This study is a feasibility study with a single-group, pretest-posttest design (n=21), in which all participants received a guided iCBT intervention for 8 weeks. Assessments were conducted at baseline (T0) and after the intervention (T1; 8 weeks after baseline). Full details of the study protocol have been published elsewhere [30]. The study protocol included 2 intervention conditions: guided iCBT and blended iCBT (a combination of online sessions and face-to-face support). However, this paper focuses exclusively on the guided iCBT condition.

Participants were recruited from the general population through the Link2Trials BV online recruitment platform, which served as the primary setting for study enrollment, between September 2022 and May 2023. Link2Trials BV is a clinical research organization specialized in patient recruitment and adherence services for clinical trials. Participants were included if they (1) were aged 55 years or older, (2) had mild to moderate depression based on a Geriatric Depression Scale-15 (GDS-15) score between 5 and 10, (3) were able to work with a computer, and (4) had access to a computer with internet. The lower age threshold of 55 years was selected to align with demographic classifications used in the Netherlands, where individuals aged 55 and older are considered part of the “younger older adult” population [31]. Additionally, several Dutch mental health services provide care trajectories for adults aged 55 years and older. Care standards note that depression in older adults may present differently due to frailty, highlighting the need for tailored approaches [32]. While this age cutoff is not based on formal clinical guidelines, it allows for a broader and more inclusive representation of older adults experiencing depressive symptoms [9]. Candidates were excluded from the study if they (1) did not have adequate proficiency in the Dutch language or (2) had suicidal thoughts (a score of 1 or higher on item 9 of the Patient Health Questionnaire 9 [PHQ-9]). Excluded participants received an email detailing the reason for their exclusion. On the advice of the Medical Ethical Committee (METC), participants who were excluded because they had moderate to severe depression were advised to contact their general practitioner. Participants identified as being at risk for suicide were advised to contact the national suicide prevention helpline [33].

The sample size was determined based on recommendations for feasibility studies, which prioritize evaluating study procedures over efficacy testing [34]. Accordingly, the primary outcomes of this study concerned feasibility aspects such as usability, engagement, and acceptability, and not the estimation of effect sizes or statistical power. While sample size guidelines for feasibility studies vary, a range of 24 to 50 participants is typically recommended to ensure adequate representation [34,35]. Given the variability in participants’ age and the diversity of depressive symptom profiles in older adults in the general population, we aimed to recruit 30 participants.

Participants were recruited through the Link2Trials BV online recruitment platform. After expressing interest via the platform, potential participants were contacted by the research team and sent an information letter along with a printed informed consent form by post. Those who were willing to take part in the study were asked to sign and return the consent form by post. The procedure of consent via the post was required by the METC, which mandated a physical signature for ethical approval. Upon receiving the signed informed consent by post, participants were sent a link to the online screening questionnaire via a secure electronic data capture system (Castor). Eligible participants received a link to the online baseline (T0) questionnaire. Once the baseline questionnaire was filled out, participants were assigned a trained coach who provided them with feedback during the intervention. These coaches were master’s-level psychology students who had received training in delivering guided iCBT interventions for older adults and were supervised by a licensed psychologist to ensure adherence to the intervention guidance protocol. The login details for Moodbuster (the online platform) and a manual on the use of the platform were shared with the participants through secure email. The posttreatment assessment (T1) was sent to the participants 8 weeks after the baseline assessment (T0). Participants received a €30 (US $34.84) voucher after completing the postassessment questionnaire, regardless of the time spent on the intervention.

Moodbuster 2.0, developed and tested through several European projects [36-38], is an online and mobile research platform designed for the delivery of online interventions. It includes both a patient and therapist portal and can be used in an unguided, guided, or blended format across various settings. For this study, Moodbuster 2.0 was tailored to better meet the needs of older adults experiencing depressive symptoms. The intervention content was shortened, and the language used throughout the modules was made age appropriate by using examples of everyday situations relevant to the age group, such as physical health changes or coping with retirement. In addition, video content was altered by featuring older adults and professionals discussing depression and CBT principles in a relatable manner.

Moodbuster 2.0 for older adults included 7 web-based CBT modules covering psychoeducation, behavioral activation, cognitive therapy, problem-solving, exercise promotion, sleep hygiene promotion, and relapse prevention. The module on sleep was included as sleep problems are a common contributor to persistent depressive symptoms in this age group [39]. Participants were advised to complete one module per week, starting with psychoeducation, followed by behavioral activation and cognitive therapy, while the other modules were used according to the preference of the participant. An online coach provided feedback once a week for each module through a secure messaging system. If participants were inactive or had not engaged with the platform for some time, the coach would proactively send messages to encourage and motivate them to resume the interventions.

All measures were conducted through web-based self-report questionnaires. Table 1 provides an overview of the assessments at each time-point. In addition to the assessments taken from participants, the Working Alliance Inventory-Short Form (WAI-SF) questionnaire was also filled out by the coaches after the assessment.

Digital literacy is the degree of a person’s knowledge, comfort, and perceived skills with a digital instrument such as a computer [56]. Digital literacy was measured using 2 subscales of the Digital Health Literacy Instrument (DHLI), developed by van der Vaart and Drossaert [56]. The 2 subscales that were used were “operational skills” and “the ability to add content.” The subscale “operational skills” assesses an individual’s ability to use basic digital tools, such as operating a computer or navigating the internet. The subscale “adding content” evaluated a person’s capacity to create and share content, such as posting messages or uploading information [56]. Each subscale consists of 3 items rated on a 4-point Likert scale, with response options ranging from 1 (“very easy”) to 4 (“very hard”), with total scores on both subscales ranging from 3 to 9. Higher scores indicate lower proficiency in operational skills and the ability to add content. The operational skills subscale demonstrated good reliability (Cronbach α=0.77), while the adding content subscale showed excellent reliability (Cronbach α=0.89) [56].

Data analyses were performed using SPSS (version 27; IBM Corp) [57]. Descriptive analyses were used to describe demographic characteristics, digital literacy at baseline, feasibility of the platform for usability, satisfaction, and engagement, as well as working alliance and technical alliance at after the assessment. To assess the difference in depression severity between the preintervention and postintervention measurements, a 2-tailed, paired sample t test was performed. The significance level used was P<.05. Effect sizes were calculated using Cohen d, with thresholds of 0.2, 0.5, and 0.8 interpreted as small, medium, and large effects, respectively [58].

The study was reviewed and approved by the METC of the VU University Medical Center Amsterdam (2021.0435). All procedures were conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants before participation. Participant data were processed in a coded and deidentified manner and stored securely in accordance with institutional and General Data Protection Regulation regulations; only authorized members of the research team had access to the data. Participants received a €30 (US $34.84) voucher as compensation for their participation. The reporting adhered to the CONSORT (Consolidated Standards of Reporting Trials) 2010 extension for pilot and feasibility studies to ensure methodological transparency [59].

Of the 231 potential participants (mean age 64.8, SD 6.5 years; n=162, 70.1% female), only 34 (14.7%) signed the paper-based informed consent form and were assessed for eligibility, of which 21 (61.8%) were eligible to take part (mean 7.76, SD 2.14). The CONSORT flowchart diagram of the study is shown in Figure 1. After 8 weeks, 90.5% (19/21) of the included participants completed the postassessment questionnaires.

The baseline characteristics of our participants are summarized in Table 2. The group was predominantly female (18/21, 85.7%), had an average age of 59.85 (SD 4.19; range 55-68) years, was highly educated (13/21, 61.9%), and all but one (20/21, 95.2%) were born in the Netherlands. Their marital status was mixed, and most of them had children (16/21, 76.2%), and all were in contact with them. Regarding their digital literacy, they subjectively reported their operational skills at 3.81 (on a scale from 3 to 9), which we interpret as “very good” as a higher score indicates lower skills (Table 2). However, they scored their ability to add content with an average score of 5.48 (SD 1.91, on a scale from 3 to 9), which is considered “moderate.” Concerning the level of depression, participants showed an average score of 8.61 (SD 4.72) on the PHQ-8, which falls within the range of mild depression.

Feasibility outcomes are presented in Table 3. Participants reported high satisfaction with the intervention, with a mean score of 20.58 (SD 5.01). Engagement, assessed using the TWEETS, was rated as moderate to high (mean 2.89, SD 0.95) with higher scores in the cognitive engagement subscale (mean 3.11, SD 1.03). However, the average usability score (SUS: mean 54.21, SD 24.00) fell below the accepted threshold (68 or higher).

The scores of the working alliance (WAI-SF) for participants are shown in Table 3. Participants rated the WAI-SF with their coach as “moderate.” In addition, coaches reported their experience working in alliance with participants. The 4 coaches guided 21 (range 1-10) participants. The coaches scored on average 3.12 out of 5. Subscale scores were highest for the bond dimension (mean 4.46, SD 0.78), followed by the task (mean 2.42, SD 0.60) and therapeutic goals (mean 2.35, SD 0.87).

Participants’ technical alliance (WAI-TECH-SF) is shown in Table 3. Participants rated the technical alliance on average 45.15 (SD 14.75) on a scale from 12 to 84.

Of the 21 participants who completed the baseline, 20 (95.2%) participants started the intervention. On average, participants logged in 19 (SD 19. 49) times, visited 62 (SD 39.89) pages out of the total 133 pages, and spent 117 (SD 159.52) minutes on the platform. Module completion varied considerably across the 8 modules. Nearly all participants completed the introduction (n=20, 95.2%) and psychoeducation (n=19, 90.5%) modules. Engagement then dropped, with 10 (47.6%) participants completing behavioral activation, and 4 (19%) participants completing cognitive therapy. Of the optional modules, 7 (33.3%) participants chose and completed problem-solving, followed by sleep and relapse prevention (n=4, 19%), and only 1 (4.8%) participant completed the exercise module. Only 1 (4.8%) participant finished all modules. Regarding time spent, the problem-solving and cognitive therapy modules showed the most time on average (mean 40.03, SD 26.57 and mean 39.31, SD 26.57 minutes, respectively). These modules also had the highest variation in completion time, showing individual differences in time spent in these modules. Further information about uptake can be found in Table 4.

At baseline, the participants reported mild depressive symptoms on the PHQ-8 (mean 8.61, SD 4.72). On average, the depressive symptoms of the group decreased by almost 2 points on the PHQ-8 (mean 6.72, SD 2.93). The paired sample t test revealed that this decrease was not significant (t₁₇=2.03, 95% CI –0.07 to 3.84; P=.06; Cohen d=0.47).