Systematic Review: Computerized Cognitive Behavioral Therapy for Adults with Depressive or Anxiety Disorders
Given the high rates of mental illness among Veterans returning from Iraq and Afghanistan, it is not surprising that the demand for VA mental health services has increased 132% since 2006. Cognitive behavioral therapy (CBT) is effective in treating mild to severe mental health symptoms. Computer-based programs grounded in CBT (computerized CBT [cCBT]) have generally been shown to produce significant reductions in depressive and anxiety symptoms, but treatment effects vary across studies. The availability of support via e-mail, instant messaging, or phone contact with a therapist may mitigate attrition and improve treatment outcomes. However, the extent to which support-related factors influence treatment response to cCBT programs is unclear.
To support the development of cCBT programs, VA commissioned HSR&D's Evidence-based Synthesis Program (ESP) to conduct a systematic review of the literature. Investigators with the ESP in Durham, NC reviewed the literature from January 1990 through August 2013 and found 47 relevant randomized controlled trials (RCTs); the majority of trials were conducted outside the United States, and only one involved U.S. military personnel or Veterans.
Results of this evidence review are summarized here and more detailed responses to three specific Key Questions follow.
As detailed below, investigators found moderate to strong evidence that cCBT – when compared with control conditions – is effective in improving end of treatment symptoms for mid-life patients (median age 39.8, range 20 to 58) with mild to moderate major depressive disorder, generalized anxiety disorder, or panic disorder. Treatment effects were smaller in studies of patients who were enrolled based on a positive depression symptom screen compared with studies of patients enrolled using a formal diagnostic assessment of depression. For PTSD and participants enrolled based on a positive anxiety symptoms screen, the evidence was insufficient to draw a conclusion. Investigators also found evidence suggesting that a greater level of therapist support was related to greater benefit, but additional head-to-head trials examining varying levels of therapist support are needed to address this issue definitively. It is suggested that VA/DoD consider this body of evidence when updating their clinical guidelines for depression and anxiety disorders.
Key Questions and Findings
For adults with depressive disorder, post-traumatic stress disorder (PTSD), panic disorder, or generalized anxiety disorder, what are the effects of computerized CBT (cCBT) interventions compared with inactive controls?
Control conditions included usual care, waitlist, or attention/information controls. Studies utilizing usual care control groups generally included little information on the treatment received, so information was insufficient to definitively characterize usual care conditions in this review as active controls. Overall, investigators found at least moderate strength of evidence that cCBT interventions improved symptoms to a greater degree than did control conditions for depressive symptoms, major depressive disorder, generalized anxiety disorder, and panic disorder. Significant treatment outcome differences by control condition type (usual care, waitlist, or attention/information control) were not evident for any disorder types.
For major depressive disorder, generalized anxiety disorder, and panic disorder, the effects measured at the end of treatment were large. However, for PTSD and anxiety symptoms, there was evidence of possible benefit, but too few trials to estimate a reliable summary treatment effect. Patterns were similar for effects on health-related quality of life on major depressive disorder, generalized anxiety disorder, and panic disorder, though effects were generally smaller. Trials conducted in other disorders generally did not report quality of life data. For the subset of trials in this review that evaluated outcomes at six months or longer, treatment effects were smaller but remained statistically significant for depressive disorders and depressive symptoms. Trials conducted in anxiety disorders generally did not report long-term follow-up data.
Findings also showed the proportion completing all cCBT content was reported in approximately two-thirds of studies and varied substantially across studies (median proportion completing all cCBT sessions was 49.5%, range 11% to 100%). Completion rates were lower for patients with depressive symptoms than for other conditions. Data on cCBT safety and adverse events were rarely reported. Only five studies reported cost data. Limited data from these studies suggested the financial costs of cCBT are less than face-to-face therapy. However, costs of providing cCBT varied substantially depending on how much support time was provided by therapists.
Also of note, the use of cCBT technology brings with it privacy and information security risks that must be addressed to ensure that these risks are eliminated – or at least communicated to Veterans using cCBT. For treatments that use electronic messaging from hospital staff to remind patients to complete modules or to address questions, secure messaging systems will need to be integrated with the treatment. Because cCBT often utilizes web-based modules, the security of information transmitted and stored on these sites also will need to be addressed.
For cCBT interventions, what level, type, and modality of user support is provided (e.g., daily telephone calls, weekly email correspondence); who provides this support (e.g., therapist, graduate student, peer); what is the clinical context (primary intervention, adjunct); and how is this support related to patient outcomes?
Most of the cCBT interventions were accessed via the Internet (79%) from non-clinical locations and were supported by a therapist. Approximately one-third included peer support in some form. The level of therapist support varied widely, ranging from minimal feedback on homework assignments via e-mail to a full therapy session via instant messaging or a chat room format. In two of the 57 intervention arms, cCBT was used as an adjunct to face-to-face therapy, but for the remaining 55 intervention arms examined, cCBT was a stand-alone treatment.
Findings also showed:
- An advertisement on the Internet was the most common means of patient recruitment (53%).
- Most trials used e-mail in some form (74%), while phone support by clinical staff (35%) and peer support via discussion board (25%) were used less often. Instant messaging was used in one study.
- The intervention components of studies classified as supported or supported with live features were highly variable, making firm conclusions difficult to draw. However, exploratory subgroup analysis using indirect comparisons showed a general association between higher levels of support and greater treatment effects.
- Analyses suggested greater support was related to better treatment outcomes in depressive symptoms, generalized anxiety disorder, and panic disorder, while treatment outcomes were good across levels of support in major depressive disorder. There were too few studies in PTSD or elevated anxiety symptoms to analyze the effects of level of support.
For adults with depressive disorder, PTSD, panic disorder, or generalized anxiety disorder, what are the effects of cCBT interventions compared with face-to-face therapy?
Seven studies directly compared cCBT with face-to-face therapy. Only one study on depressive symptoms was a non-inferiority trial designed specifically to test the hypothesis that cCBT would not perform significantly worse than face-to-face therapy. Panic disorder was the only condition with more than two studies for this comparison, and these trials showed no difference in effects on symptom severity or health-related quality of life. Two studies found no difference in treatment effects for participants with depressive symptoms, and the sample size in the single pilot study on major depressive disorder was too small to determine the strength of evidence. Findings suggest that the current literature are generally insufficient for making a determination about whether the efficacy of cCBT is comparable with traditional, face-to-face therapy.
There is a need for more RCTs investigating PTSD, and for RCTs comparing different levels and types of therapist support for cCBT. In addition, future research could examine safety and efficacy of cCBT for Veterans specifically, as well as those with more severe symptoms, older age, and less experience with computers. Information is insufficient on the effects of cCBT adherence rates or the durability of treatment effects beyond the end of therapy due to a lack of clinical trials with 6- to 12-month outcome assessments. VA and other healthcare systems should consider their IT, clinical, and policy needs when deciding whether to invest in research to address gaps in the evidence.
A Cyberseminar session on this ESP Report will be held TBD. To register, go to the HSR&D Cyberseminar web page.
This report is a product of VA/HSR&D's Quality Enhancement Research Initiative's (QUERI) Evidence-Based Synthesis Program (ESP), which was established to provide timely and accurate synthesis of targeted healthcare topics of particular importance to VA managers and policymakers — and to disseminate these reports throughout VA.
Dedert E, McDuffie JR, Swinkels C, Shaw R, Fulton J, Allen KD, Datta S, Williams JW. Computerized Cognitive Behavioral Therapy for Adults With Depressive or Anxiety Disorders. VA-ESP Project #09-010; 2013.
View the full report (**VA Intranet only**):
Please feel free to forward this information to others!
Read past HSR&D Management e-Briefs on the HSR&D website.
This Management eBrief is a product of the HSR&D Evidence Synthesis Program (ESP). ESP is currently soliciting review topics from the broader VA community. Nominations will be accepted electronically using the online Topic Submission Form. If your topic is selected for a synthesis, you will be contacted by an ESP Center to refine the questions and determine a timeline for the report.
This Management e-Brief is provided to inform you about recent HSR&D findings that may be of
interest. The views expressed in this article are those of the authors and do not necessarily
reflect the position or policy of the Department of Veterans Affairs. If you have any questions or
comments about this Brief, please email CIDER.
The Center for Information Dissemination and Education Resources (CIDER) is a VA HSR&D Resource
Center charged with disseminating important HSR&D findings and information to policy makers, managers,
clinicians, and researchers working to improve the health and care of Veterans.