). Even though it has been successfully used in group caregivers were assessed by three trained psychologists. Quantity
format in previous studies (Márquez-González et al., 2007), an of homework was assessed using a 5-point scale ranging from 0
individual format was used in this study. This program is com- (the participant did not do any homework assignments) to 4 (all
posed of four modules: (1) cognitive restructuring: Caregivers homework assignments were done by the participant). Quality was
were trained to acknowledge, analyze, and reformulate maladap- assessed using a 5-point scale ranging from 0 (no homework was
tive thoughts into new ones that facilitate a more appropriate available) to 4 (skills have been correctly implemented).
coping with dementia caregiving (e.g., “I am the only one who
knows how to care of him” would be substituted by “Dementia
Data Analysis
care is a team work; a single person cannot assume it all”); (2)
assertive skills: asking for help; (3) relaxation; and (4) increasing Univariate and multivariate outliers and normality were deter-
pleasant activities. Further information regarding this program mined using Tabachnick and Fidell’s (2001) criteria. Differences
may be found in Márquez-González et al. (2007). in the assessed variables between completers and those who
ACT. An ACT intervention was specifically designed for dropped out of the study were analyzed through differences be-
caregivers (Márquez-González et al., 2010), departing from pre- tween means and independence (␹2) tests. Between intervention
vious works by Hayes et al. (1999) and McCurry (2006). The conditions, differences prior to the intervention were assessed
program is structured around the three basic pillars of ACT ther- through analyses of variance and independence tests (␹2).
apy: (1) acceptance of aversive internal events and the circum- Change on caregivers’ primary outcomes (depression and anx-
stances activating them, (2) choosing meaningful courses of action iety) and secondary outcomes (frequency of pleasant events, dys-
consistent with one’s personal values, and (3) action oriented functional thoughts, and caregiving experiential avoidance) was
toward those values. To increase acceptance, the therapeutic pro- measured using linear mixed models with random intercepts and
cess is as follows: (a) identification of personal patterns of expe- restricted maximum likelihood estimation. Treatment group (CBT,
riential avoidance; (b) analysis of the trap of control of aversive ACT vs. CG), measurement point (pretreatment, posttreatment,
inner events; (c) acknowledging the cost of the avoidant behaviors and follow-up), and Group ⫻ Time interaction were treated as
(creative hopelessness); and (d) learning the alternative to control fixed effects, and subjects were treated as a random effect. To
(i.e., acceptance), using metaphors, cognitive defusion, and mind- assess changes from baseline to posttreatment and to follow-up, we
fulness techniques. To facilitate meaningful courses of action carried out planned contrasts. Three effect sizes were calculated
consistent with one’s values, therapists trained caregivers to iden- for primary outcomes. Following Feingold’s (2009) formula, we
tify their values and the barriers to action oriented to them. To calculated Cohen’s d effect sizes. With the aim of assessing
foster action oriented to those values, therapists encouraged care- clinical significance, we calculated the number of participants who
givers to commit to his or her values, to program and perform needed to be treated to achieve one additional favorable outcome
 CBT VS. ACT FOR CAREGIVERS 765

(number needed to treat; NNT) with Newcombe–Wilson 95% was 107.97 ⫾ 22.39, and mean ⫾ SD at postintervention was ⫽
confidence intervals following Bender (2001) and Newcombe 124.70 ⫾ 10.20; t(30) ⫽ ⫺2.95, p ⬍ .01.
(1998) criteria. The third effect size was the reliable change index Treatment enactment. No significant differences between
(RCI), which has been calculated following Jacobson and Truax treatment conditions were found for quantity (mean ⫾ SD for
(1991). CBT ⫽ 2.29 ⫾ 1.34; mean ⫾ SD for ACT ⫽ 2.65 ⫾ 1.10), t(58) ⫽
1.00, and quality (mean ⫾ SD for CBT ⫽ 2.63 ⫾ 1.74; mean ⫾ SD
Results for ACT ⫽ 2.33 ⫾ 0.86), t(58) ⫽ ⫺0.71, of homework activities
done by participants.
Outliers and Normality
No univariate (z scores ⬎ 3.29, p ⬍ .001) or multivariate Changes in Primary Outcomes, CBT Versus CG
(Mahalanobis distance; p ⬍ .001) outliers were found. Skewness Table 2 and Figure 2 present the estimated means for treatment
and kurtosis data were within normal ranges. conditions across time for all outcomes. No significant differences
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

were found in the assessed variables at preintervention. For change

Treatment Attrition in depressive symptomatology scores, contrasts comparing the
As shown in Figure 1, 94 of the 135 participants (69.63%) CBT and control conditions indicated that participants randomized
completed the intervention and were assessed postintervention. to CBT showed significantly greater reductions from baseline to
Completers spent fewer hours caring, t(132) ⫽ ⫺2.45, p ⬍ .05, posttreatment compared with those randomized to the CG condi-
and they had lower scores on the preassessment in depression, tion, time by treatment effect estimate ⫽ ⫺8.70, t(180.46) ⫽ 3.57,
t(133) ⫽ ⫺2.37, p ⬍ .05, than those who dropped out of the study p ⬍ .001. Cohen’s d for this effect was 0.98, which indicates large
prior to postassessment. Sixty-six participants were assessed at magnitude. Significant differences were also observed for change
follow-up. The proportion of completers by treatment condition in depressive symptomatology scores from baseline to follow-up,
was similar at postintervention (␹2 ⫽ 11.02, p ⫽ .20) and time by treatment effect estimate ⫽ 6.66, t(186.76) ⫽ 2.35, p ⫽
follow-up (␹2 ⫽ 5.90, p ⫽ .82). No differences were also found in .02, with Cohen’s d of 0.74.
the proportion of completers between the ACT and CBT condi- Comparison of change in anxiety from baseline to posttreatment
tions at postintervention (␹2 ⫽ 3.93, p ⫽ .42) and follow-up (␹2 ⫽ showed no significant differences between the CBT and CG con-
1.68, p ⫽ .80). Completers were caring for relatives with better ditions, time by treatment effect estimate ⫽ ⫺2.11, t(177.17) ⫽
functional status, t(133) ⫽ 2.01, p ⬍ .05, and had lower scores at 1.13, p ⫽ .261. Similarly, the CBT and CG conditions were not
preintervention in depression, t(133) ⫽ ⫺3.98, p ⬍ .01. significantly different in change from baseline to follow-up, time
by treatment effect estimate ⫽ ⫺2.64, t(183.10) ⫽ 1.21, p ⫽ .228.
Differences Between Treatment Conditions Cohen’s d for these estimates was 0.28 and 0.36, respectively.
at Preintervention In summary, CBT was superior to the CG condition for depres-
sive symptomatology (at posttreatment and follow-up), but not for
No significant differences in the sociodemographic and stressors
anxiety.
variables were found between treatment conditions at preinterven-
tion, except for time since being a caregiver (see Table 1). Care-
givers in the CG had been caring for their relatives for more time Changes in Primary Outcomes, ACT Versus CG
than those in the ACT or CBT groups.
Contrasts comparing change in depressive symptomatology
Treatment Implementation scores from baseline to posttreatment indicated significantly
greater change favoring the ACT condition, time by treatment
Satisfaction with the interventions. No differences between effect estimate ⫽ ⫺10.47, t(179.85) ⫽ 4.40, p ⬍ .001. Cohen’s d
treatment conditions were found in the rating of satisfaction with
for this comparison was 1.17, which is a large effect. However,
the intervention (mean ⫾ SD for CBT ⫽ 9.60 ⫾ 0.68; mean ⫾ SD
when comparing ACT with the CG condition on change from
for ACT ⫽ 9.48 ⫾ 0.99), with the therapists (mean ⫾ SD for
baseline to follow-up, these differences were not maintained, time
CBT ⫽ 9.75 ⫾ 0.44; mean ⫾ SD for ACT ⫽ 9.81 ⫾ 0.40), and
by treatment effect estimate ⫽ ⫺3.48, t(184.95) ⫽ 1.30, p ⫽ .196.
with the contents of the intervention (mean ⫾ SD for CBT ⫽
9.45 ⫾ 0.68; mean ⫾ SD for ACT ⫽ 9.10 ⫾ 1.11). Participants in This comparison yielded a Cohen’s d of 0.39.
the ACT condition rated the intervention exercises as more diffi- A comparison of change in anxiety showed that participants
cult than did participants in the CBT intervention (mean ⫾ SD for in ACT had significantly greater change from baseline to post-
CBT ⫽ 6.55 ⫾ 1.98; mean ⫾ SD for ACT ⫽ 4.00 ⫾ 3.39), treatment, time by treatment effect estimate ⫽ ⫺5.84,
t(62) ⫽ ⫺2.87, p ⬍ .01. t(176.58) ⫽ 3.19, p ⫽ .002. However, this effect was not
Treatment receipt. The results show that participants in both significant from baseline to follow-up, time by treatment effect
CBT and ACT conditions significantly increased their perceived estimate ⫽ ⫺3.67, t(181.38) ⫽ 1.79, p ⫽ .077. Cohen’s d for
knowledge and intervention-associated skills from pre- to postin- these estimates was 0.79 and 0.49, respectively.
tervention. For CBT, mean ⫾ SD at pretreatment was 95.60 ⫾ In summary, ACT was superior to the CG condition for
29.24, and mean ⫾ SD at posttreatment was ⫽ 140.52 ⫾ 16.63; depressive symptomatology and anxiety at postintervention, but
t(28) ⫽ ⫺6.44, p ⬍ .01. For ACT, mean ⫾ SD at preintervention not at follow-up.
 766 LOSADA ET AL.

Table 2
Estimated Means Across Time by Treatment Condition

Time
Baseline Posttreatment Follow-up
Measure Condition Mean (SE) Mean (SE) Mean (SE)

Primary outcomes
CES–D CG (n ⫽ 48) 28.10 (1.39) 25.23 (1.66) 25.27 (1.91)
CBT (n ⫽ 42) 27.88 (1.49) 16.30 (1.71) 18.39 (2.04)
ACT (n ⫽ 45) 28.18 (1.44) 14.83 (1.63) 21.87 (1.81)
Anxiety CG (n ⫽ 48) 20.65 (1.09) 18.36 (1.29) 17.39 (1.48)
CBT (n ⫽ 42) 19.38 (1.16) 14.99 (1.33) 13.49 (1.58)
ACT (n ⫽ 45) 21.20 (1.12) 13.08 (1.27) 14.27 (1.41)
Secondary outcomes
CG (n ⫽ 48)
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Pleasant events 5.25 (0.31) 5.46 (0.36) 5.70 (0.41)

CBT (n ⫽ 42)
This document is copyrighted by the American Psychological Association or one of its allied publishers.

4.57 (0.33) 5.92 (0.37) 5.39 (0.44)

ACT (n ⫽ 45) 5.16 (0.32) 6.67 (0.36) 5.67 (0.40)
Dysfunctional thoughts CG (n ⫽ 48) 23.22 (1.70) 23.05 (1.94) 21.10 (2.18)
CBT (n ⫽ 42) 24.43 (1.80) 15.67 (2.01) 19.01 (2.31)
ACT (n ⫽ 45) 24.76 (1.74) 19.62 (1.92) 17.90 (2.09)
Caregiving experiential
avoidance CG (n ⫽ 48) 43.73 (1.23) 43.68 (1.49) 43.43 (1.72)
CBT (n ⫽ 42) 44.68 (1.32) 42.19 (1.52) 42.92 (1.84)
ACT (n ⫽ 45) 43.91 (1.27) 37.21 (1.45) 38.76 (1.63)
Note. CES–D ⫽ Center for Epidemiological Studies Depression scale; CG ⫽ minimal support control group;
CBT ⫽ cognitive– behavioral therapy; ACT ⫽ acceptance and commitment therapy.

Changes in Primary Outcomes, CBT Versus ACT of 16 at baseline. Therefore, in addition to NNT using the
CES–D cutoff, we also evaluated clinically significant change
The CBT and ACT conditions differed significantly only on for our primary outcomes. Clinically significant change, as
anxiety outcome. Specifically, a significant difference favoring defined by Jacobson and Truax (1991), involves (a) moving
ACT was observed at posttreatment, time by treatment effect below a cutoff score in which individuals are said to be in the
estimate ⫽ ⫺3.73, t(173.86) ⫽ 2.01, p ⫽ .046, but not at follow-
range of normal functioning (i.e., CES–D ⬍16), and (b) dem-
up, time by treatment effect estimate ⫽ ⫺1.04, t(180.14) ⫽ 0.49,
onstrating clinically meaningful change as per the RCI. Indi-
p ⫽ .63. Cohen’s d for these estimates was 0.50 and 0.14, respec-
viduals who meet RCI criteria and who fall below the cutoff of
tively.
16 are classified as “recovered,” whereas those who meet RCI
criteria alone are classified as “improved.” Using CES–D nor-
Changes in Primary Outcomes, Clinical Effect Sizes mative data provided by Lewinsohn, Seeley, Roberts, and Allen
For depressive symptomatology, we computed NNT at post- (1997), we established RCI criteria as at least an 18-point
treatment and follow-up by defining treatment success as scor- reduction on the CES–D. At posttreatment, eight of 30 partic-
ing ⬍16 on the CES–D at the postintervention assessment time ipants (26.7%) in the CBT condition reached recovery, as did
points. Success rates at posttreatment were 10/31 (32.3%) partic- eight of 33 participants (24.2%) of participants in the ACT
ipants in the CG condition, 18/30 (60.0%) participants in the CBT condition. In contrast, none of the 31 participants (0.0%) in the
condition (NNT ⫽ 3.61), and 20/33 participants (60.6%) in the CG condition achieved recovery. An additional participant in
ACT condition (NNT ⫽ 3.53). Success rates at the follow-up were CBT (3.3%) and two participants in ACT (6.1%) were classified
7/22 participants (31.8%) in the CG condition, 8/19 participants as improved. The NNT for CBT was 3.33; for ACT, it was 3.30.
(42.1%) in the CBT condition (NNT ⫽ 9.71), and 9/25 participants At follow-up, two of 19 participants (10.53%) in the CBT
(36.0%) in the ACT condition (NNT ⫽ 23.81). For anxiety, we condition reached recovery, as did one of 25 participants (4.0%)
computed NNT at posttreatment and follow-up defining treatment in the ACT condition. One of 19 participants in CBT (5.26%)
success as scoring ⬍13 at the postintervention measurement time and two participants in ACT (8.0%) were classified as im-
points. Success rates at posttreatment were 7/31 (22.58%) in the proved. None of 22 participants (0.0%) in the CG condition
CG condition, 14/30 (46.67%) participants in the CBT condition achieved recovery or clinically improved. The NNT for CBT
(NNT ⫽ 4.15), and 16/33 (48.49%) in the ACT group (NNT ⫽ was 6.33; for ACT, it was 8.33.
3.86). Success rates at the follow-up were 8/22 participants Using POMS normative data provided by Nyenhuis et al.
(36.36%) in the CG condition, 7/19 participants (36.84%) in the (1999), we established RCI criteria as at least a 7-point reduc-
CBT condition (NNT ⫽ 20.83), and 9/25 participants (36%) in the tion on the POMS. At posttreatment, seven of 30 participants
ACT condition (NNT ⫽ 27.78). (23.33%) in the CBT condition reached recovery, as did 12 of
It is important to note that change from ⱖ16 to ⬍16 on the 33 participants (36.36%) of participants in the ACT condition.
CES–D may reflect chance change, as well as measurement In contrast, two of the 31 participants (6.45%) in the CG
error, particularly for those individuals already close to a score condition achieved recovery. An additional participant in CBT
 CBT VS. ACT FOR CAREGIVERS 767
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Figure 2. Changes in primary and secondary outcomes. CBT ⫽ cognitive– behavioral therapy; ACT ⫽
acceptance and commitment therapy.

(3.23%), seven participants in ACT (21.21%), and six partici- classified as improved. The NNT for CBT was 10.45; for ACT,
pants in the CG condition (19.35%) were classified as im- it was 7.86.
proved. The NNT for CBT was 13.16; for ACT, it was 4.03. At In summary, the proportion of caregivers recovered in de-
follow-up, five of 19 participants (26.32%) in the CBT condi- pressive symptomatology was similar for CBT and ACT at
tion reached recovery, as did nine of 25 participants (36%) in posttreatment and follow-up, whereas no participants in the CG
the ACT condition. In contrast, three of the 22 participants condition was recovered. For anxiety, the proportions of par-
(13.64%) in the CG condition achieved recovery. Two addi- ticipants in the ACT and CBT conditions recovered at postin-
tional participants in CBT (10.53%), one participant in ACT tervention or follow-up were superior to those observed in the
(4%), and three participants in the CG condition (13.64%) were CG condition.
 768 LOSADA ET AL.

Changes in Secondary Outcomes, CBT Versus CG Changes in Secondary Outcomes, CBT Versus ACT
Contrasts comparing change in engagement in pleasant events The ACT and CBT conditions were not significantly different in
from baseline to posttreatment indicated that CBT participants had change in any of the secondary outcomes from baseline to post-
significantly greater change from baseline to posttreatment, time treatment or from baseline to follow-up.
by treatment effect estimate ⫽ 1.14, t(183.91) ⫽ 2.21, p ⫽ .03.
Cohen’s d for this estimate was 0.53. However, when comparing Discussion
CBT with the CG condition on change from baseline to follow-up, The objective of this study was to analyze the differential
this effect was not significant, time by treatment effect estimate ⫽ efficacy of ACT and CBT for dementia family caregivers with
0.37, t(189.05) ⫽ 0.62, p ⫽ .54. Cohen’s d for this effect was 0.17. high depressive symptomatology. Both interventions were applied
Comparison of change in dysfunctional thoughts from baseline in an individual therapy format and compared with a minimal
to posttreatment indicated significantly greater reductions favoring support CG. The results of the study provide support for our
the CBT condition, time by treatment effect estimate ⫽ ⫺8.59, primary hypothesis: ACT and CBT interventions seem to be su-
t(169.59) ⫽ 3.47, p ⬍ .001. Cohen’s d for this comparison was
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

perior to CG in terms of statistical and clinically significant re-

0.72. However, this effect was not significant from baseline to
This document is copyrighted by the American Psychological Association or one of its allied publishers.

duction of distress (i.e., depression and anxiety).

follow-up, time by treatment effect estimate ⫽ ⫺3.30, t(172.39) ⫽ Regarding depressive symptomatology, the results of the study
1.14, p ⫽ .25. Cohen’s d for this effect was 0.28. suggest that both ACT and CBT interventions are superior to CG
Comparison of change in caregiving experiential avoidance at postintervention (with large effect sizes), although they are only
from baseline to posttreatment showed no significant differences maintained in CBT at follow-up (medium effect size). When
between CBT and control conditions, time by treatment effect examining the clinical significance of these results, approximately
estimate ⫽ ⫺2.43, t(179.13) ⫽ 1.07, p ⫽ .29. Similarly, no 60% of participants in both CBT and ACT interventions scored
significant differences between CBT and the CG condition from lower than 16 at postintervention. The obtained NNT suggests that,
baseline to follow-up were found, time by treatment effect esti- both for CBT and ACT, out of every four caregivers treated with
mate ⫽ ⫺1.45, t(187.84) ⫽ 0.55, p ⫽ .58. Cohen’s d for these each of these interventions, one would show a clinically meaning-
estimates was 0.28 and 0.17, respectively. ful reduction in depressive symptomatology compared with the
In summary, CBT was superior to the CG condition for CG condition. Using RCI criteria, results of the study show that a
leisure and dysfunctional thoughts at postintervention, but not remarkable proportion of caregivers from both ACT and CBT
at follow-up. interventions had a clinically significant change at postintervention
(scored ⬍16 on the CES–D and at least an 18-point reduction on
Changes in Secondary Outcomes, ACT Versus CG
the CES–D) or were improved but not recovered (at least 18 points
A comparison of change in engagement in pleasant events on the CES–D but still scored above 16 at posttreatment). None of
showed that ACT participants had significantly greater change the participants in the CG showed a clinically significant change or
from baseline to posttreatment, time by treatment effect estimate ⫽ were improved but not recovered at postintervention. However,
1.30, t(183.35) ⫽ ⫺2.58, p ⫽ .01, compared with the CG condi- clinically significant changes at the long-term follow-up appeared
tion. Cohen’s d for this effect was 0.62. However, no significant diluted.
differences were found for change in engagement in pleasant The findings provide additional support for the use of CBT
events scores from baseline to follow-up, time by treatment effect interventions for reducing depressive symptomatology in dementia
estimate ⫽ 0.07, t(187.53) ⫽ ⫺0.12, p ⫽ .91, with Cohen’s d of caregivers (Gallagher-Thompson & Coon, 2007). Our findings
0.03. also suggest that CBT and ACT are similar in terms of their effects
When examining change in dysfunctional thoughts from base- on depression in the short term (i.e., significant and strong effects
line to posttreatment, a significant difference between the ACT and were found for both at postintervention). However, as already
control conditions favoring ACT was observed, time by treatment noted, CBT has an advantage over ACT in the long-term mainte-
effect estimate ⫽ ⫺4.97, t(169.20) ⫽ 2.05, p ⫽ .04. Cohen’s d for nance of gains with regard to depression. These findings are
this effect was 0.41. However, this effect was not significant when similar to previous studies done with noncaregiver samples, which
comparing change from baseline to follow-up, time by treatment have found a stronger maintenance of effects of CBT interventions
effect estimate ⫽ ⫺4.74, t(171.64) ⫽ 1.74, p ⫽ .08, with Cohen’s over ACT interventions (Forman, Shaw, et al., 2012). A possible
d of 0.40. explanation for these results is related to the different difficulty
Contrasts comparing change in caregiving experiential avoid- perceived by caregivers to practice and apply learned strategies in
ance scores from baseline to posttreatment indicated significantly CBT and ACT. The analysis of the treatment implementation
greater change favoring the ACT condition, time by treatment suggests that ACT participants report a higher degree of difficulty
effect estimate ⫽ ⫺6.65, t(178.44) ⫽ 2.98, p ⫽ .01. Cohen’s d for of homework assignments than CBT participants. As has been
this comparison was 0.77. However, when comparing ACT with suggested (Forman, Shaw, et al., 2012), CBT is a parsimonious
the CG condition on change from baseline to follow-up, these and intuitive intervention, more compatible with folk psychology,
differences were not maintained, time by treatment effect esti- and thus facilitates the ability of patients to apply and generalize
mate ⫽ ⫺4.85, t(185.44) ⫽ 1.94, p ⫽ .054, with Cohen’s d of learned strategies in their everyday life. However, ACT is more
0.56. complex, as individuals need to understand abstract concepts such
In summary, ACT was superior to the CG condition for leisure, as experiential avoidance, assume counterintuitive ideas that are
dysfunctional thoughts, and experiential avoidance at postinterven- somehow opposite to the cultural and social norms (e.g., the
tion, but not at follow-up. ubiquity of suffering), and learn exercises and techniques that
 CBT VS. ACT FOR CAREGIVERS 769

require a significant degree of metacognitive abilities. Hence, it is and a decrease in experiential avoidance in ACT. Even when these
likely that ACT requires from patients more time and effort, as results suggest that these interventions may present different
well as more support from therapists, to understand its rationale, mechanisms of action, a closer look at the results reveals a more
learn its techniques, and generalize the ACT-trained abilities over complex picture. Specifically, although significant increases in
the long term. It could be possible that CBT would be a more behavioral activation and decreases in dysfunctional thoughts were
recommendable intervention for caregivers with high levels of found for both CBT and ACT, significant decreases in experiential
depressive symptomatology, but also for those with difficulties for avoidance were found only for ACT. Hence, these results are in
completing therapy. line with the conclusion by Forman, Chapman, et al. (2012) that
With respect to anxious symptomatology, participants from the ACT and CBT may have common mediator pathways but also
ACT intervention showed significantly lower scores in anxiety at different ones, or the Arch and Craske (2008) statement that ACT
postintervention than caregivers from the CBT and CG conditions. and CBT are “more similar than distinct” (p. 263). Focusing on the
However, when analyzing the number of participants who dem- common mediators of change, the results of this study may suggest
onstrated clinically significant improvement in anxiety, the per- that the effect of both CBT and ACT for caregivers’ depression
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

centage of participants from the ACT and CBT conditions (48.49% may be due to changes in the frequency of pleasant events and
This document is copyrighted by the American Psychological Association or one of its allied publishers.

and 46.67%, respectively) was superior to that observed in the CG dysfunctional thoughts. It is plausible that CBT and ACT share
condition (22.6%). As observed for depressive symptomatology, these mechanisms of action, even though the therapeutic objectives
the NNT for the ACT and CBT conditions suggested that out of of each intervention are based on different (although compatible)
every four caregivers treated with these interventions, one would theoretical models, which is consistent with previous research
show a clinically significant reduction in anxiety symptoms as (Forman, Chapman, et al., 2012). On the one hand, in CBT and
compared with caregivers in the CG condition. Again, the results ACT individuals increase their levels of engagement in pleasurable
were diluted at follow-up, although a clinically meaningful reduc- activities by using different psychological techniques: behavioral
tion was found for 36% of the participants in the ACT intervention, activation in CBT and committed actions related to values in ACT.
26.32% in the CBT intervention, and 13.64% in the CG condition. On the other hand, it is also likely that changing one’s relationship
The findings provide preliminary support for the efficacy of ACT with one’s negative thoughts using cognitive defusion and mind-
for reducing anxious symptomatology in dementia caregivers and fulness techniques in ACT produces a reduction in the frequency
contribute to the sparse number of caregiving intervention studies of dysfunctional thoughts. In this sense, a study with clinical
that measure anxious symptomatology (Cooper et al., 2007; Li et students using session-by-session measurement showed that de-
al., 2013). creases in self-reported dysfunctional thinking and willingness to
Previous studies comparing ACT’s and CBT’s efficacy for engage in behavioral activity despite unpleasant thoughts or emo-
reducing anxiety suggest that ACT may produce stronger effects tions were equivalent mediators across CBT and ACT (Forman,
than CBT on this outcome variable (Brown et al., 2011), and CBT Chapman, et al., 2012).
interventions for caregivers have been primarily developed to In the current study, however, experiential avoidance changed
target depression but not anxiety (Cooper et al., 2007). The par- significantly only in the ACT intervention. Similarly, Forman,
ticular efficacy of ACT to reduce anxiety symptoms from baseline Chapman, et al. (2012) found that increased use of psychological
to posttreatment may be related to the fact that this therapy places acceptance strategies relative to cognitive or affective change
a strong emphasis on validating and accepting negative emotions strategies mediated the outcome for ACT, whereas for CBT, the
in general, which may help caregivers to normalize their distress, mediation was in the opposite direction. It is important to note,
decrease their apprehension and worry regarding their aversive however, that the nature of this study does not allow us to make
emotions and, hence, eventually reduce their general activation or inferences about the mechanism of actions of the interventions,
anxiety. It is plausible that the ACT treatment was also useful in and future studies using mediational analyses are needed to ana-
reducing caregivers’ anxiety symptoms through training them in lyze the mediators of change in these interventions.
cognitive defusion skills for reducing experiential avoidance, Even when this study found important effect sizes of both CBT
which are taught via acceptance and mindfulness exercises. Com- and ACT at posttreatment in caregiver depression, the effects of
pared with the components of the CBT intervention, these types of these interventions were diluted at follow-up (6 months after
exercises may facilitate a more direct exposure to uncomfortable postassessment). Although CBT effects seemed to be maintained
internal sensations, feelings, and avoided thoughts, which eluci- for depression at follow-up, the effect size varied from large at
date habituation and extinction of anxious responses (Arch & posttreatment to medium at follow-up. Adding booster sessions
Craske, 2008). In fact, a decrease in caregivers’ levels of experi- has been suggested as a strategy to consolidate the outcomes of the
ential avoidance has been found only after the ACT treatment. It is interventions, by allowing the therapists to supervise the caregiv-
plausible that to reduce caregivers’ anxiety symptoms, training ers’ use of the trained skills (e.g., Moore et al., 2013). In addition,
caregivers in acceptance coping strategies (mostly used in ACT) the analysis of the treatment implementation and, specifically,
was more useful than control/change coping strategies (mostly treatment enactment suggests that additional efforts aimed at
used in CBT), which is consistent with previous research showing increasing the quantity and quality of homework assignments
that one of the most significant predictors of caregivers’ anxiety is by participants may contribute to increase the effects of the
the frequency with which they use escape-avoidance coping strat- interventions.
egies (Cooper et al., 2007). This study presents several limitations. Participants were care-
Regarding the analysis of the short-term treatment effects on givers who volunteered to participate and may not be representa-
coping variables, the results of the study provide support for our tive of the overall caregiver population. Furthermore, although the
hypothesis regarding a decrease in dysfunctional thoughts in CBT results of this study may generalize to caregivers with high levels
 770 LOSADA ET AL.

of depressive symptomatology, they may not generalize to the specific mediators of change that explain the effects of each
subset of caregivers who suffer from high levels of burden or intervention.
anxiety but with low levels of depression. Future studies should In conclusion, the results of this study confirm that psycholog-
also consider including participants with high levels in other ical interventions are well received by dementia family caregivers
outcome variables such as burden or anxiety. As is usual in and provide them with skills to cope with the demanding and
randomized clinical trials, the fact that interventions were stan- exhausting task of caring for a relative diagnosed with dementia. A
dardized and manualized clearly limits the individualization of remarkable percentage of participants in both interventions re-
each intervention to the specific needs of each caregiver. Future ported changes in symptomatology considered clinically signifi-
studies should include indicators of treatment delivery (e.g., ad- cant. Given the promising results for the treatment of anxiety and
herence to the manual) for increasing the internal validity of the depression using ACT with caregivers, the findings of this study
study. Even though the flow of participants through the study (see support this intervention as a viable and effective treatment for a
Figure 1) was similar to those reported in other studies with population for whom previous treatment has had limited success.
caregivers (e.g., Moore et al., 2013) or other populations (e.g., Hence, findings suggest that ACT may be an appropriate alterna-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

anxiety disorders; Arch et al., 2012), completers showed lower tive to CBT for helping depressed dementia caregivers.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

scores at preintervention in depressive symptomatology and pro-

vided care for a fewer number of hours per day than those who
discontinued participation. These findings suggest that a number of References
caregivers have a profile that may require additional assistance for Alzheimer’s Association. (2013). 2013 Alzheimer’s disease facts and fig-
continuing their participation in interventions such as the ones ures. Alzheimer’s & Dementia, 9, 208 –245. http://dx.doi.org/10.1016/j
described here. In addition, there was an important number of .jalz.2013.02.003
caregivers who reported being no longer interested in the inter- APA Publications and Communications Board Working Group on Journal
vention, mainly because of lack of time or difficulties finding help Article Reporting Standards. (2008). Reporting standards for research in
for caring for the relative while the caregiver attended the inter- psychology: Why do we need them? What might they be? American
Psychologist, 63, 839 – 851. http://dx.doi.org/10.1037/0003-066X.63.9
vention. The obtained figures are similar those reported in the
.839
caregiving literature (Fortinsky, Kulldorff, Kleppinger, & Kenyon- Arch, J. J., & Craske, M. G. (2008). Acceptance and commitment therapy
Pesce, 2009). Providing formal resources (e.g., home care or day and cognitive– behavioral therapy for anxiety disorders: Different treat-
care) to caregivers may increase the chances of their participation ments, similar mechanism? Clinical Psychology: Science and Practice,
in these types of interventions. Finally, an important limitation of 15, 263–279. http://dx.doi.org/10.1111/j.1468-2850.2008.00137.x
the study has to do with the design of the control group. Partici- Arch, J. J., Eifert, G. H., Davies, C., Vilardaga, J. C., Rose, R. D., &
pants in the ACT and CBT conditions had eight weekly individual Craske, M. G. (2012). Randomized clinical trial of cognitive– behavioral
sessions with a therapist, whereas the CG had a single, 2-hr group therapy (CBT) versus acceptance and commitment therapy (ACT) for
workshop. We cannot rule out the possibility that the differences mixed anxiety disorders. Journal of Consulting and Clinical Psychology,
80, 750 –765. http://dx.doi.org/10.1037/a0028310
between the intervention and control conditions were due to the
Bender, R. (2001). Calculating confidence intervals for the number needed
increased contact or increased perceived social support (Roth,
to treat. Controlled Clinical Trials, 22, 102–110.
Mittelman, Clay, Madan & Haley, 2005) rather than the compo- Blackledge, J. T., & Hayes, S. C. (2006). Using acceptance and commit-
nents of the interventions. We opted for this design considering ment training in the support of parents of children diagnosed with
that previous studies done by our research group have already autism. Child & Family Behavior Therapy, 28, 1–18. http://dx.doi.org/
shown that CBT was superior to a treatment-as-usual control group 10.1300/J019v28n01_01
(Losada et al., 2011; Márquez-González et al., 2007). In addition, Brown, L. A., Forman, E. M., Herbert, J. D., Hoffman, K. L., Yuen, E. K.,
as has been done in other studies, such as the REACH Project & Goetter, E. M. (2011). A randomized controlled trial of acceptance-
(Wisniewski et al., 2003), the workshop and written material were based behavior therapy and cognitive therapy for test anxiety: A pilot
offered to participants in the CG for ethical reasons as well as for study. Behavior Modification, 35, 31–53. http://dx.doi.org/10.1177/
0145445510390930
facilitating recruitment and retention of participants. Future studies
Burgio, L., Lichstein, K. L., Nichols, L., Czaja, S., Gallagher-Thompson,
should consider the inclusion of an attentional control group.
D., Bourgeois, M., . . . Schulz, R. (2001). Judging outcomes in psycho-
Despite these limitations, the contribution of the present study is social interventions for dementia caregivers: The problem of treatment
threefold; (a) To our knowledge, this is the first randomized implementation. The Gerontologist, 41, 481– 489. http://dx.doi.org/
clinical trial to analyze the efficacy of an ACT intervention with 10.1093/geront/41.4.481
dementia caregivers, as compared with CBT and control groups. Burke, K., Muscara, F., McCarthy, M., Dimovski, A., Hearps, S., Ander-
Results show that both treatments (CBT and ACT) improved son, V., & Walser, R. (2014). Adapting acceptance and commitment
depression in the short term, but only CBT showed maintenance of therapy for parents of children with life-threatening illness: Pilot study.
these effects in the long term. (b) This is the first study to show Families, Systems, & Health, 32, 122–127. http://dx.doi.org/10.1037/
significant effects of ACT in reducing caregivers’ levels of anxiety fsh0000012
Cooper, C., Balamurali, T. B., Selwood, A., & Livingston, G. (2007). A
symptoms in short term. (c) This study contributes to understand-
systematic review of intervention studies about anxiety in caregivers of
ing the possible mechanisms of different intervention programs for people with dementia. International Journal of Geriatric Psychiatry, 22,
producing positive outcomes in caregivers, namely specific coping 181–188. http://dx.doi.org/10.1002/gps.1656
strategies targeted by ACT and CBT. The limited understanding of Cuijpers, P. (2005). Depressive disorders in caregivers of dementia pa-
coping strategies has been noted as a sparse area of caregiving tients: A systematic review. Aging & Mental Health, 9, 325–330. http://
research (Li et al., 2013), and future studies should analyze the dx.doi.org/10.1080/13607860500090078
 CBT VS. ACT FOR CAREGIVERS 771

Feingold, A. (2009). Effect sizes for growth-modeling analysis for con- tions for future research. The Gerontologist, 33, 240 –248. http://dx.doi
trolled clinical trials in the same metric as for classical analysis. Psy- .org/10.1093/geront/33.2.240
chological Methods, 14, 43–53. http://dx.doi.org/10.1037/a0014699 Kohout, F. J., Berkman, L. F., Evans, D. A., & Cornoni-Huntley, J. (1993).
Forman, E. M., Chapman, J. E., Herbert, J. D., Goetter, E. M., Yuen, E. K., Two shorter forms of the CES–D depression symptoms index. Journal
& Moitra, E. (2012). Using session-by-session measurement to compare of Aging and Health, 5, 179 –193. http://dx.doi.org/10.1177/
mechanisms of action for acceptance and commitment therapy and 089826439300500202
cognitive therapy. Behavior Therapy, 43, 341–354. http://dx.doi.org/ Lappalainen, R., Lehtonen, T., Skarp, E., Taubert, E., Ojanen, M., &
10.1016/j.beth.2011.07.004 Hayes, S. C. (2007). The impact of CBT and ACT models using
Forman, E. M., Shaw, J. A., Goetter, E. M., Herbert, J. D., Park, J. A., & psychology trainee therapists: A preliminary controlled effectiveness
Yuen, E. K. (2012). Long-term follow up of a randomized controlled trial. Behavior Modification, 31, 488 –511. http://dx.doi.org/10.1177/
trial comparing acceptance and commitment therapy and cognitive ther- 0145445506298436
apy for anxiety and depression. Behavior Therapy, 43, 801– 811. http:// Lewinsohn, P. M., Seeley, J. R., Roberts, R. E., & Allen, N. B. (1997).
dx.doi.org/10.1016/j.beth.2012.04.004 Center for Epidemiologic Studies Depression Scale (CES–D) as a
Fortinsky, R. H., Kulldorff, M., Kleppinger, A., & Kenyon-Pesce, L. screening instrument for depression among community-residing older
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

(2009). Dementia care consultation for family caregivers: Collaborative adults. Psychology and Aging, 12, 277–287. http://dx.doi.org/10.1037/
This document is copyrighted by the American Psychological Association or one of its allied publishers.

model linking an Alzheimer’s Association chapter with primary care 0882-7974.12.2.277

physicians. Aging & Mental Health, 13, 162–170. http://dx.doi.org/ Li, R., Cooper, C., Austin, A., & Livingston, G. (2013). Do changes in
10.1080/13607860902746160 coping style explain the effectiveness of interventions for psychological
Gallagher-Thompson, D., & Coon, D. W. (2007). Evidence-based psycho- morbidity in family carers of people with dementia? A systematic review
logical treatments for distress in family caregivers of older adults. and meta-analysis. International Psychogeriatrics, 25, 204 –214. http://
Psychology and Aging, 22, 37–51. http://dx.doi.org/10.1037/0882-7974 dx.doi.org/10.1017/S1041610212001755
.22.1.37 Losada, A., Márquez-González, M., & Romero-Moreno, R. (2011). Mech-
Gallagher-Thompson, D., Coon, D. W., Solano, N., Ambler, C., Rabinow- anisms of action of a psychological intervention for dementia caregivers:
itz, Y., & Thompson, L. W. (2003). Change in indices of distress among Effects of behavioral activation and modification of dysfunctional
thoughts. International Journal of Geriatric Psychiatry, 26, 1119 –1127.
Latino and Anglo female caregivers of elderly relatives with dementia:
Losada, A., Márquez-González, M., Romero-Moreno, R., & López, J.
Site-specific results from the REACH National Collaborative Study. The
(2014). Development and validation of the Experiential Avoidance in
Gerontologist, 43, 580 –591. http://dx.doi.org/10.1093/geront/43.4.580
Caregiving Questionnaire (EACQ). Aging & Mental Health, 18, 897–
Gallagher-Thompson, D., Tzuang, Y. M., Au, A., Brodaty, H., Charles-
904. http://dx.doi.org/10.1080/13607863.2014.896868
worth, G., Gupta, R., . . . Shyu, Y. I. (2012). International perspectives
Losada, A., Montorio, I., Izal, M., & Márquez-González, M. (2006).
on nonpharmacological best practices for dementia family caregivers: A
Estudio e intervención sobre el malestar psicológico de los cuidadores
review. Clinical Gerontologist, 35, 316 –355. http://dx.doi.org/10.1080/
de personas con demencia: El papel de los pensamientos disfuncionales
07317115.2012.678190
[Study and intervention on dementia caregiving distress: The role of
González-Menéndez, A., Fernández, P., Rodríguez, F., & Villagrá, P.
dysfunctional thoughts]. Madrid, Spain: IMSERSO.
(2014). Long-term outcomes of acceptance and commitment therapy in
Mahoney, F. L., & Barthel, D. W. (1965). Functional evaluation: The
drug-dependent female inmates: A randomized controlled trial. Interna-
Barthel Index. State Journal, 4, 61– 65. Retrieved from http://www
tional Journal of Clinical and Health Psychology, 14, 18 –27. http://dx
.medchi.org/maryland-medicine-journal
.doi.org/10.1016/S1697-2600(14)70033-X
Márquez-González, M., Losada, A., Izal, M., Pérez-Rojo, G., & Montorio,
Hayes, S. C., Strosahl, K. D., & Wilson, K. G. (1999). Acceptance and
I. (2007). Modification of dysfunctional thoughts about caregiving in
commitment therapy: An experiential approach to behaviour change. dementia family caregivers: Description and outcomes of an intervention
New York, NY: Guilford Press. programme. Aging & Mental Health, 11, 616 – 625. http://dx.doi.org/
Hayes, S. C., Strosahl, K., & Wilson, K. G. (2011). Acceptance and 10.1080/13607860701368455
commitment therapy: The process and practice of mindful change (2nd Márquez-González, M., Romero-Moreno, R., & Losada, A. (2010). Care-
ed.). New York, NY: Guilford Press. giving issues in a therapeutic context: New insights from the acceptance
Hayes, S. C., Strosahl, K., Wilson, K. G., Bissett, R. T., Pistorello, J., and commitment therapy approach. In N. Pachana, K. Laidlaw, & B.
Toarmino, D., & McCurry, S. M. (2004). Measuring experiential avoid- Knight (Eds.), Casebook of clinical geropsychology: International per-
ance: A preliminary test of a working model. The Psychological Record, spectives on practice (pp. 33–53). New York, NY: Oxford University
54, 553–578. Retrieved from http://opensiuc.lib.siu.edu/tpr/vol54/iss4/5/ Press. http://dx.doi.org/10.1093/med/9780199583553.003.0003
Hesser, H., Gustafsson, T., Lundén, C., Henrikson, O., Fattahi, K., Johns- McCracken, L. M., & Jones, R. (2012). Treatment for chronic pain for
son, E., . . . Andersson, G. (2012). A randomized controlled trial of adults in the seventh and eighth decades of life: A preliminary study of
Internet-delivered cognitive behavior therapy and acceptance and com- acceptance and commitment therapy (ACT). Pain Medicine, 13, 860 –
mitment therapy in the treatment of tinnitus. Journal of Consulting and 867. http://dx.doi.org/10.1111/j.1526-4637.2012.01407.x
Clinical Psychology, 80, 649 – 661. http://dx.doi.org/10.1037/a0027021 McCurry, S. M. (2006). When a family member has dementia: Steps to
Jacobson, N. S., & Truax, P. (1991). Clinical significance: A statistical becoming a resilient caregiver. Oxford, UK: Greenwood Publishing.
approach to defining meaningful change in psychotherapy research. McGlinchey, J. B., Atkins, D. C., & Jacobson, N. S. (2002). Clinical
Journal of Consulting and Clinical Psychology, 59, 12–19. http://dx.doi significance methods: Which one to use and how useful are they?
.org/10.1037/0022-006X.59.1.12 Behavior Therapy, 33, 529 –550. http://dx.doi.org/10.1016/S0005-
Karlin, B. E., Walser, R. D., Yesavage, J., Zhang, A., Trockel, M., & 7894(02)80015-6
Taylor, C. B. (2013). Effectiveness of acceptance and commitment McNair, D., Lorr, M., & Droppleman, L. (1971). Profile of Mood States.
therapy for depression: Comparison among older and younger veterans. Manual. San Diego, CA: EdITS.
Aging & mental health, 17, 555–563. http://dx.doi.org/10.1080/ Moher, D., Schulz, K. F., & Altman, D. (2001). The CONSORT statement:
13607863.2013.789002 Revised recommendations for improving the quality of reports of
Knight, B. G., Lutzky, S. M., & Macofsky-Urban, F. (1993). A meta- parallel-group randomized trials. Journal of the American Medical As-
analytic review of interventions for caregiver distress: Recommenda- sociation, 285, 1987–1991. http://dx.doi.org/10.1186/1471-2288-1-2
 772 LOSADA ET AL.

Moore, R. C., Chattillion, E. A., Ceglowski, J., Ho, J., von Känel, R., Mills, correlates, and causes. The Gerontologist, 35, 771–791. http://dx.doi
P. J., . . . Mausbach, B. T. (2013). A randomized clinical trial of .org/10.1093/geront/35.6.771
behavioral activation (BA) therapy for improving psychological and Spira, A. P., Beaudreau, S. A., Jimenez, D., Kierod, K., Cusing, M. M.,
physical health in dementia caregivers: Results of the Pleasant Events Gray, H. L., & Gallagher-Thompson, D. (2007). Experiential avoidance,
Program (PEP). Behaviour Research and Therapy, 51, 623– 632. http:// acceptance, and depression in dementia family caregivers. Clinical Ger-
dx.doi.org/10.1016/j.brat.2013.07.005 ontologist, 30, 55– 64. http://dx.doi.org/10.1300/J018v30n04_04
Newcombe, R. G. (1998). Interval estimation for the difference between Stevens, A. B., Coon, D., Wisniewski, S., Vance, D., Arguelles, S., Belle,
independent proportions: Comparison of eleven methods. Statistics in S., . . . Haley, W. (2004). Measurement of leisure time satisfaction in
Medicine, 17, 873– 890. http://dx.doi.org/10.1002/(SICI)1097- family caregivers. Aging & Mental Health, 8, 450 – 459. http://dx.doi
0258(19980430)17:8⬍873::AID-SIM779⬎3.0.CO;2-I .org/10.1080/13607860410001709737
Nyenhuis, D. L., Yamamoto, C., Luchetta, T., Terrien, A., & Parmentier, Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th
A. (1999). Adult and geriatric normative data and validation of the ed.). Needham Heights, MA: Allyn & Bacon.
profile of mood states. Journal of Clinical Psychology, 55, 79 – 86. Teri, L., Truax, P., Logsdon, R., Uomoto, J., Zarit, S., & Vitaliano, P. P.
http://dx.doi.org/10.1002/(SICI)1097-4679(199901)55:1⬍79::AID- (1992). Assessment of behavioral problems in dementia: The Revised
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

JCLP8⬎3.0.CO;2-7 Memory and Behavior Problems Checklist. Psychology and Aging, 7,

Öst, L. G. (2008). Efficacy of the third wave of behavioral therapies: A 622– 631. http://dx.doi.org/10.1037/0882-7974.7.4.622
This document is copyrighted by the American Psychological Association or one of its allied publishers.

systematic review and meta-analysis. Behaviour Research and Therapy, Wetherell, J. L., Afari, N., Rutledge, T., Sorrell, J. T., Stoddard, J. A.,
46, 296 –321. http://dx.doi.org/10.1016/j.brat.2007.12.005 Petkus, A. J., . . . Atkinson, J. H. (2011). A randomized, controlled trial
Pinquart, M., & Sörensen, S. (2003). Differences between caregivers and of acceptance and commitment therapy and cognitive– behavioral ther-
noncaregivers in psychological health and physical health: A meta- apy for chronic pain. Pain, 152, 2098 –2107. http://dx.doi.org/10.1016/
analysis. Psychology and Aging, 18, 250 –267. http://dx.doi.org/10.1037/ j.pain.2011.05.016
0882-7974.18.2.250 Wetherell, J. L., Liu, L., Patterson, T. L., Afari, N., Ayers, C. R., Thorp,
Pinquart, M., & Sörensen, S. (2006). Helping caregivers of persons with S. R., . . . Sorrell, J. T. (2011). Acceptance and commitment therapy for
dementia: Which interventions work and how large are their effects? generalized anxiety disorder in older adults: A preliminary report. Be-
International Psychogeriatrics, 18, 577–595. http://dx.doi.org/10.1017/ havior Therapy, 42, 127–134. http://dx.doi.org/10.1016/j.beth.2010.07
S1041610206003462 .002
Radloff, L. S. (1977). The CES–D scale: A self-report depression scale for Wisniewski, S. R., Belle, S. H., Coon, D. W., Marcus, S. M., Ory, M. G.,
research in the general population. Applied Psychological Measurement, Burgio, L. D., . . . Schulz, R. (2003). The Resources for Enhancing
1, 385– 401. http://dx.doi.org/10.1177/014662167700100306 Alzheimer’s Caregiver Health (REACH): Project design and baseline
Reisberg, B., Ferris, S. H., de Leon, M. J., & Crook, T. (1982). The Global characteristics. Psychology and Aging, 18, 375–384. http://dx.doi.org/
Deterioration Scale for assessment of primary degenerative dementia. 10.1037/0882-7974.18.3.375
The American Journal of Psychiatry, 139, 1136 –1139. http://dx.doi.org/ Zettle, R. D., & Rains, J. C. (1989). Group cognitive and contextual
10.1176/ajp.139.9.1136 therapies in treatment of depression. Journal of Clinical Psychology, 45,
Roth, D. L., Mittelman, M. S., Clay, O. J., Madan, A., & Haley, W. E. 436 – 445. http://dx.doi.org/10.1002/1097-4679(198905)45:3⬍436::
(2005). Changes in social support as mediators of the impact of a AID-JCLP2270450314⬎3.0.CO;2-L
psychosocial intervention for spouse caregivers of persons with Alzhei-
mer’s disease. Psychology and Aging, 20, 634 – 644. http://dx.doi.org/
10.1037/0882-7974.20.4.634 Received October 23, 2014
Schulz, R., O’Brien, A. T., Bookwala, J., & Fleissner, K. (1995). Psychi- Revision received February 17, 2015
atric and physical morbidity effects of dementia caregiving: Prevalence, Accepted April 6, 2015 䡲