Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Aug 1.
Published in final edited form as: Aging Ment Health. 2019 Mar 27;24(8):1196–1206. doi: 10.1080/13607863.2019.1590309

Internet-Delivered Cognitive Behavioral Therapies for Late-Life Depressive Symptoms: A Systematic Review and Meta-analysis

Xiaoling Xiang 1, Shiyou Wu 2, Ashley Zuverink 1, Kathryn N Tomasino 3, Ruopeng An 4, Joseph A Himle 1,5
PMCID: PMC7529149  NIHMSID: NIHMS1572168  PMID: 30913898

Abstract

Background:

This study aimed to review and synthesize evidence related to the effectiveness of internet-based cognitive behavioral therapy (iCBT) for reducing depressive symptoms in older adults.

Method:

The authors conducted a systematic review of intervention studies testing iCBT for symptoms of depression in older adults. An initial search of PubMed, PsychINFO, and Web of Science was undertaken, followed by a manual search of reference lists of the relevant articles. The Cochrane Risk of Bias Tool was used to appraise study quality. The mean effect size for included studies was estimated in a random effects model. Meta-regression was used to examine potential moderators of effect sizes.

Results:

Nine studies met the inclusion criteria, including 1272 participants averaging 66 years of age. The study design included randomized controlled trials (k=3), controlled trials without randomization (k=2), uncontrolled trials (k=2), and naturalistic evaluation (k=2). Seven studies tested iCBT with some level of therapist involvement and 2 examined self-guided iCBT. Six studies tested interventions specifically adapted for older adults. The mean within-group effect size was 1.27 (95% CI=1.09, 1.45) and the mean between-group effect size was 1.18 (95% CI=0.63, 1.73). Participants’ age was negatively associated with within-group effect sizes (b=−0.06, p=.016).

Conclusions:

iCBT is a promising approach for reducing depressive symptoms among older adults with mild to moderate depressive symptoms. However, studies involving older adults in iCBT trials were limited, had considerable heterogeneity, and were of low quality, calling for more studies with rigorous designs to produce a best-practice guideline.

Keywords: cognitive behavioral therapy, computer-assisted therapy, depression

INTRODUCTION

Cognitive Behavioral Therapy (CBT)-based interventions are effective in treating depression in older adults with large effect sizes reported from meta-analyses (Cuijpers, van Straten, & Smit, 2006; Jayasekara et al., 2015; Jonsson et al., 2016; Orgeta, Brede, & Livingston, 2017). However, access to psychological treatment modalities for depression is often out of reach for most older adults due to shortages of specialist providers, cost, transportation, stigma, or other access issues (Choi & Gonzales, 2005; Weinberger, Mateo, & Sirey, 2009). Internet-delivered CBT (iCBT) is a promising approach to address access barriers to psychotherapy. iCBT, also known as computerized CBT (CCBT), online CBT, or eCBT, is psychotherapy based on CBT principles delivered via websites (i.e., web-based) or dedicated apps on mobile devices or tablets. Proponents of iCBT praise its low cost, efficiency, accessibility, and ability to be tailored to the specific needs of a patient while delivering the core techniques of the therapy in a standardized and consistent way from one patient to another (Eells, Barrett, Wright, & Thase, 2014). A typical iCBT program consists of six to 15 modules that mirror sessions in face-to-face CBT and requires little therapist involvement beyond guidance and feedback on homework assignments (about 10–15 minutes per patient per week) (Andersson, Cuijpers, Carlbring, Riper, & Hedman, 2014). Meta-analyses of iCBT have shown that both self-guided and therapist-guided iCBT programs are effective in reducing depressive symptoms in mixed-age samples (Hedman, Ljótsson, & Lindefors, 2012; Karyotaki et al., 2017; Păsărelu, Andersson, Bergman Nordgren, & Dobrean, 2017; Renton et al., 2014) and that therapist-guided iCBT and face-to-face CBT have equivalent overall treatment effects in mixed-age samples (Andersson et al., 2014).

As older adults are more digitally connected than ever and increasingly turning to the internet for health information (Anderson & Perrin, 2017), iCBT has the potential to significantly improve the accessibility of evidence-based psychotherapy among older adults, an age group with the lowest utilization of psychotherapy. However, the effectiveness of iCBT for late-life depression has not been assessed in systematic reviews or meta-analyses. Older adults were rarely included in randomized controlled trials of iCBT before 2010 (Crabb et al., 2012). The representation of older adults in iCBT trials has increased in recent years and a systematic assessment of the evidence base is warranted to enable informed decisions regarding the clinical implications of the findings and future research directions. This study aimed to systematically review the literature on iCBT for older adults, determine whether iCBT is effective for improving symptoms of depression in older adults, and, where possible, perform a meta-analysis.

METHODS

Inclusion and Exclusion Criteria

This review included intervention studies that tested the effectiveness of iCBT for depression or depressive symptoms among older adults. Studies that tested transdiagnostic iCBT interventions designed to treat core symptoms of anxiety and depression in mixed depression and anxiety samples were included if 80% or more of their samples had at least mild depressive symptoms or concerns of depression. Detailed inclusion and exclusion criteria were discussed as follows.

Types of study designs.

Our review included intervention studies that were randomized controlled trials (RCTs), non-randomized controlled trials with active or non-active controls, and pre-post studies without a control group. Active controls referred to a treatment-as-usual or standard-of-care control condition such as face-to-face CBT. Non-active controls referred to a waitlist or non-treatment control. Given that older adults have been historically underrepresented in iCBT trials (Crabb et al., 2012), we included a variety of study designs to capture the full scope of relevant literature.

Types of participants.

The review included studies in which all patients were aged 50 years and older and their mean age was 60 years or older. A 50-year age cut-off was used in a previous meta-analysis of CBT for depression in older adults (Gould, Coulson, & Howard, 2012), as many North American trials of older adult populations use a cut-off of 50 or 55 years. Studies using mixed-age samples were eligible only if the analyses were stratified by age groups so that an effect size specific to older adults could be calculated. Participants were classified as having a depression concern if they had (1) a diagnosis of depression according to the Diagnostic and Statistical Manual of Mental Disorders, the International Classification of Diseases, or the Research Diagnostic Criteria, or (2) symptoms of depression on self-reported or clinician-administered standardized rating scales, or (3) difficulty with depression based on self-reported measures.

Types of interventions.

We included psychological interventions from the psychotherapeutic school of CBT (Dobson, 2010; O’Donohue & Fisher, 2009). The main component of the eligible interventions must have been delivered via websites (i.e., web-based) or dedicated apps on mobile devices or tablets. Treatments delivered via teleconferencing were excluded from the review because such treatments are too similar to conventional face-to-face CBT in that therapists deliver the intervention content and follow largely the same procedures during a scheduled appointment as they do in front of a camera (Hedman et al., 2012).

Types of outcome measures.

The primary outcome of interest was depression symptomatology. Studies that used self-reported or clinician-rated measures of depressive symptoms were both included. Self-rated assessments frequently used in treatment studies included the Geriatric Depression Scale (GDS) (Yesavage et al., 1982), Patient Health Questionnaire (PHQ) (Kroenke, Spitzer, Williams, & Löwe, 2010), and Beck Depression Inventory (Beck, Steer, & Carbin, 1988). The Hamilton Depression Rating Scale (Hamilton, 1960) is an example of a clinician-rated assessment tool.

Other exclusion criteria.

We excluded studies in which the intervention was not based on CBT, there were insufficient data to compute effect sizes, and studies not published in English.

Search Strategy

An experienced librarian and a trained research assistant performed systematic searches from September 2017 to November 2017 using PubMed, PsychINFO, and Web of Science. Searches were re-run and articles extracted in January 2018. We searched using both Medical Subject Headings and text word terms to maximize retrieval of relevant articles (Jenuwine & Floyd, 2004). These included, but were not limited to, terms related to depression (e.g., depress*, dysthymi*, adjustment disorder, mood disorder, affective disorder, and affective symptoms) combined with terms related to older adults (e.g., aged, elder, late life, old, geriatric, aging, and older), CBT (e.g., cognitive-behavioral therapy, CBT, behavioral therapy, behavioral modification, cognitive therapy, cognitive restructuring, acceptance and commitment, problem solving therapy, and behavioral activation), and the internet or computers (e.g., internet, web, online, computer, computerized, mHealth, automated, iCBT, and cCBT) (see Appendix for details). The search was limited to English-only results from 1990 to the present. Subsequent to the database search, we did manual searches of the reference lists of previously identified relevant articles and prior systematic reviews of iCBT interventions.

Study Selection

Two reviewers (XX and AZ) independently screened titles and abstracts for eligibility based on aforementioned inclusion and exclusion criteria. All publications of potential relevance were retrieved in full text. The first author (XX) read the texts retrieved by both reviewers and finalized the list of articles for inclusion.

Data Extraction

Two reviewers (XX and SW) independently extracted data regarding study characteristics and outcomes and inserted these data into tables. Disagreements regarding the relevance of study data were resolved by discussion. Information extracted from the studies included (1) sample (inclusion/exclusion criteria, sample size, and demographic characteristics); (2) treatment (therapy type and model, length of sessions and frequency, delivery, therapists’ backgrounds, length of trial, and follow-up); (3) type of comparator; (4) study design; and (5) outcome and measures. Attempts were made to contact study authors when crucial study details were missing.

Quality Appraisal

The Cochrane Risk of Bias Tool (Higgins et al., 2011) was used to appraise the quality of included studies. The Cochrane Tool includes 6 main bias domains to assess 5 types of bias: selection, performance, detection, attrition, and reporting. The domains are random sequence generation and allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), and selective outcome reporting (reporting bias). We did not assess blinding of participants and personnel because such blinding is not possible in the types of treatments we examined. Two reviewers (XX and SW) independently assessed study quality. Disagreements regarding study quality were resolved through discussion.

Data Analysis

Effect size.

We calculated the standardized mean difference for effect sizes on depression symptomology using Cohen’s d (Cohen, 1988). For each study, within-group effect sizes were calculated based on the standardized mean difference between the pre- and post-treatment depression scores for the treated. For studies with a treatment-as-usual (TAU) or waiting-list-control (WLC) comparator, between-group effect sizes were calculated based on the standardized mean difference between the treatment and comparison groups at post-treatment. If mean or standard deviations were not provided, effect sizes were calculated from the t or F statistics. For studies that included multiple depression outcome measures, only the primary measure identified by the studies was used to calculate effect sizes. The intention-to-treat principle was applied to calculate effect sizes in all but one studies (see Table 1).

Table 1.

Summary of characteristics of included studies

Study name Design Na Age mean Country Treatment Tailored to older adults? Comparator No. lessons (duration) Adherence (% completed all lessons) Primary depression measure Follow-up ITT?
Dear (2013) US 20 63 Australia TG Yes NA 5 (8 weeks) 80% PHQ-9 3 months Yes
Dear (2015) US 20 67 Australia SG Yes NA 5 (8 weeks) 80% PHQ-9 3 months Yes
Hobbs (2018) NE 69 ≥65 Australia TG No NA 6 (12 weeks) 65% PHQ-9 NA Yes
McMurchie (2013) NRS 53 72 UK SG No TAU 8 (8 weeks) 73% GDS-30 1 month Yes
O’Moore (2017) RCT 69 62 Australia TG No TAU 6 (10 weeks) 84% PHQ-9 3 months Yes
Staples (2016) NE 516 66 Australia TG Yes NA 5 (8 weeks) 75% PHQ-9 3 months Yes
Titov (2015) RCT 52 65 Australia TG Yes WLC 5 (8 weeks) NA PHQ-9 3 & 12 months Yes
Titov (2016) RCT 433 66 Australia iTG Yes iSG & SG 5 (8 weeks) 92% for iTG; 79% for iSG/SG PHQ-9 3 months Yes
Tomasino (2017) NRS 40 70 USA TG; TG+PS Yes WLC 16 (8 weeks) 55% for TG; 61% for TG+PS PHQ-9 NA No
Open in a new tab

Note. NA=not applicable or not available; NE=natural evaluation; RCT=randomized controlled trial; NRS=non-randomized controlled study; WLC=wait-list control; AC=attention control; TAU=treatment as usual; iSG=self-guided iCBT with clinician interview; SG=self-guided iCBT; TG=therapist-guided iCBT; TG+PS: therapist-guided iCBT plus peer support; iTG=therapist-guided iCBT plus clinician interview; US=uncontrolled study (pre-post design); NE=naturalistic evaluation; PHQ=Patient Health Questionnaire; GAD=Geriatric Depression Scale. ITT=intention-to-treat; ITT was applied in this meta-analysis for studies that marked with “yes”.

a.

Total number of participants included in the outcome analysis in each study. This number may differ from the number of eligible participants who went through treatment allocation.

Meta-analysis.

We performed meta-analyses to investigate (a) the overall within-group effect on depression outcomes, (b) the overall between-group effect on depression outcomes, and (c) the potential moderators of the treatment effects. As we expected heterogeneity among studies, we estimated a random effects model to calculate effect sizes. In each set of meta-analysis, only one effect size from each study was included so that data points were independent. For studies with more than 2 arms, we analyzed the multiple treatment arms as one group if there was no significant difference in depression outcomes between the treatment groups and if such analysis was permitted with available data. When data were unavailable to compute combined treatment effect from multiple arms, we calculated the effect size using the treatment arm with the largest sample size. In addition, we used between-group effect sizes in the pooled analyses only when the comparators were nonactive controls or active controls (i.e., TAU or WLC) that were not equivalent to iCBT or face-to-face CBT. This was necessary to ensure that the between-group effect sizes were comparable across studies.

We performed a meta-regression to examine potential moderators of treatment effects, including patient-level characteristics (% female, mean age, a dichotomous indicator of whether persons with severe depression were excluded from the study, and if applicable, baseline mean PHQ-9 scores), treatment-level characteristics (a dichotomous indicator of whether the treatment was tailored to older adults and a dichotomous indicator of whether the treatment was therapist-guided or self-guided), and study-level characteristics (a dichotomous indicator of whether the study received a “low risk of bias” rating on 3 or more bias domains).

Study heterogeneity was assessed using the I2 index (Higgins & Thompson, 2002). The level of heterogeneity represented by the I2 index was interpreted as small (I2 ≤ 25%), moderate (25% < I2 ≤ 50%), large (50% < I2 ≤ 75%), or very large (I2 > 75%). Publication bias was assessed using Begg’s test and Egger’s test (Egger et al., 1997) and visual inspection of the funnel plot.

Statistical analyses were conducted using Stata 15.1 SE (StataCorp, College Station, TX). Specific Stata commands included metan, metabias, and metreg. All analyses used two-sided tests, and P values lower than 0.05 were considered statistically significant.

RESULTS

Study Selection

Figure 1 presents the study selection process in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis statement (Moher et al., 2009). The top reason for exclusion was that some studies did not test iCBT in older adults. Two studies out of the 9 studies included in the review used data from the same clinical trial. Dear et al. (2015) reported findings from the same clinical trial as Dear et al (2013); however, the 2015 study analyzed data from the delayed-treatment waitlist control group to assess the preliminary effectiveness of self-guided iCBT, whereas their 2013 study analyzed data from the treatment group to assess the preliminary effectiveness of therapist-guided iCBT. Therefore, these two articles were treated as separate studies in this review. Staples, Fogliati, Dear, Nielssen, and Titov (2016) was a naturalistic evaluation in which the effectiveness of an iCBT program implemented in routine clinical care was compared with the same intervention examined in a previous RCT conducted by Titov et al. (2016). These two articles were treated as different studies in this review.

Figure 1.

Figure 1.

Open in a new tab

Flowchart for the search and study selection process

Study Characteristics

All studies were published between 2013 and 2018. Types of study design included RCTs (k=3), controlled trials without randomization (k=2), uncontrolled trials (k=2), and naturalistic evaluation (k=2) where researchers assessed pre- and post-treatment changes in depressive symptoms among patients prescribed an iCBT program by their clinicians in routine clinics. Sample sizes ranged from 20 to 516, for a total of 1,272 older adults. The mean age of participants was about 66 years across studies. Seven studies were conducted in Australia and 1 study apiece was conducted in the US and the UK. (Table 1)

All the included studies tested multi-component iCBT interventions that had weekly modules typically involving psychoeducation, behavioral activation, cognitive restructuring, relapse prevention and homework assignments, although the length of the modules and the extent to which CBT techniques were covered varied across studies. Most interventions included some level of therapist involvement (k=7) whereas 2 studies examined self-guided treatment (Dear, Zou, Ali, et al., 2015; McMurchie, Macleod, Power, Laidlaw, & Prentice, 2013). Six studies tested interventions that were specifically adapted for older adults in the form of age-appropriate case stories and examples of skill use, whereas three studies tested interventions with no adaptations (Hobbs, Joubert, Mahoney, & Andrews, 2018; McMurchie et al., 2013; O’moore et al., 2018). Comparators varied among studies that included controls (k=5). One study was a 3-arm superiority trial comparing therapist-guided and self-guided iCBT with different levels of therapist involvement (Titov et al., 2016). Two studies compared treatment groups with wait-list groups (Titov, Dear, Ali, et al., 2015; Tomasino et al., 2017), while two studies compared with treatment-as-usual (McMurchie et al., 2013; O’moore et al., 2018). The number of modules ranged from 5 to 16 and intervention duration ranged from 8 to 12 weeks. Adherence to iCBT, defined as the percentage of participants completing all modules, ranged from 55% to 92%. The PHQ-9 was the most popular measure for assessment of depressive symptoms (k=8). (Table 1)

The majority of study participants were female. All but one studies included adults aged 60 or older whereas one study included participants as young as 50 years of age (O’moore et al., 2018). The proportion of participants with tertiary education (i.e., post-high school education) ranged from 45% to 98%. Four studies assessed depression inclusion criterion based on either self-reported standardized rating scales (Dear et al., 2013; McMurchie et al., 2013; Tomasino et al., 2017) or diagnostic structured interview (O’moore et al., 2018), while the rest relied on self-reported difficulties with depression (k=5). Nevertheless, in three of the studies that used self-reported difficulties with depression, over half of their study samples met study criteria for major depressive disorder or major depressive episode (Dear, Zou, Ali, et al., 2015; Hobbs et al., 2018; Titov, Dear, Ali, et al., 2015) and in the rest of the studies, over 90% of the study samples had at least mild depressive symptoms according to the PHQ-9 (Staples et al., 2016; Titov et al., 2016). In the 8 studies that used PHQ-9, baseline mean PHQ-9 scores among participants in the treatment group ranged from 10.5 to 14.0. In the one study that used GDS, baseline mean GDS-30 score was 21.1 (McMurchie et al., 2013). Most studies explicitly excluded persons with severe depression (k=5) and those with active suicidal ideation (k=8). All but one study (McMurchie et al., 2013) expected participants to have prior experience with computers and access to a computer and the internet. (Table 2)

Table 2.

Detailed characteristics of population and intervention in included studies

Study name Sample Selected Inclusion/Exclusion Criteria Intervention
Dear (2013) 60–80 years old;
65% Female;
45% tertiary education;
Baseline mean PHQ-9=14.0;
85% met DSM-IV for MDE		 Inclusion: PHQ-9>9; computer/internet access
Exclusion: illicit substances or ≥3 drinks/day; psychotic mental illness; severe depression or suicidal; taking regular benzodiazepines; if taking a medication, not on a stable dose of antidepressants for at least 1 month		 Therapist-guided iCBT: “Manage Your Mood Course”
8-week intervention: 5 online lessons, lesson summaries and homework, and automatic reminders; features age-appropriate case stories and examples of skill use; plus, weekly telephone or secure messaging with a therapist.
Therapist support: Clinical psychologist (Ph.D.); tasks including sending and reading emails and telephoning participants; therapist spent an average of 78 minutes (SD=36.1) per patient.
Dear (2015) 62–76 years old; 70% Female;
50% tertiary education;
Baseline mean PHQ-9=12.4;
85% met DSM-IV for MDE		 Inclusion: self-reported difficulties with depression; computer/internet access assumed
Exclusion: illicit substances or ≥3 drinks/day; schizophrenia or bipolar diagnosis; severe depression or suicidal; taking regular benzodiazepines; if taking a medication, not on a stable dose of antidepressants for at least 1 month		 Self-guided iCBT: “Manage Your Mood Course”
8-week intervention: 5 online lessons, lesson summaries and homework, and automatic reminders; features age-appropriate case stories and examples of skill use.
Hobbs (2018) ≥65 years old;
48% Female;
Baseline mean PHQ-9=11.1; 57% had probable MDD		 Inclusion: patients prescribed iCBT for their depression in routine clinical care; computer/internet access assumed
Exclusion: actively suicidal; drug or alcohol dependency; schizophrenia or bipolar diagnosis; being treated with benzodiazepines or atypical antipsychotics. Adherence to these exclusion criteria was unknown.		 Self-guided iCBT: “ThisWayUp”, with optional therapist contact
12-week intervention: 6 online lessons and homework; features the story of a fictional character in her mid-30s being treated with CBT; therapists are encouraged to contact patients after the completion of first two lessons plus automated emails to therapists if symptoms become severe.
Therapist support: Clinicians who prescribed iCBT, including general practitioner, psychiatrist, clinical psychologist, or allied health professional; clinicians were encouraged to contact patients after the first two lessons were completed but details were not available.
McMurchie (2013) ≥65 years;
76% Female;
Baseline mean GDS-30 for the treated=21.1		 Inclusion: depression subscale of the Hospital Anxiety and Depression Scale≥8
Exclusion: current psychotic symptoms or suicidal; dementia; receiving psychological treatment; alcohol or drug dependence		 Self-guided iCBT: “Beating the Blues”
8-week intervention: 8 lessons with homework; technical assistance available upon request.
O’Moore (2017) 50–81 years old with Osteoarthritis;
80% Female;
72% tertiary education;
Baseline mean PHQ-9 for the treated=14.0;
100% had MDD		 Inclusion: major depressive disorder based on clinician-administered MINI; internet/computer access
Exclusion: suicidal; bipolar/psychotic/substance dependence; taking antipsychotics or benzodiazepines; if taking a medication, not on a stable dose of antidepressants for at least 2 months		 Therapist-guided iCBT: “The Sadness Program”
10-week intervention: 6 online lessons and homework; access to supplementary resources; features a story of a fictional character in her mid-30s being treated with CBT; therapist available by telephone or email
Therapist support: Clinical psychologist; therapist answered questions via email and initiated telephone contact when patients’ PHQ-9 scores deteriorated significantly.
Staples (2016) 60–88 years old;
59% Female;
75% tertiary education;
Baseline mean PHQ-9=12.7; over 90% scored ≥ 5.		 Inclusion: self-reported complaint of depression/anxiety; computer/internet access assumed
Exclusion: severe symptoms or suicidal deemed to require face-to-face assessment; prefer face-to-face services		 Therapist-guided, transdiagnostic iCBT: “Wellbeing Plus Course”
8-week intervention: 5 online lessons and homework, case-enhanced stories of experiences of older adults recovering from depression and anxiety, automatic emails. Trained clinicians were instructed to have weekly email or telephone contact with participants.
Therapist support: Registered psychologists; clinicians attempted to contact all participants each week vie telephone or email to provide guidance in completing the course, but actual level of contact varied widely.
Titov (2015) ≥60 years;
73% Female;
65% tertiary education;
Baseline mean PHQ-9 for the treated=11; 59% had MDE		 Inclusion: self-reported difficulties with depression; computer/internet access assumed
Exclusion: drug use or alcohol abuse; schizophrenia or bipolar diagnosis; severe depression symptoms or suicidal; if taking a medication, not on a stable dose of antidepressants for at least 1 month		 Therapist-guided iCBT: “Manage Your Mood Course”
8-week intervention: 5 online lessons, lesson summaries and homework, and automatic reminders; features age-appropriate case stories and examples of skill use; plus, weekly telephone or secure messaging with a therapist.
Therapist support: Clinical psychologists with doctoral qualifications with supervision; therapist provided weekly telephone or email contact.
Titov (2016) 60–93 years old;
64% Female; 80% tertiary education; Baseline mean PHQ-9=10.5; 91% scored ≥ 5.		 Inclusion: self-reported complaint of depression/anxiety; computer/internet access assumed
Exclusion: severe depression symptoms or suicidal		 Therapist-guided and self-guided iCBT: “Wellbeing Plus Course”, a transdiagnostic iCBT intervention for symptoms of anxiety and depression in older adults
8-week intervention: 5 online lessons and homework, case-enhanced stories of experiences of older adults recovering from depression and anxiety, automatic emails. The treatment arm with the highest level of therapist support is an interview plus clinician-guided iCBT program featuring brief telephone interview at the beginning and weekly contact from clinicians
Therapist support (if applicable): Clinical psychologists with doctoral-level qualifications with several years of experience in internet-delivered treatments; therapist conducted a brief telephone interview taking between 10–20 mins and weekly contact via telephone or email to answer questions; therapist spent an average of 68 minutes (SD=34.4) per patient.
Tomasino (2017) ≥65 years;
68% Female;
87% White;
98% tertiary education;
Baseline mean PHQ-9 for the treated=11		 Inclusion: PHQ-8≥8 or GDS-15>7; computer/internet access
Exclusion: no access to telephone/email/internet; psychotic disorder; cognitive impairment		 iCBT: “MoodTech”, with therapist guide and peer support
8-week intervention: 16 lessons and directions to practice skills; two-character storylines; and individual coaching. One treatment arm received iCBT with weekly individual coaching and the other treatment arm received iCBT with peer support and group moderation (optional individual coaching available upon request).
Therapist support: Clinical psychologist (Ph.D); therapists provided weekly brief (10–15 minutes) coaching during which they provided technical assistance and encouragement and answered questions.
Open in a new tab

Note. MDE=Major depressive episode; MDD=major depressive disorder; PHQ=Patient Health Questionnaire; GAD=Geriatric Depression Scale; MINI=Mini International Neuropsychiatric Interview Version 5.0

Meta-analysis

The mean within-group effect size was 1.27 (95% CI=1.09, 1.45) and the mean between-group effect size was 1.18 (95% CI=0.63, 1.73). Most studies reported findings regarding clinically significant improvement, although the criteria used for evaluating improvement varied. The proportion of participants who scored at PHQ-9≥10 at pretreatment and then scored below the cut-off at posttreatment ranged from 33% to 68.7% in the treatment group (Dear, Zou, Titov, et al., 2015; Dear et al., 2013; Hobbs et al., 2018; Titov, Dear, Ali, et al., 2015). In another study, 75% of the treated participants no longer met criteria for major depressive disorder according to a diagnostic interview at the follow-up (O’moore et al., 2018). In the study that used GDS, 39.4% of the participants had GDS scores falling at least two standard deviations in the direction of improvement below the pre-treatment mean (McMurchie et al., 2013). The average percentage reduction in symptoms as assessed by the PHQ-9 ranged from 54% to 58% in two studies (Staples et al., 2016; Titov et al., 2016). There was large heterogeneity among studies (for within-group effect sizes: Q(8) = 18.49, P < .05, I2 = 56.7%; for between-group effect sizes: Q(3) = 9.37, P < .05, I2 = 68%) (Table 3).

Table 3.

Effect sizes and clinically significant improvement statistics from included studies

Study name Within-group effect size Between-group effect size % clinically significant improvement in the treateda Criterion for clinically significant improvement
Dear (2013) 1.59 (0.88, 2.30) Not applicable 58% (n=10/17) Participants who scored at PHQ-9≥10 at pretreatment and scored below the clinical cut-off at posttreatment
Dear (2015) 1.06 (0.40, 1.72) Not applicable 33% (n=5/15) Participants who scored at PHQ-9≥10 at pretreatment and scored below the clinical cut-off at posttreatment
Hobbs (2018) 0.71 (0.36, 1.05) Not applicable 58.8%b Participants who scored at PHQ-9≥10 at pretreatment and scored below the clinical cut-off at posttreatment
McMurchie (2013) 1.11 (0.60, 1.63) 0.84 (0.26, 1.42) 39.4% (n=13/33) Participants whose GDS-30 scores fall at least two SDs in the direction of improvement below the pretreatment mean.
O’Moore (2017) 1.32 (0.86, 1.79) 1.02 (0.50, 1.54) 75% (n=33/44) Participants who no longer meet diagnostic criteria for major depressive disorder according to the MINI assessment at follow-up
Staples (2016) 1.44 (1.30, 1.58) Not applicable 58% Average percentage improvements (i.e., reduction in symptoms) on the PHQ-9 from pre- to post-treatment
Titov (2015) 1.63 (1.01, 2.25) 2.08 (1.40, 2.75) 68.7% (n=11/16) Participants who scored at PHQ-9≥10 at pretreatment and scored below the clinical cut-off at posttreatment
Titov (2016) 1.40 (1.25, 1.55) Not applicable 54%−60% Average percentage improvements (i.e., reduction in symptoms) on the PHQ-9 from pre- to post-treatment
Tomasino (2017) 1.16 (0.60, 1.72) 0.82 (0.07, 1.57) Not reported --
Mean effect size 1.27 (1.09, 1.45) 1.18 (0.63, 1.73) -- --
Heterogeneity statistics Q(8) = 18.49, p=.018; I2 = 56.7% Q(3) = 9.37, p=.025; I2 = 68% -- --
Open in a new tab

Note. PHQ-9=Patient Health Questionnaire. GDS=Geriatric Depression Scale. MINI=Mini International Neuropsychiatric Interview Version 5.0. 95% confidence intervals in parentheses.

a.

For studies that reported multiple clinically significant change indices (e.g., % reliable change, % reliable recovery, and % reliable improvement), only the proportion of reliable recovery (i.e., proportion scored below clinically cut-off) is presented in this table.

b.

Authors reported the overall recovery rate for a mixed-age sample and did not report age-specific recovery rates. Authors reported no significant differences in recovery rates by age groups. Therefore, the overall recovery rate is presented in this table.

In meta-regression, participants’ age significantly moderated within-group effect sizes such that a one-year increase in mean age was associated with a 0.06-point decrease in within-group effect sizes (β=−0.06, 95% CI=−0.11, −0.02, p=.016). Studies that explicitly excluded persons with severe depression tended to report larger effect sizes, but this moderating effect was not statistically significant (β=0.35, 95% CI=−0.03, 0.73, p=.064). Within-group effect sizes did not differ by the proportion of female, baseline PHQ-9 scores, treatment tailoring, guidance, or study quality. We did not conduct meta-regression on between-group effect sizes given the small number of studies eligible (k=4).

Publication bias was not evident based on Egger’s test and Begg’s test. However, visual inspection of the funnel plot revealed a pattern: Studies with a smaller sample size (thus larger standard error) tended to report a smaller effect size than studies with a large sample size.

Quality Appraisal

Figure 2 shows the risk of bias evaluation for each included study. Figure 3 shows the distribution of risk of bias of all the included studies. Six studies received a rating of low risk of bias for 1 or 2 domains. One study had a low risk of bias rating for all 5 domains, one had a low risk of bias rating for 4 domains, and another one had a low risk of bias for 3 domains. Risk of selection bias was low for all the RCT studies (k=3) and high for all the studies with non-RCT designs (k=6). Risk of detection bias was unclear for seven studies. All studies had low risk of reporting bias.

Figure 2.

Figure 2.

Open in a new tab

Estimated risk of bias summary for each domain

Figure 3.

Figure 3.

Open in a new tab

Distribution of estimated risk of bias across all included studies

DISCUSSION

This study aimed to systematically review and evaluate studies that have tested the effects of iCBT interventions on reducing depressive symptoms among older adults. Only nine studies met the inclusion criteria, which included a variety of study designs, and only three studies utilized RCT designs. The mean within-group effect size was 1.27 and the mean between-group effect size was 1.18. Within-group effect sizes decreased as participants’ age increased. Studies that excluded persons with severe depressive symptoms tended to report larger within-group effect sizes. There was large heterogeneity among studies, and study quality was low overall. Study samples were rather homogenous and social and cultural diversity were lacking. In addition, most studies required participants to have prior experience with computers and access to the internet, and participants with more severe depressive symptoms were excluded from most studies.

The mean effect sizes reported in our study were considered large, and the proportions of clinically significant improvement were substantial. These findings suggest that iCBT is feasible and potentially effective for reducing depressive symptoms among older adults. Although older adults may experience or report more challenges with technology, no evidence suggests that older adults do not benefit from iCBT. Emerging evidence suggests that iCBT may actually be more effective for older adults, who may be more likely to adhere to treatment than their younger counterparts (Hobbs et al., 2018). Additionally, in a sample of low-income older adults receiving home-delivered meals, improvement in depressive symptoms was greater among those receiving problem-solving therapy via the internet than those receiving face-to-face treatment for depression (Choi et al., 2014). While the mechanism behind this finding is unclear, possibly it was driven by increased self-efficacy obtained through learning and effectively using a new technology.

Our effect size estimates are in line with results from previous meta-analyses of iCBT trials for depression with mixed-age samples (Hedman et al., 2012; Păsărelu et al., 2017). Hedman et al. (2012) reported a mean within-group effect size of 0.94 from 20 RCTs on iCBT for depression. Păsărelu et al. (2017) reported a large mean within-group (Hedges’s g=1.08) and bewteen-group effect sizes (Hedges’s g=0.79) in reducing depressive symptoms based on 19 RCTs testing transdiagnostic iCBT for anxiety and depression. Both of these meta-analyses invovlved a mix of self-guided and therapist-guided iCBT interventons. In comparsion, a meta-analysis of 13 RCTs focused on self-guided iCBT for depression reported a small effect size (Hedges’s g=0.27) (Karyotaki et al., 2017). To what extent therapist guidance affects treatment effects of iCBT for depression is unclear. In our review, therapist guidance did not significantly moderate within-group effect sizes. A systematic review of RCTs comparing guided vs. unguided interventions reported a small but significant effect size (Cohen’s d=0.27) favoring guided interventions when mixed mental health symptoms were examined, but the effect size was reduced and no longer statistically significant for depression trials (Baumeister, Reichler, Munzinger, & Lin, 2014). More recent studies did not find significant differences between guided vs. non-guided interventions for reducing depressive symptoms (Dear et al., 2016; Titov, Dear, Staples, et al., 2015). Nevertheless, findings from a previous systematic review suggested that guidance improved completion and adherence (Baumeister et al., 2014). For older adults, the presence of even minimal guidance may serve to increase accountability and help resolve technological challenges faster, and thus improve their adherence using iCBT. This may be particularly important for older adults with limited technology competency. Most of the older adults included in this review had prior experiences with technology use and likely required less guidance. Future studies will add value by examining the impact of guidance on the effectiveness of iCBT for depression among older adults of varying levels of technology competency.

Participants’ age was negatively associated with within-group effect sizes, suggesting that effects of iCBT may diminish slightly with age. Mechanisms for the age difference in responses to iCBT are unknown. Possible explanations include technological challenges and limited age-appropriate components. Although CBT is an established and effective treatment for late-life depression, experts on late-life depression treatment have articulated the need for a developmentally appropriate format for CBT with older adults to enhance treatment outcomes (Laidlaw & Kishita, 2015). Nevertheless, the very limited evidence from previous studies found no age-related differences in treatment effects of iCBT (Hobbs et al., 2018; Mewton, Sachdev, & Andrews, 2013). These studies, however, focused on comparing older adults (generally defined as 65 years or older) and younger adults and did not examine the within-group heterogeneity in treatment responses among older adults. It is worth noting that participants included in this review largely belonged to the young-old group (65–74). Older adults from the middle-old (75–84) and the old-old (85 and above) groups have rarely been included in iCBT trials. Given the new evidence regarding age differences from our systematic review, future iCBT trials should involve older age groups and explore age as a moderator in treatment effects of iCBT and the underlying mechanisms to inform treatment development and clinician decision making.

Prior research demonstrates that iCBT interventions, with and without therapist support, can be effective for younger and working-aged adults with severe depressive symptoms (Bower et al., 2013; Meyer et al., 2015). Most studies included in this review, however, excluded persons with severe depressive symptoms, and several studies included persons with self-reported difficulty with depression but not clinical depression. Moreover, studies that excluded persons with severe depressive symptoms tended to report larger effect sizes. Our findings, therefore, can be generalized to treatment of mild to moderate depression at best; the effectiveness of iCBT for older adults with severe depressive symptoms is unclear and requires further study.

Only nine studies met the study inclusion criteria, and these had various study designs. The top reason for exclusion during the full-text review was failure to test iCBT on older adults. Many studies retrieved during the literature search explicitly excluded older adults aged 65 years and older from the trial, often without providing adequate explanation for this exclusion criterion. All included studies were published after 2010. A 2010 review of the literature examining the representation of older adults in iCBT trials found that older adults comprised only 3% of study participants in iCBT trials (Crabb et al., 2012). Our study findings echo those of the 2010 review and highlight the lack of iCBT trials that include or focus on evaluation of these interventions with older adults. The upside of our results is that since 2010, the number of iCBT trials aimed at evaluating the efficacy of these interventions for older adults has increased. Given that internet adoption among older adults has also increased steadily and that older adults are increasingly turning to the internet for health information (Anderson & Perrin, 2017), it is important that researchers continue to include older adults in iCBT trials and evaluate such interventions with this growing segment of the population.

Although there was large heterogeneity in outcomes among the studies included in this review, there was limited diversity in participant characteristics. Most participants were educated and had tertiary education or above. The majority of studies required participants to have at least basic computer and internet literacy and access, which may have limited inclusion of individuals with less education and lower socioeconomic status. The exclusion of older adults without internet access or prior experience with computer also limit the generalizations of the study findings to older adults with low computer literacy, who may require more guidance to complete treatment and benefit from iCBT interventions. There was also a significant lack of racial and ethnic diversity among participants in the included studies. While most studies did not report racial composition, given the racial composition of the populations where the studies took place (Australia, UK, Sweden), the majority of participants were likely white. Only the US study reported racial composition of the study sample, and an overwhelming majority of the participants in this study (87%) were white. Inclusion of older adults from socially and culturally diverse backgrounds has historically been low in mental health research (Fuentes & Aranda, 2012), and differences in mental health treatment preference and responses across socially and culturally diverse groups have been reported (Fuentes & Aranda, 2012; Windsor, Jemal, & Alessi, 2015). Thus, the degree to which our findings generalize to older adults with low education, with limited computer literacy, and from racial and ethnic minority groups is unclear. While we did not extract data regarding recruitment strategies used for each of the studies, it is possible that use of limited recruitment strategies or sources contributed to the lack of diversity in the included studies. Future trials of iCBT with older adults may benefit from using a wide range of recruitment strategies to avoid sampling bias and recruit diverse participants (Lindner, Nyström, Hassmén, Andersson, & Carlbring, 2015).

Limitations

This systematic review was limited to publications in English and the majority (k=7) were conducted in Australia. As such, these findings may be impacted by factors limited to Australian or predominately English-speaking cultures. Given the paucity of research on the effectiveness of iCBT for older adults with depression, we included pilot studies, open trials, and feasibility studies in addition to RCTs. The limitations associated with these designs prevented us from drawing a firm conclusion regarding the effectiveness of iCBT for older adults. Moreover, the number of studies included was relatively small, resulting in lack of power that would be needed to detect a particular study characteristic as the cause of heterogeneity. In addition, we did not examine the specific treatment elements that could impact the effectiveness of internet interventions for older adults with depression.

CONCLUSION

iCBT is a promising approach for treating depressive symptoms among older adults with mild to moderate depressive symptoms. Nevertheless, due to the small number of studies included, low study quality, large heterogeneity, lack of controls, and the inclusion of non-clinical samples in some studies, the results from this systematic review should be regarded as preliminary. Older adults are still significantly underrepresented in iCBT trials and there is limited high quality research examining the efficacy of iCBT interventions tailored for older adults. Future research directions include recruiting socially and culturally diverse groups, the old-old, older adults with limited computer literacy, and older adults with severe depressive symptoms, clarifying the impact of therapist guidance and tailoring, and adopting rigorous research designs with controls.

Acknowledgments

Funding Sources:

This work was supported by a grant from the National Institutes of Health, University of Michigan Older Americans Independence Center Research Education Core (Grant number: AG024824).

Appendix

Search Terms

PsycINFO

( DE “Cognitive Behavior Therapy” OR DE “Acceptance and Commitment Therapy” ) AND

AG ( Aged NOT (“thirties” or “young adulthood”) ) AND

control* AND

SU ( “Computer assisted therapy” or internet )

AND Depression

Filters:

Age group = aged

Peer reviewed

1990–2017

SU ( “therapy” or “psychotherapy” ) AND

AG (aged NOT (“thirties or “young adult*”) ) AND

depression AND

SU ( “computer assisted therapy” or “internet” ) AND

control*

Filters:

Age group = aged

Peer reviewed

1990–2017

( “older adults” AND depressi* ) AND ( “cognitive behavior therapy” OR “behavior* intervention” OR “behavior* activation” OR “cognitive therapy” or “cognitive intervention” ) AND ( “computer-assisted” OR internet OR telemedicine )

Filters:

Age group = aged

Peer reviewed

1990–2017

Web of Science

((“cognitive behavior therapy” or CBT OR “cognitive therapy”) and depression and “older adult*”)

Refined by: TOPIC: (internet OR “computer assisted” OR online)

(TS=((“cognitive behavior therapy” OR CBT OR “cognitive therapy” OR “behavioral intervention”) AND (depression) AND (techno* OR computer-assisted OR internet OR online))) AND LANGUAGE: (English) AND DOCUMENT TYPES: (Article)

Refined by: WEB OF SCIENCE CATEGORIES: ( GERIATRICS GERONTOLOGY OR GERONTOLOGY )

Timespan: 1990–2017. Indexes: SCI-EXPANDED, SSCI, A&HCI, CPCI-S, CPCI-SSH, BKCI-S, BKCI-SSH, ESCI, CCR-EXPANDED.

PubMed

(((“Aged”[Mesh]) OR “Aged, 80 and over”[Mesh]) AND “Depression”[Mesh]) AND “Therapy, Computer-Assisted”[Mesh] AND “Cognitive Behavior* Theory”

(((Aged) AND Depression) AND Cognitive Behavior Therapy) AND therapy, computer-assisted

((((Older adults) AND depression) AND cognitive behavioral theory) AND computer assisted therapy) AND control*

((((cognitive therapy) AND aged) AND depression)) AND internet based

(((((“cognitive therapy”) OR (“acceptance and commitment therapy”)) AND aged) AND depression) AND internet) OR “computer assisted

Filters:

Ages: 65+

Publish: 1990–2018

English

Footnotes

Conflict of Interest:

None.

References

  1. Andersson G, Cuijpers P, Carlbring P, Riper H, & Hedman E. (2014). Guided Internet-based vs. face-to-face cognitive behavior therapy for psychiatric and somatic disorders: A systematic review and meta-analysis. World Psychiatry, 13(3), 288–295. doi: 10.1002/wps.20151 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Anderson M & Perrin A. (2017). Tech Adoption Climbs Among Older Adults. 1. Technology use among seniors. Pew Research Center; Retrieved January 15, 2018, from http://www.pewinternet.org/2017/05/17/technology-use-among-seniors/. [Google Scholar]
  3. Baumeister H, Reichler L, Munzinger M, & Lin J. (2014). The impact of guidance on Internet-based mental health interventions—A systematic review. Internet Interventions, 1(4), 205–215. doi: 10.1016/j.invent.2014.08.003 [DOI] [Google Scholar]
  4. Beck AT, Steer RA, & Carbin MG (1988). Psychometric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8(1), 77–100. doi: 10.1016/0272-7358(88)90050-5 [DOI] [Google Scholar]
  5. Bower P, Kontopantelis E, Sutton A, Kendrick T, Richards DA, Gilbody S, ... & Meyer B. (2013). Influence of initial severity of depression on effectiveness of low intensity interventions: Meta-analysis of individual patient data. British Medical Journal, 346. doi: 10.1136/bmj.f540 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Choi NG, & Gonzalez JM (2005). Barriers and contributors to minority older adults’ access to mental health treatment: Perceptions of geriatric mental health clinicians. Journal of Gerontological Social Work, 44(3–4), 115–135. doi: 10.1300/J083v44n03_08 [DOI] [Google Scholar]
  7. Choi NG, Marti CN, Bruce ML, Hegel MT, Wilson NL, & Kunik ME (2014). Six‐month postintervention depression and disability outcomes of in‐home telehealth problem‐solving therapy for depressed, low‐income homebound older adults. Depression and Anxiety, 31(8), 653–661. doi: 10.1002/da.22242 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cohen J. (1988). Statistical Power Analysis for the Behavioral Sciences. Vol 2 Hillsdale, NJ: L. Lawrence Earlbaum Associates. [Google Scholar]
  9. Crabb RM, Cavanagh K, Proudfoot J, Learmonth D, Rafie S, & Weingardt KR (2012). Is computerized cognitive-behavioural therapy a treatment option for depression in late-life? A systematic review. The British Journal of Clinical Psychology, 51(4), 459–464. doi: 10.1111/j.2044-8260.2012.02038.x [DOI] [PubMed] [Google Scholar]
  10. Cuijpers P, van Straten A, & Smit F. (2006). Psychological treatment of late-life depression: A meta-analysis of randomized controlled trials. International Journal of Geriatric Psychiatry, 21(12), 1139–1149. doi: 10.1002/gps.1620 [DOI] [PubMed] [Google Scholar]
  11. Dear BF, Zou JB, Ali S, Lorian CN, Johnston L, Terides MD, … Titov N. (2015). Examining self-guided internet-delivered cognitive behavior therapy for older adults with symptoms of anxiety and depression: Two feasibility open trials. Internet Interventions, 2(1), 17–23. doi: 10.1016/j.invent.2014.11.002 [DOI] [Google Scholar]
  12. Dear BF, Zou J, Titov N, Lorian C, Johnston L, Spence J, ... & Knight, R. G. (2013). Internet-delivered cognitive behavioural therapy for depression: A feasibility open trial for older adults. Australian & New Zealand Journal of Psychiatry, 47(2), 169–176. doi: 10.1177/0004867412466154 [DOI] [PubMed] [Google Scholar]
  13. Dear BF, Zou JB, Titov N, Lorian CN, Johnston L, Spence J, … Fogliati VJ (2015). Clinical and cost-effectiveness of therapist-guided internet-delivered cognitive behavior therapy for older adults with symptoms of anxiety: A randomized controlled trial. Behavior Therapy, 46(2), 206–217. doi: 10.1016/j.beth.2014.09.007 [DOI] [PubMed] [Google Scholar]
  14. Dear BF, Staples LG, Terides MD, Fogliati VJ, Sheehan J, Johnston L, … Titov N. (2016). Transdiagnostic versus disorder-specific and clinician-guided versus self-guided internet-delivered treatment for Social Anxiety Disorder and comorbid disorders: A randomized controlled trial. Journal of Anxiety Disorders, 42, 30–44. doi: 10.1016/j.janxdis.2016.05.004 [DOI] [PubMed] [Google Scholar]
  15. Dobson KS. (2010). Handbook of cognitive-behavioral therapies. 3rd ed. New York, NY: Guilford Press. [Google Scholar]
  16. Eells TD, Barrett MS, Wright JH, & Thase M. (2014). Computer-assisted cognitive-behavior therapy for depression. Psychotherapy, 51(2), 191–197. doi: 10.1037/a0032406 [DOI] [PubMed] [Google Scholar]
  17. Egger M, Davey Smith G, Schneider M, Minder C, Mulrow C, Egger M, … Olkin I. (1997). Bias in meta-analysis detected by a simple, graphical test. British Medical Journal, 315(7109), 629–634. doi: 10.1136/bmj.315.7109.629 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Fuentes D & Aranda MP (2012). Depression interventions among racial and ethnic minority older adults: A systematic review across 20 years. The American Journal of Geriatric Psychiatry, 20(11), 915–931. doi: 10.1097/JGP.0b013e31825d091a [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Gould RL, Coulson MC, & Howard RJ (2012). Cognitive behavioral therapy for depression in older people: A meta‐analysis and meta‐regression of randomized controlled trials. Journal of the American Geriatrics Society, 60(10), 1817–1830. doi: 10.1111/j.1532-5415.2012.04166.x [DOI] [PubMed] [Google Scholar]
  20. Hamilton M. (1960). A rating scale for depression. Journal of Neurology, Neurosurgery, and Psychiatry, 23(1), 56–62. doi: 10.1136/jnnp.23.1.56 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hedman E, Ljótsson B, & Lindefors N. (2012). Cognitive behavior therapy via the Internet: A systematic review of applications, clinical efficacy and cost–effectiveness. Expert Review of Pharmacoeconomics & Outcomes Research, 12(6), 745–764. doi: 10.1586/erp.12.67 [DOI] [PubMed] [Google Scholar]
  22. Higgins JPT, & Thompson SG (2002). Quantifying heterogeneity in a meta-analysis. Statistics in Medicine, 21(11), 1539–1558. doi: 10.1002/sim.1186 [DOI] [PubMed] [Google Scholar]
  23. Higgins JPT, Altman DG, Gøtzsche PC, Jüni P, Moher D, Oxman AD, ... & Sterne JA (2011). The Cochrane Collaboration’s tool for assessing risk of bias in randomised trials. British Medical Journal, 343(7829), d5928. doi: 10.1136/bmj.d5928 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Hobbs MJ, Joubert AE, Mahoney AE, & Andrews G. (2018). Treating late-life depression: Comparing the effects of internet-delivered cognitive behavior therapy across the adult lifespan. Journal of Affective Disorders, 226, 58–65. doi: 10.1016/j.jad.2017.09.026 [DOI] [PubMed] [Google Scholar]
  25. Jayasekara R, Procter N, Harrison J, Skelton K, Hampel S, Draper R, & Deuter K. (2015). Cognitive behavioural therapy for older adults with depression: A review. Journal of Mental Health, 24(3), 168–171. doi: 10.3109/09638237.2014.971143 [DOI] [PubMed] [Google Scholar]
  26. Jenuwine ES, & Floyd JA (2004). Comparison of Medical Subject Headings and text-word searches in MEDLINE to retrieve studies on sleep in healthy individuals. Journal of the Medical Library Association, 92(3), 349–353. [PMC free article] [PubMed] [Google Scholar]
  27. Jonsson U, Bertilsson G, Allard P, Gyllensvärd H, Söderlund A, Tham A, & Andersson G. (2016). Psychological treatment of depression in people aged 65 years and over: A systematic review of efficacy, safety, and cost-effectiveness. PloS ONE, 11(8), e0160859. doi: 10.1371/journal.pone.0160859 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Karyotaki E, Riper H, Twisk J, Hoogendoorn A, Kleiboer A, Mira A, … Cuijpers P. (2017). Efficacy of self-guided internet-based cognitive behavioral therapy in the treatment of depressive symptoms a meta-analysis of individual participant data. JAMA Psychiatry, 74(4), 351–359. doi: 10.1001/jamapsychiatry.2017.0044 [DOI] [PubMed] [Google Scholar]
  29. Kroenke K, Spitzer RL, Williams JB, & Löwe B. (2010). The patient health questionnaire somatic, anxiety, and depressive symptom scales: A systematic review. General Hospital Psychiatry, 32(4), 345–359. doi: 10.1016/j.genhosppsych.2010.03.006 [DOI] [PubMed] [Google Scholar]
  30. Laidlaw K, & Kishita N. (2015). Age-appropriate augmented cognitive behavior therapy to enhance treatment outcome for late-life depression and anxiety disorders. GeroPsych: The Journal of Gerontopsychology and Geriatric Psychiatry, 28(2), 57–66. doi: 10.1024/1662-9647/a000128 [DOI] [Google Scholar]
  31. Lindner P, Nyström MBT, Hassmén P, Andersson G, & Carlbring P. (2015). Who seeks ICBT for depression and how do they get there? Effects of recruitment source on patient demographics and clinical characteristics. Internet Interventions, 2(2), 221–225. doi: 10.1016/j.invent.2015.04.002 [DOI] [Google Scholar]
  32. McMurchie W, Macleod F, Power K, Laidlaw K, & Prentice N. (2013). Computerised cognitive behavioural therapy for depression and anxiety with older people: A pilot study to examine patient acceptability and treatment outcome. International Journal of Geriatric Psychiatry, 28(11), 1147–1156. doi: 10.1002/gps.3935 [DOI] [PubMed] [Google Scholar]
  33. Mewton L, Sachdev PS, & Andrews G. (2013). A naturalistic study of the acceptability and effectiveness of internet-delivered cognitive behavioural therapy for psychiatric disorders in older Australians. PloS ONE, 8(8), e71825. doi: 10.1371/journal.pone.0071825 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Meyer B, Bierbrodt J, Schröder J, Berger T, Beevers CG, Weiss M, ... & Hautzinger M. (2015). Effects of an internet intervention (Deprexis) on severe depression symptoms: Randomized controlled trial. Internet Interventions, 2(1), 48–59. doi: 10.1016/j.invent.2014.12.003 [DOI] [Google Scholar]
  35. Moher D, Liberati A, Tetzlaff J, Altman DG, Altman D, Antes G, … Tugwell P. (2009). Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. PLoS Medicine 6(7), e1000097. doi: 10.1371/journal.pmed.1000097 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. O’Donohue WT, & Fisher JE (Eds.). (2009). General principles and empirically supported techniques of cognitive behavior therapy. Hoboken, NJ: John Wiley & Sons Inc. [Google Scholar]
  37. O’Moore KA, Newby JM, Andrews G, Hunter DJ, Bennell K, Smith J, & Williams AD (2018). Internet cognitive–behavioral therapy for depression in older adults with knee osteoarthritis: A randomized controlled trial. Arthritis Care & Research, 70(1), 61–70. doi: 10.1002/acr.23257 [DOI] [PubMed] [Google Scholar]
  38. Orgeta V, Brede J, & Livingston G. (2017). Behavioural activation for depression in older people: Systematic review and meta-analysis. The British Journal of Psychiatry, 211(5), 274–279. doi: 10.1192/bjp.bp.117.205021 [DOI] [PubMed] [Google Scholar]
  39. Păsărelu CR, Andersson G, Bergman Nordgren L, & Dobrean A. (2017). Internet-delivered transdiagnostic and tailored cognitive behavioral therapy for anxiety and depression: A systematic review and meta-analysis of randomized controlled trials. Cognitive Behaviour Therapy, 46(1), 1–28. doi: 10.1080/16506073.2016.1231219 [DOI] [PubMed] [Google Scholar]
  40. Renton T, Tang H, Ennis N, Cusimano MD, Bhalerao S, Schweizer TA, & Topolovec-Vranic J. (2014). Web-based intervention programs for depression: A scoping review and evaluation. Journal of Medical Internet Research, 16(9), e209. doi: 10.2196/jmir.3147 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Staples LG, Fogliati VJ, Dear BF, Nielssen O, & Titov N. (2016). Internet-delivered treatment for older adults with anxiety and depression: Implementation of the Wellbeing Plus Course in routine clinical care and comparison with research trial outcomes. BJPsych Open, 2(5), 307–313. doi: 10.1192/bjpo.bp.116.003400 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Titov Dear BF, Staples LG, Terides MD, Karin E, Sheehan J, … McEvoy PM (2015). Disorder-specific versus transdiagnostic and clinician-guided versus self-guided treatment for major depressive disorder and comorbid anxiety disorders: A randomized controlled trial. Journal of Anxiety Disorders, 35, 88–102. doi: 10.1016/j.janxdis.2015.08.002 [DOI] [PubMed] [Google Scholar]
  43. Titov N, Dear BF, Ali S, Zou JB, Lorian CN, Johnston L, … Fogliati VJ (2015). Clinical and cost-effectiveness of therapist-guided internet-delivered cognitive behavior therapy for older adults with symptoms of depression: A randomized controlled trial. Behavior Therapy, 46(2), 193–205. doi: 10.1016/j.beth.2014.09.008 [DOI] [PubMed] [Google Scholar]
  44. Titov N, Fogliati VJ, Staples LG, Gandy M, Johnston L, Wootton B, … Dear BF (2016). Treating anxiety and depression in older adults: Randomised controlled trial comparing guided v. self-guided internet-delivered cognitive-behavioural therapy. BJPsych Open, 2(1), 50–58. doi: 10.1192/bjpo.bp.115.002139 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Tomasino KN, Lattie EG, Ho J, Palac HL, Kaiser SM, & Mohr DC (2017). Harnessing peer support in an online intervention for older adults with depression. The American Journal of Geriatric Psychiatry, 25(10), 1109–1119. doi: 10.1016/j.jagp.2017.04.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Weinberger MI, Mateo C, & Sirey JA (2009). Perceived barriers to mental health care and goal setting among depressed, community-dwelling older adults. Patient Preference and Adherence, 3, 145–149. doi: 10.2147/PPA.S5722 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Windsor LC, Jemal A, & Alessi EJ (2015). Cognitive behavioral therapy: A meta-analysis of race and substance use outcomes. Cultural Diversity and Ethnic Minority Psychology, 21(2), 300–313. doi: 10.1037/a0037929 [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Yesavage JA, Brink TL, Rose TL, Lum O, Huang V, Adey M, & Leirer VO (1982). Development and validation of a geriatric depression screening scale: A preliminary report. Journal of Psychiatric Research, 17(1), 37–49. doi: 10.1016/0022-3956(82)90033-4 [DOI] [PubMed] [Google Scholar]

RESOURCES