Pilot quasi-experimental single-arm study of a virtual intervention for caregivers of persons with Lewy body dementia

Abstract

Objectives:

This pilot study evaluated a web-based intervention, guided by problem-solving therapy, to address challenges faced by caregivers of individuals with Lewy body dementia (LBD).

Method:

A quasi-experimental single-arm study was conducted with 39 family caregivers (mean age: 67.62 years; 69% women; 95% White). The 8-week program, Virtual Online Communities for Aging Life Experiences Lewy Body Dementia (VOCALE LBD), included a discussion platform, peer support, training, and problem-solving practice. Measurements were taken at baseline, post-intervention, and one month later. Effect sizes and confidence intervals (CIs) were analyzed using bootstrapping, and demographic impacts were assessed through linear mixed-effects models (LMMs).

Results:

Of the 39 participants, 29 completed the study. There were no significant differences in demographics between completers and withdrawals. Participants’ engagement was high, defined as posting substantive comments at least twice weekly. Significant reductions were observed in depressive (d = −0.54), burden (d = −0.31), and stress scores (d = −0.45), while social support (d = 0.46), positive attitudes (d = 0.32), and negative attitudes towards problem-solving (d = −0.63) improved.

Conclusion:

This intervention shows promise in reducing caregiver burden and improving emotional well-being, offering a flexible, effective solution for LBD caregivers.

Introduction

Lewy body dementia (LBD) impacts approximately 1.4 million people in the U.S., with 80% receiving care at home from family and friends (Fernandes, 2011). Despite the sense of fulfillment that many derive from caregiving, it significantly challenges their health and well-being, as caregivers often experience deteriorating mental health, including depression, worsening physical health such as chronic diseases and injuries, and sleep deprivation (National Academies of Sciences et al., 2021). These health issues can diminish their ability to provide effective care (The Centers for Disease Control & Prevention, 2023), which in turn negatively affects the health of the person with dementia, increasing risks of falls (Kuzuya et al., 2006) and even abuse (Williamson & Shaffer, 2001). The interconnectedness of the caregiver’s and the care recipient’s health is particularly evident in studies on depression among dementia caregivers, where behavioral and psychological symptoms of dementia, especially neuropsychiatric issues like mood swings, agitation and aggression, lead to high caregiver burden and depression (Huang, 2022), that in turn affect health in people they care for.

Addressing the daily challenges and unmet needs of family caregivers is thus vital for the well-being of both the caregivers and the individuals they support. Adding to these challenges, current pharmacological solutions for LBD are largely experimental and rarely used in everyday clinical settings, focusing primarily on symptom management and should be used with caution as side effects may occur. For instance, classical neuroleptics, such as haloperidol, are best avoided in LBD as they may worsen motor function (Galvin, 2019). Consequently, family caregivers often bear the burden of managing severe symptoms associated with LBD (Galvin, 2019). This burden is compounded by the fact that when comparing similar stages of disease in AD and LBD, LBD caregivers report significantly poorer mental health (Yuuki et al., 2023). This increased burden is likely due to the extrapyramidal motor and autonomic symptoms unique to LBD, such as urinary incontinence (Tahami Monfared et al., 2019). Caregivers of LBD patients often report high rates of comorbidities such as hypertension, back pain, and arthritis. Depression rates in LBD caregivers are double that of AD cohorts, reaching a staggering 40–50%, and the prevalence of back pain is nearly equivalent to that in spinal cord injury caregivers (Fleisher et al., 2023). These issues are worsened by the physical demands of caregiving for someone with parkinsonian symptoms, as up to 85% of people with LBD experience falls and impaired dexterity, mobility, and transfers (Fleisher et al., 2023). Additionally, the Centers for Disease Control and Prevention notes that mental health concerns and caregiver burden increase caregivers’ risk of cardiovascular disease (Caregiving for Family & Friends—A Public Health Issue, 2019). Taken together, the unparalleled burden of LBD on patients and their caregivers underscores the critical need for improved support systems to effectively address the complex needs of LBD patients and their families.

Current interventions and gaps

Internet-based interventions for family caregivers of individuals with dementia have emerged as a promising avenue for reducing caregiver depression, paralleling the potential of traditional in-person approaches (Leng et al., 2020; Zhao et al., 2019). Psychoeducational interventions are one of the commonly offered modalities (Morgan et al., 2022), offering dementia knowledge, caregiving skills, behavior management, and resources for legal and financial advice, among others (Scerbe et al., 2023). Compared to traditional methods, these psycho-educational technology-assisted interventions offer cost-efficiency and flexibility, fitting into caregivers’ schedules (Scerbe et al., 2023; Zhu et al., 2021). Participants in internet-based psychoeducation programs reported feeling empowered and expressed a desire for continued access post-intervention (Scerbe et al., 2023; Söylemez et al., 2023; Yu et al., 2023). Notably, asynchronous digital education has proven as effective as in-person training, an essential feature for caregivers in remote or isolated situations (Scerbe et al., 2023). A systematic review and meta-analysis indicated that telehealth interventions could significantly reduce depression levels compared to traditional face-to-face approaches (Standard Mean Difference [SMD] = −0.34, p = 0.01), with lower odds of case-level depression (odds ratio [OR] 0.24, 95% CI: 0.07, −0.76) (Zhu et al., 2021). However, the overall impact of these interventions on caregiver depression appears modest, with small effect sizes (SMD = −0.23; n = 626; 95% CI:−0.38, −0.07) from a meta-analysis of four randomized control trials (Zhao et al., 2019) and (SMD = −0.21; n = 1524; 95% CI: −0.31, −0.10) from eleven randomized control trials (Leng et al., 2020). Beyond depression, web-based interventions have shown modest benefits for other caregiver outcomes, such as self-efficacy, caregiver burden, and stress (Scerbe et al., 2023). One of potential reasons for the underwhelming results in terms of their effect size of the interventions might be in the fact that majority of them are not specifically tailored to subtypes of dementia and caregiving circumstances and instead are broadly offered for caregivers of persons with Alzheimer’s Disease and Related Dementias (ADRD). Even though the current state effectively maintains broad access to the intervention and acknowledges the historically prolonged journey toward diagnoses of the specific types of neurodegenerative pathologies, advancements in diagnostic techniques—fueled in part by the advent of new therapies—and heightened public awareness alongside caregivers’ desires to engage with peers in similar situations, have laid the groundwork for the development of more refined approaches that are specifically catered to individual dementia subtypes.

Development of the tailored intervention

To address existing gaps, we designed an innovative online intervention Virtual Online Communities for Aging Life Experiences Lewy Body Dementia (VOCALE LBD) that is congruent with key components of the REACH II protocol—information provision, didactic instruction, problem-solving, and skills training (Belle et al., 2006)—for a fully remote, asynchronous format focusing on caregivers of individuals with LBD (Zaslavsky et al., 2022). This adaptation was guided by recommendations from the National Academies of Sciences, Engineering, and Medicine (NASEM) report, which advocates for the adaptation and further refinement of the REACH II protocol (Committee on Care Interventions for Individuals with Dementia and Their Caregivers et al. 2021; Larson & Stroud, 2021). The need for a tailored approach was underscored by a needs assessment among LBD caregivers, who reported a lack of specific support in generic groups and highlighted unique challenges such as dealing with hallucinations and sleep disturbances, which are common in LBD (Zaslavsky et al., 2022).

VOCALE LBD integrates techniques from problem-solving therapy (PST) along with social support. In this context, we seek to employ PST to empower caregivers with structured tools to identify and address challenges, fostering a sense of control and self-efficacy, which are crucial for coping with the complexities of caregiving (Nezu et al., 2010). Studies have shown that PST is associated with reductions in depression, anxiety, and burden (Fitzpatrick et al., 2013; Kiosses & Alexopoulos, 2014; Kirkham et al., 2016; Tao & Zhang, 2019; Washington et al., 2018), making it a relevant strategy in tailoring interventions for LBD caregivers. Social support, another key mechanism, has consistently been linked to enhanced psychological resilience and lower caregiver burden (Nemcikova et al., 2023; Ruisoto et al., 2020). By creating an online environment that incorporates both PST principles and opportunities for social engagement, the VOCALE LBD intervention seeks to address the dual goals of alleviating depression and enhancing caregiving mastery. In this project, we report results from a pilot deployment of the intervention in a single arm clinical trial.

Methods

Study design

This study was a pilot quasi-experimental single-arm study to examine acceptability, efficacy and sustained effects of the VOCALE LBD intervention. Outcome measures were collected at the beginning and end of the eight-week intervention and one month later to assess the intervention’s long-term impact. Additionally, problem-solving measures were assessed near baseline and after the intervention to track changes.

Ethics approval

The hosting university’s institutional review board (IRB) approved the study protocol (STUDY00015485). Written informed consent was obtained from all participants, and all data and results presented in this study were deidentified.

Recruitment and participants

Participants were primarily recruited from the Memory and Brain Wellness Center (MBWC) at Harborview Medical Center in Seattle, Washington. The MBWC evaluates over 1000 new patients annually and is the only major academic medical center in the five-state region of Washington, Wyoming, Alaska, Montana, and Idaho. The MBWC, which includes the MBWC clinic and the Alzheimer’s Disease Research Center (ADRC), is recognized as a Lewy Body Dementia Association Research Center of Excellence. The ADRC maintains an updated list of participants who have consented to be contacted for UW-affiliated research studies. Many of these Research Registry members have been evaluated at the MBWC clinic and have confirmed diagnoses of specific dementias, such as Alzheimer’s disease, Parkinson disease dementia (PDD), and Dementia with Lewy bodies (DLB). Participants were also recruited through posted announcements on the websites and newsletters of organizations such as Lewy Body Disease Association (LBDA) and Alzheimer.gov that serve the target population. To be eligible for participation, persons should be a family/informal caregiver of a person with a diagnosis of LBD; can read, write, and speak English; have a device that can access the Internet and be used for videoconferencing and/or phone calls; and be 18 years or older.

Interventions

VOCALE LBD is a group-based online intervention that is based on problem-solving therapy (PST), a cognitive-behavioral intervention focused on adaptive problem-solving attitudes and skills (Nezu et al., 2010). PST has been employed with different populations including adolescents (Eskin et al., 2008), caregivers (Washington et al., 2018), and older adults (Kiosses & Alexopoulos, 2014); and in diverse health contexts, including depression (Kirkham et al., 2016), diabetes (Fitzpatrick et al., 2013), and frailty (Chan et al., 2012; Chan et al., 2017). This group-based moderated peer-to-peer version of VOCALE LBD enables participants to learn and practice problem solving skills on their own time and in a place that is convenient to them, as well as benefit from sharing strategies for self-care and effective caregiving with others experiencing similar circumstances. Moreover, the VOCALE LBD intervention actively engages participants in learning and practicing problem solving skills using the ADAPT method: Adopting a positive attitude, Defining the problem, brainstorming Alternatives, Predicting consequences, and Trying out the chosen solution (Nezu et al., 2007). Each week is generally focused on practicing one new skill.

VOCALE LBD workflow

VOCALE LBD is an adaptation of the VOCALE intervention, initially refined in three rounds of deployment among older adults with pre-frailty and frailty (Chen et al., 2021; Chen et al., 2021; Teng et al., 2019), to the needs of caregivers of individuals with LBD. The VOCALE LBD intervention included weekly thematic discussion prompts that allowed participants to respond to a specific topic of interest at their leisure and provided the participants an opportunity to interact with each other through text. The first three weeks were focused on the most salient LBD caregiving experiences featured in previous literature (Armstrong et al., 2019; Galvin, 2019; Killen et al., 2016) and our formative work (Zaslavsky et al., 2022). Briefly, in this formative work, we conducted several iterations of individual interviews and focus groups with caregivers of persons with LBD to identify relevant topics for the intervention. We used open-ended prompts to elicit ideas that caregivers would be interested in discussing during the study. Sleep problems, hallucinations and delusions, and self-care emerged as the topics of greatest interest, and thus became topics for the first three weeks. The next five weeks involved psycho-educational materials using the ADAPT method based on PST, a cognitive-behavioral intervention focused on the adoption and application of adaptive problem-solving attitudes and skills (Nezu et al., 2010). A novel aspect of the VOCALE LBD intervention is engaging participants in the application of the problem solving skills to realistic caregiving scenarios. By engaging in each of the steps introduced in weeks 5–8, participants can augment their ability to address caregiving challenges. Please see Supplementary Table 1 for the schedule of topics.

The intervention also incorporated case studies that featured common challenges and problem-solving opportunities caregivers face in daily life, such as imposter syndrome, disrupted sleep, cognitive fluctuations, emotional outbursts, and so on. We employed the case studies to enable participants to practice solving realistic problems that they might face. Working with the case studies also provided other potential benefits, such as being able to practice problem solving without sharing personal experiences, as well as having an outlet to focus on someone else caregiving challenges. Weekly discussions were supervised or directly moderated by registered nurses who received training. Moderators logged in at least two times each weekday and at least once per weekend day to review new comments and add comments to address questions, provide emotional support and validation, encourage dialogue, and redirect discussions as needed to stay on topic. The moderators did not provide medical advice. Moderators also monitored for comments with inaccurate information and addressed them through private email or discussion board comments. Lastly, moderators sent personalized reminder emails once a week to participants who had not yet posted. Prior to the study, participants received remote training sessions where they were introduced to the platform. Throughout the study, educational and training materials were accessible on demand in various formats to help maintain proficiency in using the intervention platform.

Measures

Depressive symptoms

Depressive symptoms were assessed using the 20-item Center for Epidemiologic Studies Depression (CES-D) Scale (Radloff, 1977). Respondents rated the frequency of depressive feelings over the past week on a scale from 0 (rarely or never) to 3 (most or all of the time). Items 4, 8, 12, and 16 were reversed so that scores range from 0 to 60, with higher scores indicating more severe depressive symptoms. A score of 16 points or more is considered depressed (Lewinsohn et al., 1997). Cronbach’s alpha at baseline was 0.89.

Caregiver burden

Caregiver burden was measured with the short 12-item Zarit Caregiver Burden Interview (Bédard et al., 2001). Caregivers rated each item on a 5-point scale from 0 (never) to 4 (nearly always), resulting in a total score range from 0 to 48. Higher scores reflect greater caregiver burden. Cronbach’s alpha at baseline was 0.78.

Perceived stress

The full 10-item Perceived Stress Scale was utilized to assess perceived stress levels (Ezzati et al., 2014). Each item was rated on a 5-point scale from 0 (never) to 4 (very often), and items 4, 5, 7, and 8 were reversed to yield total scores ranging from 0 to 40 and higher scores indicate higher perceived stress. Scores ranging from 0–13 would be considered low stress, scores ranging from 14 to 26 would be considered moderate stress, and scores ranging from 27 to 40 would be considered high perceived stress. Cronbach’s alpha at baseline was 0.90.

Loneliness

Loneliness was measured using a short 3-item version of the Revised UCLA Loneliness Scale (Russell et al., 1980). Scores range from 0 to 6, with higher scores indicating greater feelings of loneliness. Cronbach’s alpha at baseline was 0.83.

Social support

Social support was assessed using a 9-item questionnaire from the Medical Outcomes Study (Sherbourne & Stewart, 1991). Participants rated the availability of social support on a 5-point scale from “none of the time” to “all of the time.” Total social support scores were computed by summing the item scores, ranging from 9 to 45, with higher scores indicating more social support. There are four subdomains: Emotional/information support subscale (Items 1, 2, 5, and 7), Affection support subscale (Item 9), Tangible support subscale (Items 3, 6), and Positive social interaction subscale (Items 4 and 8). Cronbach’s alpha at baseline was 0.90.

Self-efficacy

Self-efficacy related to health management was measured using a 5-item Health Self-Efficacy Measure (Lee et al., 2008). Respondents indicated their level of agreement with each statement on a scale from 1 (strongly disagree) to 5 (strongly agree). Scores range from 5 to 25, with higher scores reflecting stronger health self-efficacy. Cronbach’s alpha at baseline was 0.77.

Mastery

Seven-item Pearlin Mastery Scale was used to measure mastery (Pearlin & Schooler, 1978). On a scale of 1–4 (with 1 being “strongly disagree” and 4 being “strongly agree”), the measure assessed how strongly do you agree or disagree. Items 1 through item 5 were reversed so scores range from 7 to 28 and higher scores indicate greater mastery. Cronbach’s alpha at baseline was 0.85.

Problem solving

Problem-solving is assessed through five subscales, comprising a total of 25 items with 5 items per subscale. These subscales include two problem orientation dimensions (positive attitudes and negative attitudes) and three problem-solving styles (rational, impulsive/careless, and avoidant) (Nezu et al., 2007). Rational problem-solving involves systematic and deliberate efforts in application of problem-solving skills. Impulsivity/carelessness reflects incomplete, hasty attempts, while avoidance is marked by procrastination and inaction. Positive problem orientation and rational problem-solving are strengths, while the other three are considered weaknesses. At baseline, Cronbach alphas were: Positive Attitudes (0.44), Rational Problem Solving (0.61), Negative Attitudes (0.45), Impulsivity (0.50), and Avoidance (0.67).

There is some criticism of Cronbach’s alpha as a measure of internal consistency due to potential underestimation of reliability (Malkewitz et al., 2023; Zakariya, 2022). One measure that is commonly recommended as an alternative is coefficient omega (McDonald, 1999). Thus, we also provided coefficient omega at two time points (baseline, post-intervention) for reference: Positive Attitudes (0.58, 0.81), Rational Problem Solving (0.80, 0.88), Negative Attitudes (0.50, 0.88), Impulsivity (0.64, 0.79), and Avoidance (0.79, 0.63). The reliability of each subdomain was assessed using Cronbach’s alpha and McDonald’s omega using psych package (Revelle, 2024).

Analysis plan

Descriptive statistics characterized outcomes for participants who completed, or started but did not complete, the allocated intervention. Categorical data were reported as proportions in each response category. T-tests and bootstrap t-tests (e.g. Total CESD, Total mastery, Total UCLA, Social support, Emotional/information support, Affection support, Tangible support, Positive social interaction support, and Burden [short version]) were used to compare continuous outcomes. Fisher’s exact test of independence compared categorical responses between completers and withdrawers.

Supplementary Table 2 shows the variables of interest based on the prior literature as outlined in the Introduction. The key outcomes of interest were depression and caregiver mastery (Fleisher et al., 2023; Huang, 2022; Scerbe et al., 2023), and the mechanisms of action were social support (including emotional/information support, affection support subscale, tangible support subscale, and positive social interaction) (Nemcikova et al., 2023; Ruisoto et al., 2020), problem solving skills (Nezu et al., 2010), and self-efficacy (Scerbe et al., 2023). Loneliness, stress, and caregiver burden (Galvin, 2019; Scerbe et al., 2023) were also examined as contextual factors.

Longitudinal data were analyzed across three time points using Cohen’s d to quantify effect sizes for psychosocial, social, and problem solving skills, focusing on changes from Time 1 to Time 2 and Time 1 to Time 3. To address potential data non-normality and provide robust estimates, we employed bootstrapping with 1,000 resamples using the boot package in R (Canty & Ripley, 1999; Davison & Hinkley, 1997) to determine the 95% confidence intervals of the effect sizes.

Linear mixed-effects models (LMMs) with restricted maximum likelihood estimation via the lme4 package in RStudio (Bates et al., 2015) were used to analyze the impact of demographic factors on the intervention outcomes. Each model included a within-subject factor time (levels: T1, T2, and T3). For each outcome, both unadjusted and adjusted models were reported. p-value <0.05 were considered significant.

Results

The study sample included 39 participants, with 29 completing the study and 10 withdrawing. There were no significant differences in age (completers: M = 67.69; withdrawals: M = 67.4, p = 0.94), gender distribution (women: completers = 72.4%, withdrawals = 60%, p = 0.69), education levels (graduate-level education: completers: 31%, withdrawals: 60%, p = 0.13), ethnicity (white: completers = 93%, withdrawals = 100%, p = 1), or the relationship status (spouse/partner: completers = 86%, withdrawals = 90%, p = 1). Psychological measures indicated no significant difference in depression scores (completers: M = 13.03; withdrawals: M = 17.6, p = 0.28), self-efficacy scores (completers: M = 19.90; withdrawals: M = 19.7, p = 0.86), or caregiver burden scores between completers and withdrawals (completers: M = 20.24; withdrawals: M = 20.9, p = 0.79) (see Table 1).

Table 1.

Means and standard deviations on outcome measures at baseline for the full sample, study completers, and withdrawals.

	Full sample (n = 39)	Complete (n = 29)	Withdraw (n = 10)	P (complete vs. withdraw)	P (complete vs. withdraw) bootstrap t test
Age M (SD)	67.62(9.71)	67.69(8.66)	67.4(12.83)	p = 0.94
Gender				p = 0.69
Women (%)	27 (69%)	21 (72.4%)	6 (60%)
Men (%)	12 (31%)	8 (27.6%)	4 (40%)
Education				p = 0.13
High school or associate degree	8 (21%)	6 (21%)	2 (20%)
Baccalaureate	16(41%)	14 (48%)	2 (20%)
Graduate school	15(38%)	9 (31%)	6 (60%)
Ethnicity				p = 1
White	37 (95%)	27 (93%)	10 (100%)
Other	2 (5%)	2 (7%)	0 (0%)
Relationship				p = 1
Child	5 (13%)	4 (14%)	1 (10%)
Spouse/partner	34 (87%)	25 (86%)	9 (90%)
Psychosocial outcomes
Total CESD	14.21(9.33)	13.03(8.10)	17.6 (12.09)	p = 0.19	p = 0.28
Self-efficacy	19.85(3.03)	19.90(3.07)	19.7(3.06)	p = 0.86
Total mastery	21.16(4.54)	22.41(3.86)	17.11(4.34)	p = 0.001**	p = 0.01*
Total PSS10	17.36 (6.49)	16.38(6.36)	20.2 (6.32)	p = 0.11
Total UCLA	5.51(1.88)	5.10(1.65)	6.7(2.06)	p = 0.018 *	p = 0.067
Social support	30.97(8.30)	32.52(7.89)	26.5(8.17)	p = 0.046 *	p = 0.063
Emotional/information	14.38(4.12)	14.83(4.14)	13.1(3.98)	p = 0.26	p = 0.26
Affection	3.69 (1.30)	3.93 (1.25)	3 (1.25)	p=0.0496*	p = 0.047*
Tangible	6.23(2.30)	6.52(2.20)	5.4(2.50)	p = 0.19	p = 0.24
Positive social interaction	6.67(2.26)	7.24(2.12)	5(1.89)	p = 0.005	p = 0.005*
Burden (short version)	20.41 (7.38)	20.24(7.62)	20.9 (6.98)	p = 0.81	p = 0.79

Note.

^*p < 0.05,

^**p < 0.01,

^***p < 0.001.

However, completers had significantly higher mastery scores (completers: M = 22.41; withdrawals: M = 17.11, P = 0.01), and the difference in loneliness scores approached significance (completers: M = 5.10; withdrawals: M = 6.7, P = 0.067). Social support measures showed completers had significantly higher affection support (completers: M = 3.93; withdrawals: M = 3, P = 0.047), higher positive social interaction scores (completers: M = 7.24; withdrawals: M = 5, P = 0.005), and higher and almost significant overall social support (completers: M = 32.52; withdrawals: M = 26.5, P = 0.06). A flow diagram illustrating the participant flow is provided in Figure 1.

Figure 1.

Participant flowchart.

Table 2 showed effect sizes of various psychological and social support variables from pre-intervention to post-intervention and follow-up among participants. Significant reductions in CESD scores were observed from pre-intervention (M = 14.21) to follow-up (M = 9.10, d = −0.54, 95% CI: −0.86, −0.23). Burden scores decreased significantly from pre-intervention (M = 20.41) to post-intervention (M = 17.38, d = −0.31, 95% CI: −0.60, −0.03). PSS10 decreased from pre-intervention (M = 17.36) to post-intervention (M = 14.63, d = −0.30, 95% CI: −0.69, 0.06) and follow-up (M = 13.72, d = −0.45, 95% CI: −0.88, −0.07). Social support increased significantly from pre-intervention (M = 30.97) to post-intervention (M = 34.81, d = 0.44, 95% CI: 0.17, 0.76) and follow-up (M = 35.90, d = 0.46, 95% CI: 0.10, 0.85). Among social support, tangible support increased significantly from pre-intervention (M = 6.23) to post-intervention (M = 7.0, d = 0.29, 95% CI: 0.04, 0.61) and follow-up (M = 7.48, d = 0.42, 95% CI: 0.10, 0.83); emotional/information support increased significantly from pre-intervention (M = 14.38) to post-intervention (M = 16.38, d = 0.50, 95% CI: 0.20, 0.83) and follow-up (M = 16.41, d = 0.43, 95% CI: 0.044, 0.80). In problem-solving skills, positive attitudes improved from pre-intervention (M = 18) to post-intervention (M = 19.31, d = 0.32, 95% CI: −0, 0.72); negative attitudes significantly decreased from pre-intervention (M = 11.4) to post-intervention (M = 9.34, d = −0.63, 95% CI: −1.03, −0.33).

Table 2.

Effect sizes within intervention.

Variable	Pre-intervention	Post-intervention	Follow up	Pre-post effect size Cohen’s d	Pre-follow-up effect size Cohen’s d (n = 29)
	Mean (SD) (n = 39)	Mean (SD) (n = 32)	Mean (SD) (n = 29)	(n = 32)
Total CESD	14.21 (9.33)	10.31 (5.70)	9.10 (5.63)	−0.36 (−0.75, 0.023)	−0.54 (−0.86, −0.23)
Total self-efficacy	19.85 (3.03)	20.5 (2.94)	20.93 (3.08)	0.17 (−0.20, 0.59)	0.34 (−0.07, 0.80)
Total mastery	21.16 (4.54)	21.78 (4.46)	21.17 (4.56)	−0.04 (−0.36, 0.29)	−0.29 (−0.745, 0.08)
Total PSS10	17.36 (6.49)	14.63 (5.14)	13.72 (5.41)	−0.30 (−0.69, 0.06)	−0.45 (−0.88, −0.07)
Social support	30.97 (8.30)	34.81 (6.98)	35.90 (6.38)	0.44 (0.17, 0.76)	0.46 (0.10, 0.85)
Emotional/Information	14.38 (4.12)	16.38 (3.06)	16.41 (2.98)	0.50 (0.20, 0.83)	0.43 (0.044, 0.80)
Tangible	6.23 (2.30)	7(2.81)	7.48 (2.35)	0.29 (0.044, 0.61)	0.42 (0.10, 0.83)
Positive social interaction	6.67 (2.26)	7.28 (2.22)	7.93 (1.87)	0.19 (−0.12, 0.54)	0.34 (−0.09, 0.78)
Affection	3.69 (1.3)	4.16 (1.17)	4.07 (1)	0.26 (−0.15, 0.68)	0.12 (−0.33, 0.54)
Burden (short version)	20.41 (7.38)	17.38 (7.53)	18.21 (8.05)	−0.31 (−0.60, −0.03)	−0.26 (−0.61, 0.043)
UCLA Loneliness	5.51 (1.88)	5.19 (1.64)	4.86 (1.79)	−0.07 (−0.34, 0.23)	−0.14 (−0.53, 0.23)
Problem-solving	Time 1 (n = 35)	Time 2 (n = 32)	Effect size (n = 32)
Positive attitudes	18.00 (3.19)	19.31 (3.99)	0.32 (−0.00, 0.72)
Rational problem solving	18.31 (3.40)	19.28 (4.05)	0.25 (−0.06, 0.59)
Negative attitudes	11.40 (2.88)	9.34 (3.05)	−0.63 (−1.03, −0.33)
Impulsivity	9.60 (2.79)	9.78 (2.43)	0.02 (−0.26, 0.38)
Avoidance	10.91 (3.65)	10.56 (3.14)	−0.12 (−0.44, 0.23)

The following outcomes showed little change throughout the study. Self-efficacy showed minimal improvement, increasing from pre-intervention (M = 19.85) to post-intervention (M = 20.5, d = 0.17, 95% CI: −0.20, 0.59) and follow-up (M = 20.93, d = 0.34, 95% CI: −0.07, 0.80). Mastery showed almost no change from pre-intervention (M = 21.16) to post-intervention (M = 21.78, d = −0.04, 95% CI: −0.36, 0.29), and follow-up (M = 21.17, d = −0.29, 95% CI: −0.75, 0.08). UCLA loneliness scores showed minimal improvement, decreasing from pre-intervention (M = 5.51) to post-intervention (M = 5.19, d = −0.07, 95% CI: −0.34, 0.23) and follow-up (M = 4.86, d = −0.14, 95% CI: −0.53, 0.23). In problem-solving skills, rational problem-solving showed minimal increase from pre-intervention (M = 18.31) to post-intervention (M = 19.28, d = 0.25, 95% CI: −0.06, 0.59). Impulsivity and avoidance showed minimal changes with effect sizes d = 0.02 and d = −0.12, respectively.

The linear mixed modeling results (Table 3) indicate significant changes in several psychological and social support variables from pre-intervention to post-intervention and follow-up. For CESD scores, significant reductions were observed at both post-intervention (B = − 3.15, p = 0.012) and follow-up (B = −4.49, p < 0.001) in the adjusted model. Burden scores also showed significant reductions at post-intervention (B = − 2.59, p = 0.02) and follow-up (B = − 2.25, p = 0.048) in the adjusted model. PSS10 scores decreased significantly at post-intervention (B = −2.25, p = 0.025) and follow-up (B = −3.19, p = 0.003). Social support increased significantly at post-intervention (B = 3.56, p = 0.001) and follow-up (B = 3.77, p = 0.001) in the adjusted model. Emotional/information support and tangible support also saw significant increases. Positive attitudes towards problem-solving improved significantly at post-intervention (B = 1.25, p = 0.04), while negative attitudes showed a significant decrease (B = −1.95, p < 0.001). Non-significant outcomes included self-efficacy, which showed no significant changes at post-intervention (B = 0.58, p = 0.29) and follow-up (B = 1.06, p = 0.062). Mastery, UCLA loneliness, impulsivity, avoidance, and rational problem-solving did not show significant changes across the time points.

Table 3.

The linear mixed modeling results.

Parameter	Unadjusted model		Adjusted model		Unadjusted model		Adjusted model
	B	P	B	P	B	P	B	P
CESD					Burden (12-item version)
Intercept	14.21	<.001	13.65	0.12	20.41	<.001	23.49	0.01
Post	−3.10	0.01	−3.15	0.012	−2.57	0.02	−2.59	0.02
Follow up	−4.44	0.00	−4.49	<.001	−2.24	0.049	−2.25	0.05
Age			−0.03	0.79			−0.06	0.64
Female			3.83	0.12			1.12	0.66
PSS10					Mastery
Intercept	17.36	<.001	20.15	0.00	21.16	<.001	24.57	<.001
Post	−2.17	0.03	−2.25	0.02	0.16	0.83	0.15	0.84
Follow up	−3.12	0.00	−3.19	0.00	−0.62	0.42	−0.63	0.42
Age			−0.07	0.42			−0.04	0.61
Female			2.75	0.13			−1.30	0.39
Social support					Self-efficacy
Intercept	30.97	<.001	27.03	<.005	19.85	<.001	17.30	<.001
Post	3.56	0.001	3.56	0.001	0.58	0.28	0.58	0.29
Follow up	3.78	<.001	3.77	<.001	1.07	0.057	1.06	0.062
Age			0.055	0.66			0.03	0.55
Female			0.34	0.89			1.02	0.27
UCLA loneliness					Problem subscale: positive attitudes
Intercept	5.51	<.001	7.12	0.002	18	<.001	16.22	0.001
Post	−0.2	0.46	−0.21	0.44	1.24	0.04	1.25	0.04
Follow up	−0.41	0.14	−0.41	0.14
Age			−0.03	0.32			0.02	0.78
Female			0.46	0.44			0.81	0.52
Problem: rational problem solving					Problem: negative attitudes
Intercept	18.31	<.001	12.61	<.01	11.4	<.001	15.65	<.001
Post	0.97	0.12	0.98	0.11	−1.93	<.001	−1.95	<.001
Age			0.07	0.30			−0.05	0.30
Female			1.73	0.18			−0.79	0.45
Problem subscale: impulsivity					Problem subscale: avoidance
Intercept	9.6	<.001	13.88	<.001	10.91	<.001	17.07	<.001
Post	0.11	0.79	0.09	0.82	−0.38	0.53	−0.41	0.50
Age			−0.06	0.17			−0.09	0.13
Female			0.01	0.99			−0.37	0.75

Discussion

Our intervention assessed various dimensions of the caregiver experience, including burden, mastery, perceived stress, social support, and problem-solving skills. Participants who completed the study demonstrated higher mastery, reduced loneliness, and increased social support, suggesting that caregivers with stronger support networks and better coping resources might be more likely to complete the intervention. Significant improvements were observed in caregiver burden, depression, and perceived stress, social support, emotional/information support, tangible support, and both negative and positive attitudes in problem-solving, in contrast, mastery scores, self-efficacy, and certain aspects of social support and problem-solving saw only marginal gains. These findings indicate that while our intervention successfully enhanced caregivers’ emotional well-being, there are opportunities to further strengthen the skill-building and psychoeducational components to promote more substantial improvements in mastery and self-efficacy.

A significant decline was observed in stress, burden, and depression, while problem-solving and social support scores showed improvement, aligning with existing research that suggests problem-solving interventions are effective for reducing caregiver depression, anxiety, and burden (Fitzpatrick et al., 2013; Kiosses & Alexopoulos, 2014; Kirkham et al., 2016; Tao & Zhang, 2019; Washington et al., 2018). Our intervention effects were comparable to community-based REACH II intervention, which showed effect sizes of −0.42 for depression, −0.52 for caregiver burden, and 0.28 for social support (Cho et al., 2019). The reductions in these variables are clinically meaningful, given that nearly 40% of caregivers for those with moderate-advanced LBD report high burden and/or depression (Armstrong et al., 2024). Our effect sizes are larger than those of typical internet-based programs (Zhao et al., 2019) and are comparable to cholinesterase inhibitors in reducing LBD behavioral symptoms, with a mean NPI change score of −1.73 (effect size ~ −0.49) (Meng et al., 2019). Significant improvement was observed in two subdomains of social support: emotional/informational support and tangible support (e.g. someone to help with daily chores or take you to the doctor). These findings highlight the importance of tailored, interactive interventions in fostering meaningful connections among caregivers. In exit interviews, participants mentioned that they particularly appreciated the opportunity to connect with and form a community with others in similar situations, sharing experiences, ideas, and resources. For example, caregivers whose care recipients had less advanced LBD found value in learning about potential future challenges from case studies and peers. As one participant noted: “I think [the program] gave me a lot more knowledge about how to potentially deal with some situations. But also I kind of felt that just given the state of my wife’s condition, that it’s going to be more valuable as it gets worse and in the future, so it gave me a little bit of confidence in that I’d be able to handle it coming up… It gave me some calm now that I’m going to be able to handle it in the future when I really need a lot of these skills.” These shared experiences may have contributed to the observed increases in emotional/informational support and tangible support, as well as participants’ confidence in handling future caregiving challenges.

However, there was no significant improvement in affection support or positive interaction support, which measure aspects such as having someone to share enjoyable activities with or feeling loved and wanted. This may be due to the nature of our group-based, online intervention, which may not foster the same level of personal, positive social interaction or emotional connection. It highlights the importance of assessing LBD caregivers’ specific needs and tailoring interventions to address these needs effectively. Prior research suggests that only support tailored to individual needs is most effective; for example, providing assistance with activities of daily living may not sufficiently alleviate caregiver burden in advanced LBD, where caregiver needs are more diverse and complex (Armstrong et al., 2024).

Despite these improvements, total mastery scores did not significantly increase, similar to previous findings from a 16-week peer mentoring intervention for LBD caregivers that improved knowledge and dementia attitudes but not mastery skills (Fleisher et al., 2023). This may be partly attributed to the complexity of LBD, making it challenging for caregivers to develop the necessary skills and capacities. Ikeda et al. (2024) investigated whether the treatment needs of patients with Dementia with Lewy bodies (DLB) differ according to the clinical department visited by patients. They highlighted that physicians across specialties often prescribe different treatments for LBD, reflecting the diverse range of initial symptoms: cognitive dysfunction and psychiatric symptoms in the psychiatric department; cognitive dysfunction, autonomic neuropathy, and psychiatric symptoms in the geriatric department; and parkinsonism and psychiatric symptoms in the neurology department. This complexity to LBD care may contribute to the limited improvement in mastery and self-efficacy among caregivers.

Additionally, caregiver preparedness—defined as the extent to which caregivers feel equipped to handle their caregiving responsibilities (Archbold et al., 1990)—may play an important role in enhancing mastery and self-efficacy. Without a strong sense of preparedness, caregivers may struggle to feel in control or confident, which are crucial elements for developing mastery and self-efficacy. Preparedness involves not only understanding the demands of caregiving but also developing practical skills, such as managing symptoms and accessing appropriate resources. Interventions that focus on building caregiver preparedness alongside mastery and self-efficacy might therefore provide a more comprehensive support structure for LBD caregivers. Another possible explanation for the limited impact on self-efficacy and mastery could be related to the format of the intervention. The online asynchronous format may provide fewer opportunities for interactive problem-solving practice, such as live roleplaying exercises, which are effective for developing coping skills and self-efficacy. Further, mastery involves a more profound sense of control over one’s life and emotions, recovery of self-esteem, and finding satisfaction despite challenges, which are more complex psychological outcomes that likely require more time, sustained effort, and repeated experiences of success to develop (Younger, 1991). Addressing these aspects in future interventions might require a more interactive or intensive format to support skill development in real-time scenarios.

Understanding the mechanisms behind our findings may require further analysis of the relationships involving moderators and mediators. Problem-solving and social support could act as mediators or moderators between caregiver experiences and outcomes like burden, depression, and stress, as problem-based theory suggests that improved reactions to stress through these skills can reduce stress (Nezu et al., 2010). While we did not find a significant relationship between mastery and outcomes, prior studies indicate that self-efficacy, social support, and problem-solving are mediators of depression in caregivers (Tang et al., 2015). Mastery has also been linked to reduced burden, anxiety, depression (Chan et al., 2018) and distress in older caregivers (Kabia et al., 2022). Intervention components that reinforce mastery might need to be incorporated via practicing long-term skill-building.

Limitations

The study has several limitations that highlight areas for future research. One limitation is that, given the nature of this pilot study and the absence of a control group, the findings are preliminary and exploratory. Another limitation is the absence of a control group, which limits the ability to rigorously compare outcomes and assess the intervention’s effectiveness. Including a control group in future studies would provide stronger evidence of the intervention’s impact. This exploratory study focused on collecting a range of psychobehavioral measures important to caregivers. As such, partitioning into primary and secondary outcomes was deprioritized in favor of gaining a comprehensive picture of the intervention’s effects and potential mechanisms of change. While the sample size was calculated based on the power analysis of exploratory mastery variable, which may limit sufficiency for other outcomes, this design allowed for a broader investigation, revealing potential signals mechanisms that might have been overlooked with a more targeted approach. Furthermore, the study’s participants were drawn from a research registry and required internet access. While approximately 70% of older adults currently have internet access (Anderson & Perrin, 2017), those without it may respond differently to the intervention, potentially affecting the generalizability of the findings. Another significant limitation is the lack of racial and ethnic diversity among participants. It is important for future research to investigate how caregivers of individuals with LBD from diverse racial and ethnic backgrounds experience the VOCALE-LBD intervention.

Conclusion

This pilot study demonstrates the feasibility and efficacy of the VOCALE-LBD web-based asynchronous intervention in supporting caregivers of individuals with LBD, highlighting its clinical importance in improving mental health in an underserved population. To build on these promising results, future research should include a control group for more rigorous evidence, explore the long-term effects and scalability of the intervention, and incorporate a more diverse caregiver population to enhance the generalizability and applicability of the findings. In addition, withdrawal patterns suggested that additional resources improved retention, providing additional support for caregivers with fewer resources or greater barriers to participation could be instrumental for increasing retention, and thus maximize the impact and reach of future interventions.

Supplementary Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/13607863.2025.2462758.