We recruited participants from the Bedford, Massachusetts, and Manchester, New Hampshire, Department of Veterans Affairs Medical Centers and surrounding communities using posted informational flyers. We also reviewed Veterans Affairs (VA) medical records to identify veterans with dementia and mailed recruitment letters to their homes and followed up with phone calls. Criteria for study inclusion were (1) providing informal care for someone with self-reported or physician-confirmed diagnosis of dementia, (2) at least 1 caregiver-reported NPS for the care recipient at the time of enrollment, (3) provision of a minimum of 5 hours per week of direct caregiving, and (4) access to a computer or tablet and internet or Wi-Fi. Participants could be caregivers of veterans with dementia or veterans caring for someone with dementia.

Pilot and feasibility work typically aims to explore recruitment, engagement, and practicality rather than statistical power [30,31]. Therefore, for this pilot study, the focus was on testing feasibility and acceptability rather than testing hypothesis-driven significance with associated power calculations for sample size. To accomplish these goals, we selected a target enrollment of 24 caregivers.

This study was approved and monitored by the VA Bedford Health Care System Institutional Review Board (study no. 1598706). Written informed consent was obtained from all study participants. Privacy and confidentiality of all research participants’ data and identities were maintained using multiple methods, including using a Health Insurance Portability and Accountability Act–compliant web-based platform with end-to-end encryption, coding data so that only the research team had access to identifiable information, storing data securely behind the VA firewall and restricting access to the study team, analyzing coded data and reporting data in aggregate form, and establishing ground rules during group sessions to help participants protect their privacy. Participants received US $25 for completing each of the baseline and postintervention follow-up visits, and US $50 for completing the 3-month follow-up visit, receiving a possible total of US $100 to compensate for their time.

Initially, we intended to conduct preintervention, in-person sessions to obtain written consent and complete baseline assessments, as well as to provide in-person instruction about technology use. However, prior to the start of the study, the COVID-19 pandemic forced us to convert all study visits to a virtual format. We added a preintervention live video session to provide technology instruction before the start of groups. Individual assistance to help participants access the initial tutorial was provided as needed by phone. Follow-up data collection visits were conducted by video immediately and 3 months post-treatment. Participants were compensated for their time.

A trained interviewer met with each eligible caregiver via video to explain the study, answer questions, review the Health Insurance Portability and Accountability Act authorization, and sign the informed consent forms. The interviewer also provided initial training on using the videoconferencing platform. During the same visit or a second visit, depending on participant preference, the interviewer administered the baseline questionnaires. Questions were presented visually via PowerPoint (Microsoft Corp) one at a time and read aloud. Participant responses were recorded in VA REDCap (Research Electronic Data Capture), a secure data management software, in real-time. After baseline assessment, a printed TeleCARE manual was mailed to the participant. All caregivers began the intervention within 6 weeks of baseline assessment. The collection of immediate and 3-month follow-up data followed the same procedures as baseline data collection. Additional supplemental visits took place by video or phone as needed to help participants troubleshoot technology issues. All video visits were conducted via Health Insurance Portability and Accountability Act–compliant, encrypted, VA-approved videoconferencing platforms.

TeleCARE was adapted from the original manualized CARE intervention [15]. As in CARE, TeleCARE groups met weekly for 90-minute sessions. At the beginning of each session, a doctoral-level occupational therapist leading the groups went through a series of prompts specific to telehealth to ensure safety and confidentiality during the session. Participants were given time during each session to discuss homework, troubleshoot any problems that may have arisen when attempting to use new skills, and offer each other support and advice.

The intervention was delivered to caregivers in 6 groups that ran sequentially from June 2021 to September 2022. We started with an initial TeleCARE test group (TeleCARE-test) to troubleshoot any issues before the formal pilot. TeleCARE-test participants reported that the 5 sessions did not provide adequate time to cover all the intervention material and allow for discussion. In response, we reviewed all intervention sessions to determine the number of additional sessions needed. The finalized program was extended to 7 sessions. Table 1 provides a summary of each session included in the final manualized intervention. Other intervention-specific modifications to adapt the program to the telehealth setting were made based on the experiences of the study team and feedback from the TeleCARE-test group. At the beginning of each session, we added unique steps to ensure participant safety in the telehealth setting. These steps included gathering information about where the participant was located in the event of a psychological (eg, expressed suicidal intent) or medical (eg, cardiac arrest) emergency. Unlike in-person interventions, telehealth interventions require the provider to know where the participant is located if the need arises to alert emergency services (eg, local police). The group leader also asked for updated contact information at the start of each session for participants joining from a new, unknown location. This information was provided to the group leader through secure chat to maintain privacy. The interventionist encouraged all participants to sign into the platform with their first name to facilitate group rapport. It became clear that video-specific etiquette guidelines should be added to the first session, including muting oneself when not speaking, using the raise hand feature, remaining primarily stationary during the group (eg, not walking around with the video device), and reducing background distractors (eg, television, pets, children). Participants were also asked to locate themselves in a suitable space to reduce privacy concerns (ie, a location where they could not be overheard discussing personal information). Finally, the need for broad technological support for older adults enrolled in TeleCARE. In response, we added a technology tutorial session before the intervention started, as noted above, and access to ongoing support as needed.

Following CONSORT (Consolidated Standards of Reporting Trials) guidelines [32] and the proposed guidelines for reporting feasibility study outcomes [33], we collected multiple measures of study feasibility, including recruitment rates, session attendance rates, study completion rates, and a formal rating of the acceptability of the intervention. Given our interest in the telehealth adaptation, we also collected data on technology use. Participants completed outcome measures at baseline, 1 to 2 weeks immediately after the intervention, and 3 months postintervention.

At baseline, caregivers completed a demographic and historical information survey, providing information about themselves and the care recipient. Caregivers also completed [1] the clinical dementia rating sum of boxes [34], a 5-point interview that characterizes 6 domains of the care recipient’s cognitive and functional performance and disease severity.

At baseline, immediate follow-up, and 3-month follow-up, caregivers were asked to report on the presence of the care recipient’s NPS and functional abilities. The Lawton Instrumental Activities of Daily Living (IADL) Scale [35] assesses the functional capacity of the care recipient in 8 domains of functioning. Scores total from 0 to 8, with higher scores indicating greater functional independence. The NPS inventory questionnaire [36] measures both the presence and severity of the care recipient’s NPS and the caregiver’s distress related to NPSs. Caregivers were also asked to complete measures related to their own well-being. These included the Beck Depression Inventory-II [37], a 21-item scale to assess the severity of depressive symptoms, and the Beck Anxiety Inventory [38], a 21-item scale to assess the severity of anxiety. The revised Caregiver Self-Efficacy Scale [39] is a 15-item scale exploring caregiver self-efficacy in obtaining respite, responding to patients’ disruptive behaviors, and controlling upsetting thoughts. The positive aspects of caregiving (PAC) questionnaire [40] consists of nine items that load onto two factors, self-affirmation and outlook on life, and assesses positive reflections on caregiving. The Relative Stress Scale [41] consists of 15 items and can be divided into 3 subgroups of emotional distress, social distress, and negative feelings. The meaning in life questionnaire [42] is a 10-item self-report assessment designed to measure the presence of and search for meaning in life.

At immediate follow-up only (ie, immediately after completing the TeleCARE intervention), participants also completed the client satisfaction questionnaire (CSQ-8) [43] to capture the acceptability of the intervention.

The primary objective of this pilot study was to assess the acceptability and feasibility of TeleCARE. Descriptive statistics were used because the goal was to explore trends rather than test hypotheses, as appropriate for the pilot stage of the current project. As such, we explored numerical trends and effect sizes of outcome variables. Descriptive statistics are reported to characterize the sample and summary scores for each measurement time point (baseline, post-treatment, and a 3-mo follow-up). Cohen d effect sizes were calculated to compare changes from baseline to post-treatment and from baseline to 3-month follow-up. All statistical analyses were conducted using SAS software (version 9.4; SAS Institute Inc). Given the early stage of this investigation, data from the TeleCARE test group were excluded; analyses only included participants in groups 2 to 6 who completed the final format of the TeleCARE intervention (n=17). An intent-to-treat approach was not used for this preliminary feasibility investigation.

Participants included 17 White caregivers (100%), who were predominantly female (n=13, 76%) and spouses (n=14, 82%), with a mean age of 71.8 (SD 12.58) years. All caregivers lived with the care recipient (n=17, 100%), and nearly all caregivers (n=16, 94%) reported providing the majority of care. Care recipients had a mean age of 77.7 (SD 4.84) years, were White (n=17, 100%), male (n=15, 88%), and had a mean clinical dementia rating sum of boxes score of 10.8, indicative of moderate dementia severity [44]. See Table 2.

Of the 109 caregivers contacted for recruitment, 24 did not respond, 15 did not meet eligibility criteria, 38 declined participation, and 8 had other reasons (Figure 1). Among those who declined (38/109, 35%), the most common reason was that they felt that the intervention was not needed (20/38, 53%). The second most common reason was lack of time (11/38, 29%). The remaining provided no reason (3/38, 8%), said they were not interested (3/38, 8%), or reported a technological barrier (1/38, 3%). Of those that gave other responses (8/109, 7%), 5 said that their loved one was deceased when we contacted them, 2 had been admitted to long-term care, and 1 had recently been diagnosed with cancer. Overall, 24 caregivers (24/109, 22%) were enrolled in the study, with 4 in the TeleCARE-test group 1 and 20 in intervention groups 2 to 6. Across groups 1 to 6, 20 caregivers (20/24, 83%) completed the study. Of the 4 noncompleters, 1 was from the TeleCARE-test group 1, and 3 were from groups 2 to 6. Two of 4 noncompleters withdrew because they disliked the group setting. The remaining 2 noncompleters had medical issues arise that prevented continued participation (1 during the intervention and 1 prior to the start of the intervention but after study enrollment). Because the TeleCARE test group 1 was treated as a test group and subsequent changes were made to the intervention after their feedback, completers from TeleCARE test group 1 (3/4) were not included when reporting intervention attendance or exploratory quantitative analysis presented below. Of those within groups 2 to 6 (n=17), 10 attended all 7 sessions (10/17, 59%), 5 attended 6 of 7 sessions (5/17, 29%), and 2 attended 5 of 7 sessions (2/17, 12%). Occasionally, a participant arrived late, most commonly due to technological difficulties.

To assess the acceptability and feasibility of the telehealth adaptation, data were collected to better understand participant experiences related to the use of technology. Overall, 11 of 17 participants (65%) required some degree of technology support. As mentioned above, a study staff member was assigned to provide personalized technology assistance as needed. Examples of when assistance was needed included helping participants turn on the power if a tablet was provided, access their email and click the video-meeting link, retrieve and/or reset their password, turn on their camera, use the mute and unmute function, and use the raise hand feature. One study participant had no prior experience with technology and needed assistance at all levels, from setting up an email account to using the device. The largest amount of technological assistance was required at session 1, dropping considerably by session 2. Participants needed less support with each passing session, although some continued support remained necessary at almost all time points. Participants reported feeling a sense of accomplishment and increased comfort with technology at the conclusion of the study. At the 3-month follow-up visit, 3 of 17 participants (18%) required assistance with technology, which was an increase compared with session 7 of the intervention. This finding suggests that most study participants were able to retain the taught skills, even after 3 months, while a minority of participants experienced a decay in skills. Figure 2 provides a visual depiction of the amount of time spent providing technological assistance across all sessions for intervention groups 2 to 6.

Smith et al [45] proposed the following guidelines for the interpretation of overall CSQ-8 scores into 4 levels of client satisfaction: “poor” (scores 8‐13), “fair” (scores 14‐19), “good” (scores 20‐25) and “excellent” (scores 26‐32). The mean CSQ-8 score was 22.3 (SD 1.08), suggesting good intervention acceptance.

As mentioned previously, this study was not adequately powered to detect statistical significance, as appropriate for the pilot stage, and the focus was on identifying numerical trends and examining effect sizes. Table 3 provides the mean scores of measures completed at every time point (n=17; also see Table S1 in Multimedia Appendix 1). Between baseline and immediate post-treatment time points, caregivers reported a decline in the care recipient’s ability to complete IADLs. There was no substantial change in the number of NPS reported by the caregiver, though there was a trend towards a decrease in the reported severity of NPS (Cohen d=0.16). Regarding caregiver psychological well-being, trends between baseline and post-treatment time points included a decrease in depression and anxiety symptoms and an increase in PAC and meaning and purpose. In fact, the largest effect between baseline and post-treatment was a decrease in anxiety. Participants reported little change in relative stress. Interestingly, despite improvements in psychological well-being, participants reported a decrease in caregiving self-efficacy between baseline and post-treatment, with the decrease in self-efficacy having the second largest effect size of the outcome measurements. Except for a statistically significant decrease in caregiver self-efficacy (t₁₆=2.17, P=.045, Cohen d=0.21), no outcomes were statistically significant.

Between baseline and 3-month follow-up, participants reported further decline in IADLs for the care recipient and an increase in NPS. In addition, most improvements reported between baseline and immediate post-treatment were lost by the 3-month follow-up. Participant meaning and purpose in life was the only improvement maintained from post-treatment to the 3-month follow-up. None of the changes between baseline and the 3-month follow-up visits were statistically significant.

In this pilot study, we were able to recruit dementia caregivers and engage them in a telehealth adaptation of CARE (TeleCARE). Twenty caregivers completed TeleCARE, including 3 from an initial test group that informed intervention modifications, leaving 17 caregivers in the final intervention group. Caregivers were predominantly older female spouses of the care recipient. Most enrolled caregivers attended all sessions, suggesting that caregivers who engaged in the program stayed engaged throughout. Likewise, satisfaction with the intervention was good. Preliminary quantitative trends suggested immediate postintervention improvements in caregiver outcomes. Overall, TeleCARE was deemed feasible and acceptable, with preliminary quantitative results suggesting that TeleCARE has the potential for a larger-scale study. Future studies of TeleCARE could harness digital recruitment methods for scalability.

Importantly, intervention adaptations had to be made to successfully implement TeleCARE, including both intervention telehealth-specific modifications and the provision of broad technology support. Related to the modifications, previous research emphasizes the importance of safety and risk management, participant privacy, and platform etiquette in telehealth contexts. Telehealth inherently presents unique challenges compared to in-person care, including limited immediate clinical support and barriers to emergency response due to geographic variability and licensure restrictions [46-48]. Presession safety protocols, such as verifying participant location and emergency contacts, are increasingly recommended to mitigate risks during psychological or medical crises [47]. In addition to safety considerations, given evidence that clinical outcomes in telehealth are often comparable to in-person care, researchers emphasize the importance of replicating in-person standards through structured telehealth delivery [49], such as minimizing distractions, limiting movement, and encouraging stationary participation. Privacy concerns also remain central, with recommendations for participants to join from secure, confidential environments to protect sensitive information and foster psychological safety [48,50].

Compared with younger counterparts, older adults tend to have lower digital literacy, which presents a barrier to engaging in telehealth services [51]. Even after adding an initial (before the start of the intervention) technology tutorial session, the majority of study participants (11/17, 65%) required technological support, and though less support was needed after each passing session, some participants required ongoing support throughout. Digital literacy training programs for older adults have emerged in response to the recognized need to improve access to technology-based services [52-54]. Many studies have found that older adults want to engage with technology and not feel left behind as digital opportunities increase [55]. At study completion, participants reported feeling a sense of accomplishment related to learning to use the necessary technology and increased comfort with technology, highlighting the ability of older adults to engage with technology with individualized and ongoing support.

It is worth noting that half of the caregivers who declined to join the group stated that they did so due to a lack of time, despite the program being less than a 2-month, 90-minute weekly commitment offered via telehealth, thereby minimizing the time needed to travel. Dementia caregivers consistently identify time constraints as a major barrier to participating in interventions [56-58]. Many caregivers report prioritizing the needs of the care recipient over their own and neglecting self-care, thereby reducing opportunities to participate in supportive services [59-61]. Recent research suggests that brief, flexible, telehealth-based interventions may help boost caregiver engagement [62]. The current findings suggest that offering interventions via telehealth may be only one small step toward improving intervention engagement and that brevity and flexibility may play a bigger role in helping to engage busy caregivers. In addition, finding innovative ways to increase respite time for caregivers so they can engage in supportive services offers promise. For example, Iacob and colleagues [63] demonstrated that an online scheduling and planning program with virtual coaching and education helped increase caregiver respite [63].

Although the study was not powered to statistically test changes between pre- and posttreatment, quantitative results suggested trends toward a decrease in NPS severity in care recipients, a decrease in caregiver depression and anxiety, and an increase in caregiver-expressed PAC and meaning and purpose in life after intervention completion. However, most improvements were not sustained by the 3-month follow-up. Also of note is the reduction in caregiving self-efficacy following intervention engagement, which may be unexpected. However, when learning new skills, individuals often experience initial failures early in the learning process. Although these failures are part of the growth necessary to learn new skills, failures can lead to temporary reductions in self-efficacy [64]. It is possible that caregivers in TeleCARE experienced reduced self-efficacy related to their initial unsuccessful efforts applying new skills. Caregivers may also have attributed their successes largely to the presence of the interventionist, causing self-efficacy to decline postintervention when transitioning to applying skills independently. In addition, skills may decay between the conclusion of an intervention and longer-term follow-up (in the current study, 3 mo), which can lead to further reductions in self-efficacy [65,66]. Future studies aimed at better understanding the conditions for gains in self-efficacy and sustainment of those gains are warranted. In addition, future trials of TeleCARE could incorporate theoretically driven self-efficacy enhancement techniques based on Bandura’s four identified self-efficacy enhancing methods (mastery experience, modeling, social persuasion, altering emotional/somatic states) [67], while retaining the basic intervention elements.

The major limitation of the current study is the homogeneous sample, which reduces generalizability. All of the participants in this study were white, mostly female, and primarily spouses of the individuals receiving care, with at least some college education. This fact limits the applicability of our findings to caregivers from diverse ethnic, cultural, and socioeconomic backgrounds.

Studies have found that the perception of caregiver differs for different cultures and, for those with more deprived socioeconomic disadvantage, knowledge about dementia may be low [68-70]. Intervention adaptations that take into account ethnic or cultural considerations and differences related to medical privacy, preferences for keeping the person with dementia at home, the role of religion, and the use of holistic or non-Western medicine approaches to care would be important to consider in future trials of TeleCARE [71-73]. TeleCARE may need to be adapted for caregivers with lower levels of education and lower income associated with limited or inadequate health literacy [74,75]. Of course, adapting the intervention for non-English speaking populations would improve reach. The original CARE intervention was successfully culturally and linguistically tailored for a Latino population and could serve as a basis for a TeleCARE adaptation in future trials [76]. Importantly, and aligned with the focus of telehealth adaptations, racial and ethnic minority caregivers and those with lower educational attainment are less likely to have access to technology (eg, stable internet connectivity, computer, smartphone) and the skills needed to engage with technology, and specific efforts to explore the use of telehealth for diverse cohorts of dementia caregivers is critical [77-80]. Building community partnerships for intervention adaptation, outreach, and recruitment can help strengthen future trials.

The predominantly female participants in the current intervention limit our ability to fully capture male caregivers’ experiences. Male caregivers may differ in their use of coping strategies, emotional responses, and support-seeking behaviors, with literature suggesting a greater tendency to use problem-focused strategies, seek help earlier, and prioritize having time for themselves, engaging in personal activities that may buffer caregiving stress [81,82]. As caregiving intensity increases, male caregivers may experience higher burden than their female counterparts, who often develop caregiving-related stress earlier in the process [81-83]. Thus, interventions that have predominantly female participants may differ significantly from a predominantly male group. The current study also included predominantly spousal caregivers. Adult child caregivers face distinct challenges, including the need to balance caregiving responsibilities with employment and parenting obligations [84,85]. Future research should continue exploring these subgroup differences to tailor interventions accordingly.

In summary, we successfully adapted an existing evidence-based, in-person, manualized dementia caregiver intervention (CARE) for synchronous video telehealth demonstrating good feasibility and acceptability. Successful adaptation required revising the manualized protocol to include procedures to ensure participant safety, etiquette instructions to improve group experience, and technological support to increase accessibility. These findings suggest that when attempting to transition an in-person program to video telehealth, these broad program-nonspecific adaptations should be considered. The important need for ongoing, available technological assistance was particularly salient.