This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Mental Health, is properly cited. The complete bibliographic information, a link to the original publication on https://mental.jmir.org/, as well as this copyright and license information must be included.
Mental ill-health presents a major public health problem. A potential part solution that is receiving increasing attention is computer-delivered psychological therapy, particularly during the COVID-19 pandemic as health care systems moved to remote service delivery. However, computerized cognitive behavioral therapy (cCBT) requires active engagement by service users, and low adherence may minimize treatment effectiveness. Therefore, it is important to investigate the acceptability of cCBT to understand implementation issues and maximize potential benefits.
This study aimed to produce a critical appraisal of published reviews about the acceptability of cCBT for adults.
An umbrella review informed by the Joanna Briggs Institute (JBI) methodology identified systematic reviews about the acceptability of cCBT for common adult mental disorders. Acceptability was operationalized in terms of uptake of, dropping out from, or completion of cCBT treatment; factors that facilitated or impeded adherence; and reports about user, carer, and health care professional experience and satisfaction with cCBT. Databases were searched using search terms informed by relevant published research. Review selection and quality appraisal were guided by the JBI methodology and the AMSTAR tool and undertaken independently by 2 reviewers.
The systematic searches of databases identified 234 titles, and 9 reviews (covering 151 unique studies) met the criteria. Most studies were comprised of service users with depression, anxiety, or specifically, panic disorder or phobia. Operationalization of acceptability varied across reviews, thereby making it difficult to synthesize results. There was a similar number of guided and unguided cCBT programs; 34% of guided and 36% of unguided users dropped out; and guidance included email, telephone, face-to-face, and discussion forum support. Guided cCBT was completed in full by 8%-74% of the participants, while 94% completed one module and 67%-84% completed some modules. Unguided cCBT was completed in full by 16%-66% of participants, while 95% completed one module and 54%-93% completed some modules. Guided cCBT appeared to be associated with adherence (sustained via telephone). A preference for face-to-face CBT compared to cCBT (particularly for users who reported feeling isolated), internet or computerized delivery problems, negative perceptions about cCBT, low motivation, too busy or not having enough time, and personal circumstances were stated as reasons for dropping out. Yet, some users favored the anonymous nature of cCBT, and the capacity to undertake cCBT in one’s own time was deemed beneficial but also led to avoidance of cCBT. There was inconclusive evidence for an association between sociodemographic variables, mental health status, and cCBT adherence or dropping out. Users tended to be satisfied with cCBT, reported improvements in mental health, and recommended cCBT. Overall, the results indicated that service users’ preferences were important considerations regarding the use of cCBT.
The review indicated that “one size did not fit all” regarding the acceptability of cCBT and that individual tailoring of cCBT is required in order to increase population reach, uptake, and adherence and therefore, deliver treatment benefits and improve mental health.
Globally, mental health problems and mental disorders are a major public health concern [
The onset of a global pandemic of the SARS-CoV-2 virus occurred during the conduct of this umbrella review. Government-directed measures designed to control the transmission of the virus such as social distancing, quarantine, and self-isolation were implemented widely. These measures necessitated a move to remote-delivered therapies [
Mental health interventions such as CBT require active engagement from participants in order to affect therapeutic change. Adherence and the degree to which an individual engages with, and completes, a web-based intervention is a potential limitation. Poor adherence limits exposure to a web or computerized program, and an insufficient “dose” may impact treatment effectiveness and reduce the likelihood of an improved outcome for people with mental health problems, particularly if they drop out of a program. It is important to enhance understanding about acceptability and adherence in order to gain insights about the implementation of cCBT and to increase its effectiveness [
The search strategy for the umbrella review [
#1 ((internet or web or online) adj3 (cognitive or behavio*)).ti,ab,kf. OR (iCBT or i-CBT or ePsych* or e-Psych or cCBT or c-CBT).ti,ab,kf
#2 computer communication networks/ or internet/ or blogging/ or social media/ OR cell phones/ or smartphone/ or text messaging/ or videoconferencing/ or webcasts as topic/ or wireless technology/ OR Telemedicine/ OR (eLearning or blended learning).ti,kf. OR (videoconferenc* or video conferenc*).ti,kf OR (synchronous or asynchronous or (electronic adj2 deliver*)).ti,kf. OR android.ti,ab,kf. OR (app or apps or blog*).ti,ab,kf. OR (cell phone or cellphone or chat room or computer* or cyber* or digital or technology based or DVD).ti,ab,kf. OR CD-ROM.ti,ab,kf. OR (eHealth or electronic health or email*).ti,ab,kf. OR (ePortal or eTherap* or forum* or gaming or information technolog* or instant messag* or messaging or internet* or ipad or iphone or ipod or podcast or smart phone or smartphone or social network* site* or social networking or mHealth or mobile or multimedia or online* or personal digital assistant or PDA or SMS or social medi* or software or telecomm* or telehealth* or telemed* or telemonitor* or telepsych* or teletherap* or text messag* or texting or virtual* or web* or WWW).ti,ab,kf,hw.
#3 (behavio* or cognitive).ti. OR ((cognitive or behavio*) adj2 (activat* or component? or defusion or modif* or restructur* or technique* or intervention or treatment* or therap* or train*)).mp. OR (psychotherap* or psychological therap* or cognitive behavio*).mp. OR ((acceptance* or commitment*) adj3 therap*).mp. OR (rational emotive or RET or problem sol* or PST or problem focus* or solution focus* or trauma focus* or psychoeducat* or psychodrama or mindfulness* or third wave or self control).mp. OR (self* adj3 (control or efficacy)).mp. OR (stress manage* or exposure or reality therap*).mp. OR (anxiety adj3 (management or therap* or train*)).mp. OR (relaxation or guided imagery or present cent* or person cent* or person* construct* or therapeutic process* or schema? or schemata).mp. OR (thought* adj3 suppress*).mp. OR rumination.mp.
#4 #2 AND #3
#5 #1 OR #4
#6 “Systematic Review”/ OR systematic review.ti,ab,kf.
#7 #5 AND #6
#8 program evaluation.mp. OR program evaluation/ OR process evaluation.mp. OR “process assessment (health care)”/ OR (process evaluation* OR qualitative component* OR qualitative aspect* OR qualitative approach* OR systematic evaluation* OR participant observation OR simulation OR implementation audit).mp. OR (audit OR feedback).tw. OR qualitative research/ OR qualitative.mp. OR qualitative research.mp.
#9 #7 AND #8
#10 depression/ OR depression.mp. OR anxiety/ OR anxiety.mp. OR Obsessive-Compulsive Disorder/ OR obsessive compulsive disorder.mp. OR Mental Disorders/ OR common mental disorders.mp. OR Occupational Stress/ OR Stress Disorders, Post-Traumatic/ OR stress.mp. OR Stress, Psychological/ OR mental health.ti OR mental illness.ti. OR psychiatric.ti OR mood disorder.tw OR neurotic disorder.sh
#11 #9 AND #10
The target population was people with common mental disorders or experiencing an increase in psychological symptoms (eg, depression, anxiety, and distress) [
Reports of interventions had to indicate that that they were grounded in CBT in terms of, for example, aiming to improve or maintain mental health outcomes by changing aberrant thought patterns and behaviors. CBT interventions delivered via the internet or via computer only were included [
The scoping activity and the review team discussions identified the following key outcomes or indicators of acceptability: percentage of uptake or refusal following invitation to cCBT; adherence (number of completed intervention sessions or components); proportion of “drop-outs” (including reasons for, time of, and factors associated with dropping out); satisfaction with intervention components or mode of delivery; and user, carer, or health care professional reports about their views of cCBT.
There were no language nor year restrictions. Three non-English language papers were identified, but from examining available English abstracts, they did not meet the review criteria. Double counting of primary studies across reviews is an inherent bias of an umbrella review, and where present, efforts have been made to indicate study overlap. Some reviews included physical health conditions and non-cCBT interventions, and these findings were excluded from the umbrella review.
We followed best guidance on the conduct of umbrella reviews [
The database searches identified 234 titles: 23 full-text papers remained following duplicate removal and abstract screening, and a further 8 titles were identified through citation lists of the 23 reviews. Of the 31 papers, 9 met the eligibility criteria for inclusion in the umbrella review (see
Most review studies focused on adult populations (n=119), while a smaller number examined adolescent populations (n=8) and one study included both adults and adolescents [
Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart.
We examined 178 cCBT interventions across the reviews, with 23 studies examining more than one cCBT condition. Two reviews were not focused only on cCBT and included a small number of non-cCBT interventions (n=1) or did not specify the type of therapeutic intervention (n=2). Detailed descriptions about interventions including their content, format, and delivery were reported infrequently. It was possible to discern, where this information was reported, that there were 53 guided and 50 unguided interventions; guidance included email support (n=11), telephone support (n=9), face-to-face contact (n=9), and a discussion forum (n=3). Technical support was provided in 25 studies (see
Data on the proportion of people in each study who availed of the use of a cCBT intervention were provided in 5 reviews. Definitions of uptake varied widely and referred to qualitatively different types of individuals; for example, 1 review included the proportion of eligible individuals who were randomized and spontaneous users who accessed cCBT websites [
Adherence definitions and rates.
Study | Adherence definition | Rates of uptake | Proportion completing treatmenta | Drop-outs |
Kaltenthaler et al (2008) [ |
Patient recruitment; number of patients who accessed website and agreed to take part in study; drop-outs; number of patients who dropped out of the study | 25% agreed to take part | Type of program (% participants did not complete components, number of studies): Beating the Blues (26%-45%, 5 studies); MoodGym (17%-75%, 2 studies) website mass recruitment; ODIN (34%, 2 studies); COPE (29%-32%, 2 studies); Recovery Road (32%, 1 study); Five Areas Approach (30%, 1 study); BALANCE (11%, 1 study); two unnamed interventions (0%, 37%) | Mean 31.75% (SD 16.52%); range 0%-75% |
Waller and Gilbody (2009) [ |
Invited: those who received information about the study; recruited: those specifically asked to take part in the study; finished study: % of those who started and completed the study; finished modules: % of those who completed all modules | Median 38%; range 4%-84% | Started and subsequently completed: median 83%, range 26%-100%; overall completion: median 56%, range 12%-100% | Individuals in cCBT were twice as likely to drop-out than control groups (OR 2.03, 95% CI 0.81-5.09) |
Melville et al (2010) [ |
Drop-out: “…leaving treatment before its completion…at one of a number of points throughout treatment”; pretreatment drop-out: before beginning of treatment; treatment dropout: prior to completion of treatment sessions; follow-up drop-out: prior to completing follow-up assessments; rates of drop-out; number of eligible participants (denominator) and the number of individuals who terminated at any point from registering the treatment and completing follow-up questionnaires (numerator) | Not reported | Not reported | Pretreatment drop-out: 4%-52% (median 10%; weighted mediane 21%); treatment dropout: 0%-78% (median 10%; weighted median 21%); follow-up drop-out: 0%-18% (median 4%; weighted median 8%) |
Vallury et al (2015) [ |
Rates of uptake and rates of adherence | Uptake rate: 44%-86%; refusal rate: 9%-24%; did not start: 7% | Completed treatment: 33%-100%; completed follow-up at 3 months: 63% | Not reported |
Beatty and Binnion (2016) [ |
Quantitative studies: adherers were those who completed program or completed the posttreatment assessment; nonadherers were those who did not complete program or did not complete the posttreatment assessment; numbers of modules/sessions/assessments completed, duration of logins, time using program, number of logins, number of homework assignments completed, accessing the program; qualitative studies: those who do not complete treatment, barriers to adherence | Uptake rate: 41.3%; did not start: 1%-63% | Completed all sessions: 16.9%-83.0%, 35.8%-66.0% for unguided or unsupported interventions, 58.0%-74.4% for guided or supported interventions; completed only one session or module: 27%-90%, 94% for guided, 95% for unguided; completed some but not all sessions or modules: 10%-99%, 67%-84% for guided interventions, 54.4%-93.0% for unguided interventions; completed all assessments: 26.0%-90.8% | Early drop-out (during first sessions): 10.0%-56.9%, 37.1% for the treatment group, 32.1% for the control group; drop-outs at end of study: 4.25%-38.00%, 56.3%-75.0% for the treatment group, 29%-48% for the control group, 35.9% for guided, 33.7% for unguided, 56% of those who immediately accessed, 80% of those who delayed access |
Rost et al (2017) [ |
Uptake, drop-out, or completion rates as a means of assessing user acceptance | 39%-97% | Mean 67.17% (SD 20.29%), range 26.7%-100%; 8.1%, 56% guided intervention; 16.28%, 36% unguided intervention | Mean 31.5% (SD 19.49%), range 0%-63% |
Twomey and O’Reilly 2017 [ |
Proportion of participants withdrawing before final data collection, proportion of individuals who completed intervention | Not reported | 10%-100% | Withdrew before posttreatment data collection: 0%-64% |
aThis was defined differently across reviews; in some cases, this included proportions of those who started and subsequently completed treatment or follow-up assessments.
bOf the 16 studies, 10 were overlapping.
cOf the 36 studies, 15 were overlapping.
dOf the 19 studies, 6 were overlapping.
eWeighted for study sample size.
fOf the 11 studies, 1 was overlapping; 4 studies included adolescent populations.
gOf the 36 studies, 6 were overlapping; 6 studies included populations with physical health or other conditions and thus were excluded.
hOf the 29 studies, 6 were overlapping; 2 studies included adolescents.
iOf the 11 studies, 5 were overlapping.
Findings from quantitative studies (as categorized by review) referring to factors associated with adherence or dropping out.
Factor | Intervention (population) | Number of studies or participants | Results or findings (number of studies) | Heterogeneitya |
|
|
|
|
|
|
Internet-based cognitive behavioral therapy (CBT; subthreshold depression) | 8 studies | No difference between completers and noncompleters [ |
No association |
|
|
|
|
|
|
Internet-based treatment (psychological dysfunction or distress related to psychiatric conditions) | 2 studies | No association (1); men more likely to drop-out (1) [ |
Mixed results |
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 15 studies | Women more likely to adhere (7), men more likely to adhere (1), no association with gender (7) [ |
Not reported |
|
|
|
|
|
|
Internet-based treatment (psychological dysfunction or distress related to psychiatric conditions) | 2 studies | No association (1)d, younger age associated with drop-out (1) [ |
Mixed results |
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 14 studies | No association (6)d, older age associated with adherence (4), younger age associated with adherence (3), mixed findings (1) [ |
Not reported |
|
|
|
|
|
|
Internet-based treatment (psychological dysfunction or distress related to psychiatric conditions) | 2 studies | No association (2 [ |
No association |
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 13 studies | No association (8)d, higher education associated with higher adherence (4), lower education associated with higher adherence (1) [ |
Not reported |
|
|
|
|
|
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 1 study | No association [ |
Not enough evidence |
|
|
|
|
|
|
Internet-based treatment (psychological dysfunction or distress related to psychiatric conditions) | 1 study | Being partnered associated with drop-out (1) [ |
No association |
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 8 studies | No association with having a partner (6), being partnered associated with adherence (2) [ |
No association |
|
|
|
|
|
|
Computerized CBT (cCBT; mild or moderate depression) | 1 study | Getting a job provided as reason for dropping out (2) [ |
No association |
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 7 studies | No association (7) [ |
Not reported |
|
|
|
|
|
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 1 study | Oceania or Europe residency associated with higher completion of modules than residency in North America, South America, and Africa (1) [ |
Not enough evidence |
|
|
|
|
|
|
cCBT (prevention of anxiety and depression) | 2 studies | Adherence and/or attrition among rural compared to urban participants, rurality had no effect (1) or negative effect (1) on retention to cCBT [ |
Not enough evidence |
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 2 studies | Within Ireland and Australia, no association (2) [ |
Not reported |
|
|
|
|
|
|
cCBT (mild or moderate depression) | 2 studies | Family reasons (3) or change in circumstances (15%) or moving house (10%) provided as reasons for dropping out [ |
Limited evidence |
|
cCBT, (common mental health disorders) | 9 studies (126 participants) | Personal circumstances were stated as a reason for decliningd [ |
Not reported |
|
|
|
|
|
|
cCBT (mild or moderate depression) | 2 studies | Being too busy (8) provided as reason for dropping out [ |
Limited evidence |
|
cCBT (depression) | 6 studies | Lack of time reported (6) as reason for dropping out [ |
Not reported |
|
|
|
|
|
|
cCBT (mild or moderate depression) | 1 study | Ill-health (15%) provided as reason for dropping out of study [ |
Not enough evidence |
|
|
|
|
|
|
Internet-based CBT (subthreshold depression) | 8 studies | No difference between completers and noncompleters [ |
Not enough evidence |
|
|
|
|
|
|
cCBT (mild or moderate depression) | 1 study | Improvement in symptoms was reported as a reason for dropping out (2)e [ |
Mixed evidence |
|
cCBT (common mental health disorders) | 2 studies (2 participants) | Perceived increased risk (potential increase in symptoms) reported as reasons for drop-out d [ |
Mixed evidence |
|
Internet-based treatment (psychological dysfunction or distress related to psychiatric conditions) | 3 studies | Lower symptom severity associated with dropping out of study (3 [ |
Not reported |
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 20 studies | At baseline: no association (10), lower symptom severity associated with increased adherence (6) and increased module completion (1); higher symptom severity associated with higher adherence (3) [ |
Not reported |
|
|
|
|
|
|
cCBT (mild or moderate depression) | 1 study | Improvement in condition (10%) provided as reason for dropping out of study [ |
Mixed evidence |
|
Internet-based treatment (psychological dysfunction or distress related to psychiatric conditions) | 2 studies | No association (2) [ |
Not reported |
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 3 studies | Longer duration associated with higher adherence (2), no association (1) [ |
Not reported |
|
|
|
|
|
|
Internet-based treatment (psychological dysfunction or distress related to psychiatric conditions) | 2 studies | Comorbid depression and anxiety no association (2) [ |
Not enough evidence |
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 4 studies | Diagnosis of anxiety or depression associated with higher adherence (3), and alcohol dependency associated with higher adherence among waitlist control group (1) [ |
Not enough evidence |
|
|
|
|
|
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 2 studies | No association (2; studies not targeting alcohol dependency) [ |
Not enough evidence |
|
|
|
|
|
|
cCBT (mild or moderate depression) | 4 studies | Perception that intervention is not useful (11), unhelpful (10), didn’t like treatment (n not reported), inappropriate for needs (1) provided as reason for dropping out of studye [ |
Mixed evidence |
|
cCBT (common mental health disorders) | 11 studies (101 participants) | Therapy was reported as a reason for dropping out, but it is not clear what this meant or which group it referred to [ |
Not reported |
|
Internet-based treatment (psychological dysfunction or distress related to psychiatric conditions) |
2 studies | No association (2) [ |
Not reported |
|
cCBT (prevention of anxiety and depression) | 1 study | Treatment preference fulfilment was associated with adherence to the study for rural residing participants [ |
Not reported |
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 9 studies | Positively associated with higher adherence (7), no association (2) [ |
Not reported |
|
cCBT (depression) | 6 studies | Treatment being perceived as inconvenient (4) was reported as a reason to drop out [ |
Not reported |
|
|
|
|
|
|
cCBT (mild or moderate depression) | 2 studies | Low motivation (8), inability to commit (n not reported), and no desire to continue (n not reported) provided as reasons for dropping out [ |
Limited evidence |
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 4 studies | Treatment readiness associated with higher adherence (2), and intention to complete treatment associated with higher adherence (1); intention to complete treatment no association (1) [ |
Not reported |
|
|
|
|
|
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 3 studies | No association (2), taking responsibility for one’s own choices was associated with higher adherence to a bulimia self-guided program (1) [ |
Not enough evidence |
|
|
|
|
|
|
cCBT (mild or moderate depression) | 2 studies | Internet-related issues (5), changed mind about PC delivery (1) provided as reasons for dropping out [ |
Mixed evidence |
|
cCBT (common mental health disorders) | 5 studies (14 participants) | Information technology issues were not commonly reported as a reason for dropping out [ |
Not reported |
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) |
4 studies | Higher adherence was associated with website usability (1) and a positive attitude to a computerized self-guided format (1), no association between adherence and computer literacy level (2) [ |
Not reported |
|
cCBT (depression) | 4 studies | Computer or technical issues (4) were reported as reasons for dropping out of the study [ |
Not reported |
|
|
|
|
|
|
cCBT (mild or moderate depression) | 1 study | Preference for face-to-face help (8) provided as reason for dropping out [ |
Guided intervention associated with higher adherence |
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 8 studies | Guided interventions were associated with higher adherence than unguided interventions (4), phone support was associated with higher adherence than email support (1), no difference in adherence between guided and unguided interventions (3) [ |
Not reported |
|
|
|
|
|
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 3 studies | General practitioner referral (2) or the media (1) associated with higher adherence [ |
Not enough evidence |
|
|
|
|
|
|
cCBT (mild or moderate depression) | 2 studies | Other help sought provided as reason for dropping out (2), treatment not demanding (n not reported) [ |
Not enough evidence |
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 2 studies | Gratitude intervention group was twice as likely to complete treatment than a monitoring and restructuring intervention group (1); tailored feedback to increase self-efficacy, personalization of intervention team (eg, photo with “we”) increased adherence for participants who accessed all intervention components (1) [ |
Not reported |
|
|
|
|
|
|
cCBT (mild or moderate depression) | 2 studies | Hard to attend (13) or journey too long (3) provided as reasons for dropping out [ |
Not enough evidence |
|
|
|
|
|
|
cCBT (mild or moderate depression) | 16 studies | Duration influenced drop-out (direction of effect not reported in review); duration of interventions ranged from 1-33 sessions; the authors note that it is difficult to make comparisons between cCBT programs regarding drop-out rates because of differences in study design, populations, and methods for defining drop-outs and level of detail provided in a study [ |
Duration associated with adherence |
|
|
|
|
|
|
cCBT (mild or moderate depression) | 1 study | Intervention group had higher drop-outs compared to participants in the information website group [ |
Control groups more likely to adhere than intervention groups |
|
cCBT (common mental health disorders) | 9 studies | Participants in cCBT were twice as likely to drop out than participants in the control group (OR 2.03, 95% CI 0.81-5.09) [ |
Not reported |
|
Self-directed psychological intervention (psychological outcomes for mental and physical disorders) | 4 studies | Control (waitlist) group membership predicted higher adherence than intervention group membership (1), intervention group membership associated with adherence (1), no association with group membership (2) [ |
Not reported |
aAt least 50% of studies needed to provide evidence of or absence of an association; there was not enough evidence if <5 studies available.
bAssociation with adherence.
cAssociation with drop-out.
dOverlapping studies.
eNumber of participants (where reported in reviews).
The definitions of adherence varied between and within reviews. For example, some reviews included studies that defined adherence in terms of “completers” or individuals who completed an outcome assessment at the end of receipt of cCBT while other studies defined “treatment completers” as individuals who completed all “modules” or a specified “dose” of cCBT even if they did not undertake a research outcome assessment [
Heterogeneity in terms of the proportion of individuals who completed cCBT was present for guided and unguided cCBT (guided cCBT: completed full program, 8%-74%; completed 1 module only, 94%; completed some modules, 67%-84%; and unguided cCBT: completed full program, 16%-66%; completed 1 module only, 95%; completed some modules, 54%-93%) [
Two reviews reported (without providing details about timing) that, on average, around one-third of users “dropped out” [
The proportion of service users who dropped out of guided (36%) and unguided (34%) cCBT programs was similar [
Across 2 reviews, 7 of 16 studies (and 1 overlapping study) reported the absence of an association between age and adherence [
The association between mental health symptom severity and adherence was unclear [
cCBT features and adherence were examined in 4 reviews. Guided compared to unguided cCBT was associated with higher adherence (4/7 studies), or there was no association (3/7) [
Behavioral factors and adherence were examined in 6 reviews. Negative perceptions about cCBT (eg, as unhelpful or inappropriate) were associated with dropping out [
Generally, users appeared to be highly satisfied [
Predominantly based on qualitative data (with the exception of quantitative data in 1 review [
A qualitative review containing a meta-synthesis identified 2 “core constructs” to explain factors that impeded or facilitated participation in cCBT: “sensitivity of self and identity” and “the dialectical nature of user experience.” First, the preferences, needs, and challenges that an individual with mental ill-health faces (eg, reduced motivation and concentration) need to be considered before deciding that cCBT is an appropriate option or to inform how to tailor the delivery of cCBT. Second, the delivery of cCBT needs to take into account the contradictory presence of perceived benefits and drawbacks (eg, the option to self-select modules may be empowering and burdensome, and cCBT may be perceived as enhancing confidentiality and privacy or as an obstacle to meeting the need for face-to-face support from therapists and peers) [
This umbrella review investigated the acceptability of cCBT, in terms of the proportion of people who availed of, adhered to, or dropped out of cCBT as well as with reference to qualitative reports of the reasons for dropping out and users’ and therapists’ views. Nine reviews provided relevant data, though the aforementioned indicators of acceptability were defined and operationalized heterogeneously across reviews. The extent of this variation made it difficult to “sum up” the results from the 9 reviews and to compare acceptability across populations, conditions, programs, and settings. Several reviews reported that cCBT retention proportions (eg, face-to-face CBT drop-out range, 5%-38% and cCBT mean drop-out, 33%) were comparable and as variable as psychological therapy delivered face-to-face [
Often, reasons were not given for dropping out of studies, so it is difficult to draw firm conclusions regarding why adherence to cCBT was difficult for some individuals. It is important to understand the reasons for dropping out and whether they vary at different stages. For example, did individuals withdraw because cCBT was not what they expected, treatment was too intensive, symptoms improved (whether or not improvement was attributed to cCBT), or for other reasons separately or in combination. Included reviews did not present data regarding why individuals chose not to take up cCBT at all or that illuminated a deeper understanding about reasons for dropping out. Developing strategies to increase uptake, adherence, and retention with respect to cCBT requires a clear understanding about the factors associated with adherence or reasons for dropping out. Most reviews gathered quantitative data, and it is likely that a mixed methods research approach would help to advance our understanding about the reasons for problems and challenges. Indeed, the quantitative results were inconsistent regarding sociodemographic variables, health status, cCBT characteristics, and behavioral factors and their association with cCBT adherence.
The qualitative synthesis presented by Knowles et al [
There is a need for further research to improve understanding about “dosage” and dropping out. The review points clearly to the conclusion that “one size does not fit all” and that cCBT even in relatively small “doses” may work well for some patients including patients who drop out at various stages, whereas a “full dose” and perhaps even more again is needed by other patients. For example, a recent randomized controlled trial and intention-to-treat analysis found a beneficial effect for the cCBT treatment of insomnia despite a high drop-out rate and increasingly lower adherence to modules over the course of the study. It is important to note that the study was comprised of a student population and most students experienced subclinical insomnia [
Only limited attention was given to the potential for theory to illuminate understanding about the acceptability of cCBT [
Though a protocol was developed for this review, it was not registered on a registry such as Prospero. However, this umbrella review was informed by an established rigorous methodology for the “summing” of the increasing number of systematic reviews about a given topic and involves independent screening, appraisal, and data extraction. There was a consensus among review authors that study or program heterogeneity and variability and inadequate reporting of details in primary studies restricted meaningful analytical comparisons between cCBT programs, and we relied upon authors’ reports of study and data type (eg, qualitative or quantitative) and use of validated outcome measures of acceptability or satisfaction. Moreover, there is a lot of variance in terms of the labels and descriptions used to describe internet or computerized interventions such as cCBT, and this lack of consensus may contribute to difficulties in research synthesis [
Collectively, the results of the reviews indicate that “one size does not fit all” regarding the acceptability of cCBT and that individual patient and service user group tailoring of cCBT are required to increase the population reach and uptake of cCBT and adherence and so, deliver treatment benefits and improve mental health. In turn, tailoring to create as close a patient-treatment match as possible is likely to involve a blending process involving a human guide or therapist.
Supplementary table 1: quality appraisal.
Supplementary table 2: included review characteristics.
Supplementary table 3: satisfaction and acceptability.
Supplementary table 4: qualitative findings.
A Measurement Tool to Assess Systematic Reviews
cognitive behavioral therapy
computerized cognitive behavioral therapy
Consolidated Standards of Reporting Trials
internet-delivered cognitive behavioral therapy
information technology
National Institute for Health and Clinical Excellence
Preferred Reporting Items for Systematic Reviews and Meta-Analyses
Template for Intervention Description and Replication
This study was supported by the Academy of Finland (grants 315343 and 330527) and the Strategic Research Council at the Academy of Finland (grants 327145 and 327148 for the DigiIn Project). MD and AK received funding from the Economic and Social Research Council (ESRC) (grant ES/S00744X/1) during the writing of this article.
None declared.