Caregiver Experiences With an Internet-Delivered Insomnia Intervention: SHUTi-CARE Trial Primary Qualitative Analysis

Abstract

Background

Digital health interventions show potential to increase caregivers’ access to psychosocial care; however, it is unclear to what extent existing interventions may need to be tailored to meet caregivers’ unique needs.

Purpose

This study aimed to determine whether—and if so, how—an efficacious Internet-delivered insomnia program should be modified for caregivers. The generalizability of these findings beyond the tested program was also examined.

Methods

Higher-intensity family caregivers (N = 100; age M = 52.82 [SD = 13.10], 75% non-Hispanic White, 66% ≥college degree) received access to an Internet-based cognitive-behavioral therapy for insomnia (CBT-I) program. Participants who completed one or more intervention “Cores” provided open-ended feedback on their experience; nonusers (completed no Cores) shared their barriers.

Results

Most caregivers who used the program (n = 82, 82%) found it feasible, citing its user-friendly, fully automated online format. Many reported that CBT-I strategies were helpful, although some faced challenges in implementing these strategies due to the unpredictability of their caregiving responsibilities. Opinions were divided on the utility of tailoring the program for caregivers. Nonusers (n = 18, 18%) primarily cited concerns about time burden and lifestyle compatibility as usage barriers.

Conclusions

Delivering fully automated behavioral interventions through the Internet appears suitable for many caregivers. Extensive tailoring may not be required for most caregivers to benefit from an existing online CBT-I program, although additional guidance on integrating CBT-I strategies in the context of challenging sleep schedules and environments may help a subset of caregivers. Future research should explore how such tailoring may enhance digital health intervention uptake and effectiveness for caregivers.

Most family caregivers found an online insomnia program feasible and helpful even without caregiver-specific tailoring.

Introduction

Roughly 53 million adults in the USA provide essential practical, emotional, and/or medical support to a loved one with a chronic illness or disability [1]. Many of these caregivers experience significant mental and physical health strains as a result of this role [2–4]. There is significant promise for digital health to improve the accessibility and scalability of behavioral interventions for caregivers [5–7]. Realizing the full potential of digital health to positively impact caregivers’ health will require understanding how Internet interventions can be made the most efficacious for the most caregivers—an objective identified as a high priority in caregiving research [8–11].

To date, most digital health intervention research among caregivers has developed and tested interventions targeted to caregivers within specific disease contexts, like cancer or dementia [12–17]. This kind of tailoring to a specific target population (e.g., an intervention for dementia caregivers) is intended to improve the interventions’ pertinence to that population. According to behavior change theory, tailoring increases information salience, which leads to greater attention to the information and motivation to enact behavior change [18]. These gains are expected to improve intervention efficacy and therefore its overall impact, despite narrowing the intervention’s reach (e.g., excluding cancer caregivers). Meta-analyses of the direct effects of tailoring from empirical studies have found only small gains in outcomes relative to generic materials. These effects tend to be sustained, but decaying, out to 18 months post-intervention [19, 20].

Determining how to sufficiently tailor an intervention is important, given that restricting reach to too narrow of a target population risks unnecessarily limiting intervention impact across the full population of caregivers. Caregivers who might have benefitted from an intervention but do not find the tailored content pertinent to their context can be alienated and disengaged from treatment [6]. While most digital health programs for caregivers studied to date have been targeted to caregivers for people with dementia, aging adults, or people treated for cancer [10, 21], these groups collectively represent only about half of all caregivers [1]. This means that about half of all caregivers providing care in less-studied contexts would be unlikely to find interventions pertinent to their situations.

Indeed, certain interventions may not require tailoring to caregiving at all. Many of the factors associated with caregivers’ positive perceptions of a digital health program’s utility and appropriateness for their lifestyle are general best practices. These include providing evidence-based information, having an easy-to-use interface, and mitigating privacy concerns [6]. Such factors are known drivers of engagement with digital health programs in the general population [22, 23]. Moreover, not all individuals who provide care identify strongly with the caregiving role [24], so tailoring may decrease pertinence to these individuals.

On the other hand, certain modifications may be important to ensuring interventions are feasible and appropriate for caregivers. Caregivers have identified providing stories of caregivers’ experiences, information specific to their caregiving context, and having the opportunity to connect with other caregivers as valuable components in digital health interventions [6]. Moreover, caregivers’ unpredictable schedules and their frequent prioritization of care recipient needs over their own can interfere with their ability to commit to and complete interventions [25–27].

The challenge of balancing intervention reach and efficacy for caregivers is clearly present within digital healthcare for insomnia. Insomnia is a prevalent and impairing condition among family caregivers [2, 4]. Fortunately, there are highly effective Internet-delivered insomnia programs [28], such as SHUTi (Sleep Healthy Using the Internet [29, 30]). There is robust evidence that this program is effective in treating insomnia in a wide range of populations without tailoring, having demonstrated strong treatment effects in more than 25 clinical trials across the world, including those with medically diverse populations [31–33]. However, caregiver-specific tailoring may be warranted, as caregivers report unique maladaptive thoughts about the consequences of their insomnia [34–36] and often have unique barriers to implementing behavioral therapeutic techniques to treat insomnia due to nighttime care responsibilities [35–37]. Preliminary data from a feasibility trial of SHUTi without tailoring among cancer caregivers showed that participants reported substantial sleep improvements from the program but were less engaged than users in most prior SHUTi trials [38].

Taken together, the decisions of whether and how to tailor interventions for caregivers are complex, yet there is strikingly limited data to guide these decisions. Toward addressing this limitation, this single-group trial assessed caregivers’ perceptions of their experience with the SHUTi intervention, summarizing what program characteristics caregivers found helpful and should be retained, as well as those that were impractical and should be modified. The generalizability of these findings is further explored—the extent to which caregivers’ feedback is (i) specific to caregivers or generalizable to a wider range of potential users, and (ii) specific to SHUTi or generalizable to other digital health interventions or cognitive-behavioral therapy for insomnia (CBT-I) in general. The goal of this investigation is to begin to establish how to leverage the wealth of existing evidence-based interventions to make the greatest impact on caregivers’ pressing psychosocial needs.

Methods

This article presents the primary qualitative findings from the SHUTi-CARE (Caregiver Acceptability REsearch) study, a single-group trial examining medium- and high-intensity family caregiver engagement with SHUTi, an Internet-delivered insomnia intervention. A priori primary quantitative aim results are presented elsewhere [39]. The trial (ClinicalTrials.gov NCT04986904) and protocol [40] were pre-registered. Where applicable, reporting in this manuscript follows the CONSORT-EHEALTH [41] and COREQ checklists [42]; see Supplementary Table S1. Study materials and supplemental tables are publicly available at https://bit.ly/SHUTiCARE-project. Deidentified data are available by data use agreement with the University of Virginia (UVA), per institutional requirements.

Participants

Key details of the trial are summarized here; complete information is available in the study protocol [40]. Procedures were approved by Institutional Review Boards (IRB) at UVA (HSR210255) and the University of Pittsburgh (Office of Research Protections IRB STUDY21080076). The sample size of 100 family caregivers was determined to provide adequate statistical power to evaluate the primary quantitative aim (see [39]). Participants were recruited from across the USA via the study website from January to December 2022.

For the purpose of this study, “family caregiving” was defined in accordance with the definition used by the National Alliance for Caregiving (NAC [1]). Specifically, caregivers self-reported providing unpaid care to an adult family member (or family-like individual) and/or providing care to a child beyond what would be expected due to the child’s medical, behavioral, or developmental condition or disability. Caregivers were eligible to participate if they met the following criteria: provided medium- or high-intensity care (based on NAC Level of Care Index [1]; henceforth collectively described as “higher-intensity”); expected to provide this level of care for 3 or more months; age ≥18; reported clinically significant insomnia symptoms on the seven-item Insomnia Severity Index (ISI ≥10 [43, 44]); had regular Internet access; were willing to receive study emails; spoke and read English; and resided within the USA. Consistent with prior SHUTi trials [45], exclusion criteria were: irregular sleep schedule (e.g., paid shift work, typical bedtime before 8 pm or later than 2 am); current behavioral or psychological insomnia treatment; severe difficulties using a computer; and certain medical or psychiatric contraindications (e.g., restless leg syndrome, history of mania).

Procedures

Individuals expressed interest in the study by completing an online prescreening form. All assessment was captured through Qualtrics Highly Sensitive Data (Qualtrics, Provo, UT), a web-based and HIPAA-compliant survey tool. To ensure the authenticity of participants, the identities of pre-eligible individuals were verified using TLOxp (TransUnion), a web-based people search service [46]. Verified individuals completed a phone call with the research coordinator to finalize eligibility screening. Confirmed eligible participants who wished to enroll completed informed consent and provided a digital signature via DocuSign. Enrolled participants completed the online baseline assessment (surveys and sleep diaries). All participants who completed the baseline assessment received access to SHUTi for 9 weeks, after which time participants completed the post-assessment. Participants were compensated U.S. $40 per completed assessment.

Intervention

The SHUTi intervention has demonstrated robust effects on sleep metrics in over 25 clinical trials (e.g., [29, 31, 33, 45]). A full description of the intervention has been published, along with details on the development process that includes rigorous and iterative user-centered design and collaboration with software engineers, a user interface/user experience consultant, and an instructional designer [30]. A summary of intervention content and other details can be found in the study protocol [40]. Key intervention details are summarized here.

SHUTi is a fully automated Internet-delivered insomnia program with six sequential “Cores.” There is no caregiving-specific content. Cores deliver the primary treatment mechanisms of CBT-I: basic sleep education, sleep restriction, stimulus control, cognitive restructuring, sleep hygiene, and relapse prevention. Between the first and second Core, users must complete five online daily sleep diaries within 7 days to receive their “sleep window” for the sleep restriction technique. Users may continue to enter sleep diaries to receive updated sleep windows based on their sleep improvement. When a Core is completed, the next Core is made available 1 week later, similar to provider-delivered CBT-I. The program utilizes interactive elements like quizzes, videos, and games to deliver content in a highly engaging way. Automated email reminders also support engagement. Technical assistance is available, but no clinical support is provided. Participants access the program via a password-protected website from any Internet-connected computer, tablet, or smartphone.

Measures

SHUTi Users: program satisfaction and recommendations for caregiver-specific tailoring

At post-assessment, participants who completed one or more SHUTi Cores responded to nine open-ended prompts: four regarding their satisfaction with the program and five regarding their experience with SHUTi as a caregiver. See Supplemental Table S2 for prompts. A summary of qualitative findings and a brief questionnaire were prepared and distributed to SHUTi Users who endorsed willingness to be recontacted by the study team on their post-assessment survey for synthesized member checking [47], a structured form of participant validation of results.

SHUTi Nonusers: barriers to program uptake

At post-assessment, participants who did not complete any SHUTi Cores completed a checklist of potential reasons why they did not use the program with follow-up prompts. See Supplemental Table S3 for barriers and prompts.

Data Analysis

We used a rapid content analysis approach with matrix analysis methods to efficiently derive themes from the >600 open-ended responses shared by SHUTi Users [48, 49]. Each response was deductively coded by two independent reviewers (K.S., K.P., M.M., J.G., and H.D.) according to an a priori codebook. Discrepancies were resolved by a third coder or the consensus of the team. Quotes were then organized into two matrices according to the a priori codes: One matrix for quotes expressing satisfaction (or interest in keeping program elements the same) and one for dissatisfaction (or interest in modifying program elements), with each matrix divided according to level of generalizability by population (specific to caregivers vs. applicable to the general population) and by intervention characteristics (specific to Internet-delivered CBT-I [e.g., SHUTi] vs. CBT-I by any delivery method vs. digital health delivery of any intervention). Matrices facilitated synthesizing themes across similar quotes.

Nonuser feedback was too limited to facilitate coding and is presented in full in Supplementary Table S3.

Results

Participant Characteristics

Of 516 unique applicants submitting online prescreening interest forms, 253 were ineligible for the study (49.0%), 65 were pre-eligible but either could not be contacted to schedule or declined a full eligibility screening (12.6%), and 81 were determined to be fraudulent (15.7%). Of the 118 caregivers completing a phone screening, 16 were deemed ineligible (13.6%) and 102 were eligible and enrolled. One caregiver was lost to follow-up and one withdrew. See Fig. 1 for the CONSORT chart.

Fig. 1.

CONSORT chart.

The 100 caregivers of the analyzed sample were generally middle aged (M = 52.82). The majority were female (87%), White (79%), and non-Hispanic (94%), with at least a college degree (66%). One-third of the sample reported a household income of $75,000 or greater. Caregivers had typically been providing long-term care (months providing care M = 68.46), and 38% provided care to more than one care recipient. Primary care recipients of enrolled caregivers most commonly needed care due to Alzheimer’s disease or dementia (18%), cancer (11%), or a mobility problem (10%). See Table 1 for full sample details.

Table 1.

Sample Descriptives

(N, %, unless specified)	Total sample (N = 100)	Nonusers (N = 18)	Users (N = 82)
Caregiver (participant)
Age (M [SD])	52.82 (13.10)	51.11 (12.01)	53.20 (13.37)
Gender
Female	87 (87.0)	17 (94.4)	70 (85.4)
Male	11 (11.0)	1 (5.6)	10 (12.2)
Othera	2 (2.0)	0 (0.0)	2 (2.4)
Race
Asian	4 (4.0)	0 (0.0)	4 (4.9)
Black/African American	14 (14.0)	2 (11.1)	12 (14.6)
White	79 (79.0)	14 (77.8)	65 (79.3)
Other or Multiple	3 (3.0)	2 (11.1)	1 (1.2)
Ethnicity
Hispanic/Latino/a	6 (6.0)	2 (11.1)	4 (4.9)
Not Hispanic/Latino/a	94 (94.0)	16 (88.9)	78 (95.1)
Household income
<$30,000	18 (18.0)	1 (5.6)	17 (20.7)
$30,000–50,000	23 (23.0)	1 (5.6)	22 (26.8)
$50,000–75,000	19 (19.0)	2 (27.8)	14 (17.1)
$75,000–100,000	12 (12.0)	3 (16.7)	9 (11.0)
>$100,000	21 (21.0)	7 (38.9)	14 (17.1)
Prefer not to answer	7 (7.0)	1 (5.6)	6 (7.3)
Education
High school degree/GED	9 (9.0)	0 (0.0)	9 (11.0)
Associate’s degree/some college	25 (25.0)	5 (27.8)	20 (24.4)
Bachelor’s degree	26 (26.0)	5 (27.8)	21 (25.6)
Some graduate education	6 (6.0)	0 (0.0)	6 (7.3)
Graduate degree	34 (34.0)	8 (44.4)	26 (31.7)
Health literacy
Quite a bit, A little	11 (11.0)	3 (16.7)	8 (9.8)
Extremely	89 (89.0)	15 (83.3)	74 (90.2)
Months providing care to CRb (M [SD])	68.46 (43.69)	88.62 (46.44)	64.27 (42.21)
11 or more years	20 (20.0)	7 (38.9)	13 (15.9)
Total number of care recipients
1	62 (62.0)	11 (61.1)	51 (62.2)
2	32 (32.0)	4 (22.2)	28 (34.2)
3	4 (4.0)	2 (11.1)	2 (2.4)
4 or more	2 (2.0)	1 (5.6)	1 (1.2)
Care provided to
Adult(s) only	84 (84.0)	14 (77.8)	70 (85.4)
Child(ren) only	4 (4.0)	0 (0.0)	4 (4.9)
Adult(s) and child(ren)	12 (12.0)	4 (22.2)	8 (9.8)
Caregiving intensity score (M [SD])	8.69 (2.80)	9.44 (3.75)	8.52 (2.54)
Medium intensity (score = 5)	14 (14.0)	3 (16.7)	11 (13.4)
High intensity (score > 5)	86 (86.0)	15 (83.3)	71 (86.6)
Primary care recipient (CR)
CR agec (M [SD])	64.65 (24.70)	61.33 (26.98)	65.38 (24.28)
90 and older	8 (8.0)	2 (11.1)	6 (7.3)
CR is the caregiver’s…
Spouse, partner	24 (24.0)	3 (16.7)	21 (25.6)
Parent, parent-in-law	46 (46.0)	10 (55.6)	36 (43.9)
Son, daughter	16 (16.0)	5 (27.8)	11 (13.4)
Other	14 (14.0)	0 (0.0)	14 (17.1)
Primary condition requiring care
Alzheimer’s disease/dementia	18 (18.0)	1 (5.6)	17 (20.7)
Autism spectrum disorder	6 (6.0)	2 (11.1)	4 (4.9)
Back problems	5 (5.0)	2 (11.1)	3 (3.7)
Brain damage/injury	5 (5.0)	2 (11.1)	3 (3.7)
Cancer	11 (11.0)	2 (11.1)	9 (11.0)
Developmental/intellectual disorder	3 (3.0)	0 (0.0)	3 (3.7)
Feeble, unsteady, falling	3 (3.0)	1 (5.6)	2 (2.4)
Lung disease, emphysema, Chronic obstructive pulmonary disease (COPD)	3 (3.0)	0 (0.0)	3 (3.7)
Mental/emotional illness	3 (3.0)	1 (5.6)	2 (2.4)
Mobility problem	10 (10.0)	1 (5.6)	9 (11.0)
Old age	8 (8.0)	1 (5.6)	7 (8.5)
Parkinson’s	6 (6.0)	2 (11.1)	4 (4.9)
Stroke	5 (5.0)	1 (5.6)	4 (4.9)
Otherd	14 (14.0)	2 (11.1)	12 (14.6)

^aInclusive of transgender, genderqueer, other, prefer not to say.

^bCare duration was assessed up to 11 years (last option = “11 or more years”)—recoded to 11 years for computing M, SD.

^cCR age was assessed up to 90 (last option = “90 or older”)—recoded to 90 years for computing M, SD.

^dIncludes conditions where total N < 3 (arthritis; diabetes; hypertension; heart disease, heart attack) and “Other.”

SHUTi User Feedback

See Supplementary Table S4 for a listing of themes (in bold below), codes (in italics below), and representative quotations. Caregivers are identified in-text and in Supplementary Tables with an identifying number specific to this manuscript (U indicates SHUTi User, N indicates Nonuser), the primary care recipients’ condition requiring care, reported proximity to the care recipient (sleep in the same room [bedpartners], live together in same home but sleep separately [in-home], or other living situation [live apart]), and for SHUTi Users, their final Core completed. A summary of findings was emailed to SHUTi Users with a brief survey for synthesized member checking. Of 66 caregivers emailed by the study team, 23 returned responses (35%; see Supplementary Table S5 for characteristics). All but one respondent (n = 22, 96%) endorsed that the summary matched their experience, 20 recommended no changes to the summary (87%), and 20 recommended no additions (87%). Contrasting or clarifying comments from member checking responses are integrated within the summary below.

Program feasibility

Several codes emerged related to the feasibility of caregivers successfully carrying out the SHUTi program. Caregivers generally expressed satisfaction with the way the SHUTi program was delivered. Specifically, caregivers liked how information was presented throughout the program in an interesting, comprehensive, and easy-to-use way (Delivery Format). They identified Program Flexibility and Delivery Convenience as key features promoting feasibility—namely, that completing the program and assessments on their own time and pace via the Internet was “very valuable” (U10, stroke, in-home, Core 4). One caregiver shared, “I see the built-in flexibility making the program workable for anyone, caregiver or otherwise, who is motivated to participate” (U6, mobility problems, live apart, Core 2). Furthermore, caregivers largely found the Program Manageable, by not interfering with their caregiving or other daily responsibilities, and Easy-to-Use, with a step-by-step approach of practical strategies. Reminders were spontaneously raised by several caregivers as an “essential” part of the intervention that helped them stay “on track,” which was reiterated by a respondent during member checking (U7, cancer, bedpartner, Core 6). In fact, some caregivers requested more frequent alerts. One caregiver wrote, “Give more alerts for following Cores and the goals of each Core…” because “…being a caregiver it’s easy to forget and get tied up with outside problems” (U20, autism spectrum disorder, in-home, Core 6).

Caregivers also shared barriers and suggestions to improve program feasibility. Specifically, several caregivers expressed interest in additional ways to access program information (Delivery Format), with particular interest in an audio format to permit multitasking. This recommendation may help address the primary barrier to program feasibility, namely, being Too Busy to complete Cores, submit daily diaries, or remember to follow program recommendations. Feelings of daily overload and interest in different program delivery modalities are not unique to caregivers; additionally, as described above, many caregivers noted the time requirements were manageable. However, quotes highlight how caregivers may particularly benefit from ensuring that interventions are highly efficient. One caregiver described that for her, the program was “Just a little too much. Caregivers are doing double duty. I run two homes. Work a full-time job… then organize my parent’s life all while trying to do things to decrease stress in my life. Maybe the program needs [to be] shortened or trimmed for us?” (U24, cancer, live apart, Core 3).

Program appropriateness

Several codes also emerged within a theme about the appropriateness—or the overall relevance and compatibility—of SHUTi for caregivers. Several issues were raised by caregivers as not adequately addressed by the SHUTi program, which were largely specific to caregiving, but also generalizable to CBT-I more broadly. Caregivers raised the Barrier of Nighttime Caregiving as incongruent with program recommendations, and they felt program content did not adequately address this issue. As one caregiver shared, “Often, I would be awake in the middle of the night not because of my insomnia, but because my daughter with disabilities was awake and needed my attention…” (U25, developmental/intellectual disorder, bedpartner, Core 6). Relatedly, caregivers often specifically requested Help for Sleep Interruptions, like “how to deal with unpredictable disrupted sleep” (U27, other disorder, bedpartner, Core 3), as well as how to record awakenings on their sleep diaries that were not the result of insomnia. Additional barriers were noted related to their Sleep Environment and having a Schedule Dependent on Others. Caregivers noted certain safety precautions for their care recipients made for less-than-ideal sleep environments, like being unable to use earplugs in case of calls for help or needing to keep lights on to mitigate fall risk. Co-sleeping presented issues related to both the sleep environment and sleep schedule; one caregiver supporting her husband with dementia noted that she could not delay her bedtime as her husband would become agitated if she did not go to bed with him (U3, Alzheimer’s disease/dementia, bedpartner, Core 6). Other caregivers noted how their sleep was disrupted by their bedpartner care recipients’ health issues.

When asked directly about tailoring SHUTi to better meet caregivers’ needs (see Supplementary Table S2 open-ended prompts 8 and 9), there was no clear mandate. Caregivers' comments were relatively evenly divided between indicating tailoring was unwarranted (n = 33; e.g., “I was very happy with the way it is now” [U47, lung disease/emphysema/Chronic obstructive pulmonary disease (COPD), in-home, Core 6]), sharing suggestions to refine SHUTi that would exclusively help caregivers (n = 28; e.g., “Tailoring the program to specific needs of caregivers would be much more useful” [U28, old age, bedpartner, Core 6]), and sharing suggestions that could benefit anyone using SHUTi (n = 32; e.g., “Maybe break out the Core sections into smaller sections” [U31, Parkinson’s disease, in-home, Core 6]).

This diversity of opinions was also reflected in the overarching split between caregivers who indicated that the Program Was A Good Fit for them so Personalization would not be necessary, compared with those who indicated that the Program Was Not A Good Fit for them and that Personalization would have been beneficial. Many caregivers provided brief comments indicating the program worked well for their needs (e.g., “It fit perfect” [U48, cancer, live apart, Core 2]), and that the existing SHUTi program provided a sufficient diversity in strategies for each individual to “…use the advice that fits one’s situation and reject what does not” (U29, brain damage/injury, in-home, Core 6). In addition, one caregiver noted that “each caregiver’s experience is different – might be better to keep it general” (U17, other disorder, bedpartner, Core 6). Other caregivers endorsed that the program was not the right fit for them, without further reasons (e.g., “It didn’t fit for me” [U33, cancer, bedpartner, Core 2]). Some caregivers addressed how providing caregiving-specific content could “enhance [their] experience” (U42, Alzheimer’s disease/dementia, live apart, Core 6), as well as requested adding more personalized recommendations based on deeper assessment of each caregiver’s individual needs.

Anxiety, stress, and overload

Caregivers spontaneously raised the concepts of Cognitive Load and Anxiety and Stress throughout their responses. They described ways that these factors made it difficult to sleep and engage with SHUTi, as well as ways SHUTi improved these issues. Anxiety, stress, and feeling overloaded were mentioned as key factors related to caregivers’ sleep difficulties—one caregiver wished for “more specific ways to turn your brain off of your caregiving responsibilities” (U21, Alzheimer’s disease/dementia, live apart, Core 6). Feeling overwhelmed was also cited as a barrier to completing the program, as one caregiver shared that it became “one more thing ‘to-do’ on [my] already almost overloaded to-do list” (U24, cancer, live apart, Core 3). At the same time, caregivers also noted a key benefit of the program was gaining skills for how to “shut your mind down and not take your worries to bed” (U38, Alzheimer’s disease/dementia, live apart, Core 6). Cognitive strategies presented to address maladaptive beliefs about sleep were also reported to be “very helpful in life situations, not just for sleep problems” (U2, Alzheimer’s disease/dementia, live apart, Core 6).

Behavioral components

Across responses, caregivers frequently discussed the many behavioral therapeutic components of CBT-I as implemented through SHUTi. Many caregivers expressed appreciation of the program’s provision of an array of Healthy Sleep Tools “to get better sleep” (U38, Alzheimer’s disease/dementia, live apart, Core 6) and “manage future problems” (U21, Alzheimer’s disease/dementia, live apart, Core 6). As one respondent to the member checking shared, “I mainly benefited from the suggested guidelines of how to prepare for sleep, sleep hygiene, maintain my healthy practices throughout the day, and to stay out of bed except when sleeping” (U37, other disorder, in-home, Core 6).

Of behavioral techniques, caregivers were by far the most divided about the Sleep Window, or the sleep restriction therapeutic technique. Most caregivers expressed that this technique was their favorite part of the program and identified it as a highly effective strategy. As one caregiver shared, she did “NOT enjoy staying up that late (2 AM) BUT, the sleep that I get now [is] much better quality than the joke for sleep I had been living with for years… THANK YOU for teaching me a new way, and enhancing my life tenfold” (U40, cancer, live apart, Core 6). A respondent to the member checking commented that “My applying [the] SHUTi process did not typically affect my care. I found the sleep schedule difficult initially, but it did seem to help when I would stick to it…” (U49, Parkinson’s disease, live apart, Core 6). Relatedly, many caregivers endorsed a key benefit of the program as developing a better Sleep Routine. As one caregiver shared, following the program helped her have “a realistic schedule and the permission to not have to go to sleep at an unreasonable time…” (U4, mobility problems, in-home, Core 2). During member checking, another respondent stated, “I found it easy to follow and develop a bedtime routine. I’ve also been able to improve and maintain that routine” (U46, Parkinson’s disease, live apart, Core 6).

Although fewer in number relative to those who favorably reviewed the sleep window, there was a sizeable proportion of caregivers who described concerns with this component of the program. Caregivers endorsed concerns with the technique that are frequently reported in CBT-I regardless of targeted population—for example, wanting more flexibility and leniency so “one bad night’s sleep” would not affect their window (U4, mobility problems, in-home, Core 2), or feeling “very, very fatigued and sleepy trying to get through the days” (U50, Parkinson’s disease, in-home, Core 6). However, caregivers also noted some caregiving-specific issues following the technique (e.g., “…it got hard to stay on task with the recommended sleep window due to problems resurfacing with my child…” [U41, mental/emotional illness, in-home, Core 4]). These challenges frequently overlapped with the sleep environment and schedule dependency issues addressed above. Additionally, some caregivers expressed specific concerns about daytime sleepiness, as one caregiver shared, “…limiting the window of sleep when you’re so exhausted was scary when I was worried about being able to care for my mom … or was [too] tired to drive her” (U24, cancer, live apart, Core 3). Through the member checking process, one caregiver added, “I would like to add that the sleep window (with such restricted sleep) made me groggy and unable to think straight. I felt that I could not care for my husband very well and had to stop it after two nights…. I could not take the risk of not being able to help my husband” (U3, Alzheimer’s disease/dementia, bedpartner, Core 6). These concerns echoed the concept endorsed by a few caregivers of Needing All Sleep Possible, where they were unwilling or unable to give up on any sleep in the short term to realize longer-term benefits.

Key to the sleep restriction technique, many more caregivers endorsed satisfaction than dissatisfaction with the Sleep Diaries. Many caregivers identified the sleep diary as their favorite aspect of the intervention that helped them learn about their sleep and track their adherence to program recommendations (e.g., “The diary made me much more aware of my sleep habits and reduced the stress of worrying about sleep” [U44, mobility problems, live apart, Core 5]). Additionally, caregivers largely indicated they found the diaries simple to keep, as one caregiver shared, “… The diaries are easy to enter into the system and [it] doesn’t take much time when it’s completed each day” (U2, Alzheimer’s disease/dementia, live apart, Core 6). However, some caregivers expressed concerns with the diaries that are common across CBT-I trials, such as “some days I was just too busy and tired to record or enter all the data” (U9, developmental/intellectual disorder, in-home, Core 4), and that wearables helped them accurately complete their diaries. This was reiterated by a respondent during member checking that “The sleep diary would’ve been next to impossible without a Fitbit” (U31, Parkinson’s disease, in-home, Core 6).

The other central behavioral strategy in CBT-I, Stimulus Control, was also identified by caregivers as a helpful strategy that made a positive impact on their sleep. One caregiver shared, “The most helpful things were learning about the impact of listening to radio or TV while in bed… stopping this behavior helped me to fall asleep easier” (U28, old age, bedpartner, Core 6). The same member checking respondent who expressed she had stopped practicing sleep restriction due to daytime sleepiness affecting her caregiving also shared that, “I love the idea that bed is for sleeping, not reading or doing games on my phone. When I lie awake, I get up and read on the couch. When I get sleepy, I go back to bed and quickly get back to sleep” (U3, Alzheimer’s disease/dementia, bedpartner, Core 6). Additionally, the few mentions of Sleep Hygiene identified these strategies as appreciated and useful.

SHUTi Nonuser Feedback

Of the 14 SHUTi nonusers who completed post-assessment, caregivers typically only endorsed one barrier preventing their uptake of SHUTi (range: 0–3; see Supplementary Table S3). The most frequently cited barriers by caregivers were concerns that SHUTi would take too much time (n = 5, 36%) and that it would not work with their lifestyle (n = 3, 21%)—as one caregiver responded, she was “just extremely busy and overworked right now, so finding time to sit down and really read and comprehend what was being provided became difficult” (N3, cancer, in-home). Additionally, two caregivers indicated that caregiving responsibilities specifically impeded their ability to take up SHUTi: one caregiver shared, “At the time I was to start the program, my [mother-in-law] suffered a stroke… I am even more of a caretaker now. I have no time, at all” (N11, other disorder, in-home).

Discussion

Findings from this first-of-its-kind single-group trial of an Internet insomnia program among higher-intensity family caregivers suggest that fully automated Internet interventions are a good fit for delivering psychosocial care to this population. When asked directly regarding the utility of tailoring the program for caregivers, caregivers in this sample were divided regarding whether such tailoring would be necessary or even beneficial. Where ideas for improving SHUTi were raised, these were often considerations that would benefit all program users rather than caregivers alone. Caregiving-specific challenges to completing SHUTi were largely related to the key therapeutic techniques of CBT-I, such as difficulty with the behavioral strategies due to sleep interruptions from nighttime caregiving, sleep environment limitations, and schedule dependencies on the care recipients. Findings suggest that extensive tailoring may not be required for most caregivers to benefit from an existing online CBT-I program; however, additional guidance on integrating behavioral strategies in the context of challenging sleep schedules and environments may help a subset of caregivers get the most from CBT-I by any method (e.g., fully automated online or in-clinic with a provider).

Caregivers who completed one or more of the SHUTi Cores largely indicated that receiving the intervention fully automated online was a good fit for caregivers’ lifestyles. One caregiver noted that this was SHUTi’s “best feature,” because if “you start doing one Core and your [care recipient] needs you, … that’s okay because you can close it and finish it another time…” (U5, other disorder, bedpartner, Core 6). Caregivers commonly cited this autonomy to use the program at their own convenience as important to their ability to engage with the program. This praise for flexible and on-demand programs corroborates findings across multiple trials of digital health interventions for specific caregiving populations [6]. As such, asynchronous and self-guided program features should be prioritized for digital health interventions for caregivers for the greatest flexibility. Conversely, the inclusion of any required synchronous intervention components (e.g., telephone calls) should be carefully considered—these will not only limit program flexibility, but also implementation likelihood due to added intervention complexity [50].

Caregivers also highlighted the automated email reminders as beneficial. SHUTi sends emails daily as reminders to complete a sleep diary entry and twice weekly to encourage Core engagement and completion. While there are sometimes concerns about “bothering” participants with reminders, caregivers who suggested changes to reminders actually requested more reminders. Given this positive view of reminders, as well as evidence that email reminders are an effective tool for engaging users with online programs [51, 52], frequent reminders should be routinely integrated into digital health programs for caregivers.

While the fully autonomous online delivery and regular reminders helped caregivers use SHUTi, competition with caregivers’ many other responsibilities remained a leading barrier for both program engagement and uptake. It is worth noting, however, that the quantitative findings from this trial showed no associations between greater caregiving stress or responsibilities with lower engagement—in fact, worse care recipient functioning was associated with greater engagement [39]. Being too busy for digital health interventions is not a barrier exclusive to caregivers [22, 53], but ensuring interventions are as efficient as possible may particularly benefit caregivers who have many competing priorities. For this reason, the Multiphase Optimization Strategy (MOST) methodological framework, used to develop highly efficient interventions [54], may be a particularly useful approach within caregiver digital health intervention research. Given that almost two-thirds of the SHUTi nonusers indicated that their primary barrier to uptake was related to time or balancing responsibilities, limiting the intervention burden will be critical to ensuring digital health interventions are most accessible for all caregivers.

In addition to limiting any unnecessary components, designing interventions that can be consumed through multiple modalities (e.g., online, by smartphone, audio) may improve uptake across caregivers. This request has been found previously in user-centered design studies of Internet interventions for dementia caregivers [55, 56]. It should be noted that, while a few caregivers expressed they would have liked to access SHUTi by smartphone or tablet, this is an existing capability of the SHUTi program. Reiterating throughout a program how it can be accessed on different devices may be necessary.

While caregivers largely found the delivery method of SHUTi feasible, they also commonly raised challenges with the fit of the CBT-I therapeutic techniques as a result of their caregiving contexts. These included nighttime awakenings due to caregiving, unideal sleep environments due to care recipient safety or comfort needs, and schedules dependent on the needs and preferences of the care recipient. Notably, such challenges were not restricted by disease context; for example, there were caregivers for individuals with cancer or Parkinson’s disease who expressed challenges adhering to sleep restriction, and other caregivers in the same disease contexts who found sleep restriction feasible and beneficial. This variability questions the utility of disease context as a key tailoring factor in CBT-I, as opposed to cross-cutting factors like the presence of nighttime caregiving responsibilities. Several clinical modifications to traditional CBT-I may benefit caregivers—in any disease context—who have such challenges with behavioral strategies. Integrating problem-solving training may help caregivers respond effectively to their unique and evolving challenges [57, 58]. Caregivers may also be directed to focus on implementing behavioral strategies to the best of their ability, rather than on complete adherence, as prior research suggests that improving sleep behaviors—even without strict adherence—can result in treatment gains [59, 60]. Lastly, caregiver stress and anxiety may be addressed by existing therapeutic components of traditional CBT-I—for example, caregivers may be particularly encouraged to take up regular relaxation practices and to utilize cognitive strategies for managing distressing thoughts about caregiving in addition to sleep.

Such program modifications may help caregivers with specific nighttime challenges to better enact and adhere to CBT-I recommendations; however, such tailoring is likely not necessary for all caregivers. Many caregivers expressed that the existing SHUTi program that has no caregiving-specific content was a good fit for their lifestyles, even while they provided high-intensity care. Furthermore, quantitative findings from this trial found this standard program was highly effective among caregivers, with no consistent differences in outcomes related to factors like caregiving stress or burden [39]. This suggests that the wealth of existing evidence-based digital health programs could be leveraged to address caregivers’ urgent psychosocial needs. Doing so could help speed the implementation of digital healthcare for caregivers, given that research developing caregiver-targeted interventions has largely not resulted in implementation-ready programs [50, 61]. Where caregivers endorsed suggestions for program improvements, these were commonly suggestions that could benefit all users, like increasing personalization or program efficiency. These findings reiterate that existing robust digital health programs that have been refined through years of user-centered design research may sufficiently meet many caregivers’ needs.

Limitations

Findings should be considered in light of several limitations. Relative to estimates of the overall population of family caregivers in the USA [1], caregivers in the SHUTi-CARE trial were more likely to have completed higher education and be non-Hispanic White. There was, however, significant economic diversity within this sample, with participants relatively less affluent than the overall population of caregivers. Given recruitment of caregivers into trials is a known challenge [62, 63], future studies may consider contracting with an experienced clinical recruitment agency to identify a more diverse pool of caregiver participants. While it is a strength of this study that results were returned to participants for their review, it should be noted that only about one-third of the participants contacted returned surveys within the two-week review period. Moreover, those who returned member checking surveys were disproportionately non-Hispanic White and had completed the program (i.e., completed through Core 4). Member check respondents were, however, more commonly providing care to more than one care recipient. Lastly, this trial also exclusively recruited family caregivers. While caregivers’ responses were coded according to their applicability to non-caregivers, direct comparisons between caregivers’ and non-caregivers’ perceptions of the SHUTi program cannot be drawn.

Conclusions

In this unique single-group trial of a robust Internet insomnia program among higher-intensity family caregivers, participants who used the program endorsed that fully automated Internet interventions can be highly feasible for caregivers given their flexibility and convenience. While the delivery format was largely described as a good fit for the caregivers’ lifestyles, many caregivers found that the behavioral strategies of CBT-I were challenging to implement given sleep interruptions from nighttime caregiving, sleep environment limitations, and schedule dependencies on the care recipients. Overall, caregivers expressed split opinions regarding the potential benefit or appeal of caregiver-specific tailoring, and caregivers’ suggestions for improving the program were largely not specific to certain disease contexts—or caregivers at all. While there was not a clear signal that SHUTi should be tailored specifically for caregivers, it is worth testing whether program additions like strategies for problem-solving and managing caregiving stress may encourage more caregivers to take up and complete the program. Taken together, the ultimate impact of a digital health intervention across all caregivers should be carefully considered to determine the value of potential tailoring, weighing whether modifications are expected to improve program efficacy to sufficiently outweigh limitations to treatment reach.

Funding

This trial is funded by the National Institutes of Health (NIH) National Center for Advancing Translational Sciences (NCATS R21TR003522; principal investigator: KS). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Compliance with Ethical Standards

Authors’ Statement of Conflict of Interest and Adherence to Ethical Standards K.S., K.P., M.M., J.G., J.K., and H.D. have no conflicts to report. Over the past 3 years, D.B. has served as a paid or unpaid consultant to Sleep Number, Idorsia, and Eisai. All consulting agreements have been for a total of less than $5,000 per year from any single entity. Consulting has focused on insomnia, behavioral sleep treatments, and measurement of sleep characteristics. D.B. is an author of the Pittsburgh Sleep Quality Index, Pittsburgh Sleep Quality Index Addendum for PTSD (PSQI-A), Brief Pittsburgh Sleep Quality Index (B-PSQI), Daytime Insomnia Symptoms Scale, Pittsburgh Sleep Diary, Insomnia Symptom Questionnaire, and RU_SATED (copyrights held by University of Pittsburgh). These instruments have been licensed to commercial entities for fees. He is also co-author of the Consensus Sleep Diary (copyright held by Ryerson University), which is licensed to commercial entities for a fee. L.R. has equity ownership in BeHealth Solutions, LLC, who originally licensed the Sleep Healthy Using the Internet (SHUTi) program from the University of Virginia. Somryst, a commercial prescription digital therapeutic for insomnia, was developed based on the SHUTi program by Pear Therapeutics, which subsequently sold their license to Nox Health. Nox Health has a royalty agreement with BeHealth Solutions, LLC and the UVA Licensing and Venture Group.

Authors’ Contributions Kelly M. Shaffer (Conceptualization [lead], Data curation [equal], Formal analysis [lead], Funding acquisition [lead], Investigation [equal], Methodology [lead], Project administration [supporting], Writing – original draft [lead], Writing – review & editing [lead]), Kate Perepezko (Data curation [equal], Formal analysis [equal], Writing – review & editing [equal]), Jillian V. Glazer (Data curation [equal], Formal analysis [equal], Project administration [lead], Writing – review & editing [equal]), Meghan K. Mattos (Formal analysis [equal], Funding acquisition [equal], Methodology [equal], Writing – review & editing [equal]), Julie Klinger (Formal analysis [supporting], Project administration [supporting], Writing – review & editing [equal]), Daniel J. Buysse (Funding acquisition [equal], Methodology [equal], Writing – review & editing [equal]), Lee M. Ritterband (Funding acquisition [equal], Methodology [equal], Supervision [equal], Writing – review & editing [equal]), and Heidi Donovan (Data curation [equal], Formal analysis [equal], Funding acquisition [equal], Methodology [equal], Resources [equal], Supervision [equal], Writing – review & editing [equal])

Study Registration This trial was registered at ClinicalTrials.gov (Identifier NCT04986904).

Analytic Plan Pre-registration The protocol and analysis plan for this study were pre-registered (RR1) at doi:10.2196/34792.

Analytic Code Availability The codebook and coding form are available at https://bit.ly/SHUTiCARE-project.

Materials Availability Materials used to conduct the study are available at https://bit.ly/SHUTiCARE-project.

Data Availability

Deidentified quotations are available via data use agreement with the University of Virginia.