Issues in the dissemination of cognitive–behavior therapy

Abstract

In the past 40 years, cognitive–behavior therapy (CBT) has emerged as the initial treatment of choice for patients with mild to moderate depression, anxiety disorders and other problems. In this paper, we discuss issues related to the dissemination and implementation of CBT in various practice settings as well as the use of manuals, computers, the telephone, and the Internet to aid dissemination and implementation. We review key aspects of CBT dissemination, such as the reach of CBT, models of dissemination, and obstacles and barriers to dissemination including patient interest, therapist training and research priorities. The effectiveness of manualized programs is considered, as well as the increasing sophistication of computer-assisted therapy. Stepped-care approaches are discussed as a viable solution to some of these barriers. We provide two examples of successful CBT dissemination, the Staying Free program, a smoking cessation program for inpatients, and the Improving Access to Psychological Therapies program in Britain, which aims to improve access to psychological therapy. We argue that two critical factors will determine the success of implementation of CBT in this century: 1) mandated outcomes and 2) leadership.

• Anxiety
• CBT dissemination
• CBT implementation
• Depression
• Stepped-care

In the past 40 years, cognitive–behavior therapy (CBT) has emerged as the initial treatment of choice for patients with mild to moderate depression, anxiety disorders and other problems. Even when other therapies have proved effective, CBT has the advantage of using well-defined treatment practices that can be easily taught to a variety of therapists and whose implementation can be monitored. In spite of its efficacy and potential advantages over other treatment practices, CBT appears to be underutilized in practice settings. In this paper, we discuss issues related to the dissemination and implementation of CBT in various practice settings as well as the use of manuals, computers, the telephone and the Internet to aid dissemination and implementation. We provide several examples of successful CBT dissemination. We argue that two critical factors will determine the success of implementation of CBT in this century: 1) mandated outcomes and 2) leadership.

What is the reach of CBT?

The “reach” of an intervention can be considered in terms of the number of individuals within a defined population who would benefit from and receive it. There is no doubt that depression and anxiety are undertreated and that a number of individuals suffer needlessly. For instance, a recent World Health Organization (WHO) study found that depression had a worse impact on a person's functioning than angina, asthma, diabetes or arthritis [1]. The report estimated that 40% of disability worldwide is due to depression. Undertreatment can be attributed to a number of factors, including the fact many depressed and anxiety patients are never identified. If identified and interested, these individuals may face various barriers to treatment including inability to pay, reluctance to seek help because of stigma or difficulty finding culturally sensitive care, to name only a few problems. In the UK, for instance, only a third of people with diagnosable depression and less than quarter of those with anxiety disorders receive treatment of any sort [2]. In the USA, only about 40% of patients with mood or anxiety disorders receive any treatment [3]. Other studies suggest that evidence-based psychotherapies are not practiced widely in community settings [3], [4].

Little data exists specifically on the use of CBT for treating depression and anxiety. In the USA, the most common treatment is pharmacotherapy, which is mostly provided by primary care doctors and psychiatrists who are not usually trained to provide CBT. The percentage of patients in need of CBT who have been provided with some basic cognitive and behavioral skills is not known. A survey of 2281 practicing psychologists in the USA (http://www.psychotherapynetworker.com, last accessed May 8, 2008) found that almost all reported using a variety of approaches—only 4% said they used one model exclusively. Sixty-nine percent of respondents said that they used CBT at least part-time or in combination with other therapies. The high rate of self-reported CBT use, and the few studies of its use in practice, raise the question as to what constitutes an “adequate” dose or course of CBT, an important issue for dissemination.

The adequacy of course can be defined in a number of ways, such as focusing on: 1) the therapist's practice (is the therapist demonstrating competence, for instance, as measured by comparing tape recorded sessions against some criteria), 2) the recommended number of sessions being provided for a condition (e.g. 10–16 for mild–moderate depression, 6–8 for panic disorder), 3) attainment of specific outcomes (e.g. 50% reduction in depression), 4) learning and practice of prescribed techniques [e.g. use of cognitive techniques as prescribed [5]], or even 5) completion of treatment program materials. To our knowledge, no data is available on these issues for community-based practice, although this may soon change with the implementation of the UK Improving Access to Psychological Therapies (IAP) program discussed below 2.

Models of dissemination

Improved reach is the endproduct of successful dissemination. Given the potential benefit of CBT and its apparent underutilization, there has been considerable discussion as to how to improve its reach. Dissemination of evidence-based techniques is a major focus across all aspects of medicine, psychology and public health. As we move forward with the dissemination and implementation of CBT, consideration of such models could guide dissemination practice. Rogers [6], whose model has been influential, describes five factors that affect adoption: perceived advantage, consistency with current practices where adoption is to occur, complexity of innovation (the simpler the better), time-frame for implementation (gradual is better) and visibility once adopted (the more the better). The model serves as a heuristic framework but has not been tested against other models, nor used, for CBT implementation. For instance, this model would imply that an organization interested in disseminating CBT would examine the perceived advantage to the therapist of using CBT over her other current practices. Will the use of CBT generate more business or make the work more enjoyable? Does the therapist believe that the use of CBT will achieve better and faster outcomes, and does this constitute an advantage to the therapist?

Social marketing, the application of marketing techniques for dissemination of socially valuable activities and products, is another widely used model [7]. Social marketing considers the market need, market segments, strategies to reach those markets and evaluation of effects. In a social marketing model, those interested in promoting the use of CBT might advertise directly to a particular market segment (e.g. those with anxiety disorders) similar to the pharmaceutical industry. Other models include community organization, blends of dissemination practices (FCP) and the RE-AIM model, which addresses reach, efficacy/effectiveness, adoption, implementation and maintenance of interventions (http://www.re-aim.org, last accessed May 10, 2008). Detailed guidelines for implementing interventions have been written based on these models. For instance, Smith & Taylor describe in detail all the steps involved in implementing a CBT-based inpatient smoking cessation program (Table 1) [8]. The model is based on techniques borrowed from Rogers, social marketing and the RE-AIM model.

Table 1.  A list of factors involved in implementation.

Foundation

Evidence base for intervention

Clinical guidelines, quality assurance and insurance

Program administration and management

Program parameters

Target population and reach

Champions advisory committees, stakeholders, community resources

Financing and budgeting

Program development

Identifying providers

Identifying patients

Training providers

Program evaluation

Data and measurement

Program promotion

Ramp up and delivery

Policy, procedures, protocols and systems

Quality assurance (monitoring outcomes, improving program, etc.)

In spite of innumerable studies of dissemination of evidence-based techniques across many disorders, a recent review reports, “There is an imperfect evidence base to support decisions about which guideline dissemination and implementation strategies are likely to be efficient under different circumstances” [9].

What are obstacles/barriers to dissemination?

Many researchers and clinicians have discussed barriers to disseminating evidence-based interventions relevant to the dissemination of CBT (cf., [10–12]). The interest or appropriateness of CBT to patient populations is one factor. In the highly influential STAR-D trial in the USA, patients who showed an inadequate response to the first pharmacology treatment were offered CBT and/or pharmacotherapy as a next step [13]. However, less than 30% of patients were willing to use CBT alone or in combination with medication. These results are biased by a sample selected to participate in a pharmacology study, but having failed the first step, one might expect a more enthusiastic response to a non-pharmacological approach. Is lack of interest in CBT another barrier to dissemination?

In addition, researchers themselves may be impeding dissemination and implementation. There is the assumption that empirical support is sufficient to motivate adoption and that effective innovations will naturally “diffuse” throughout the system with little additional effort on the part of the researcher. Research studies typically involve very selected populations and the results can be perceived by the end users (the therapists) as too limited and not relevant to their clinical population [14]. A research culture trained to focus on “getting it right” versus “getting it out” may also impede dissemination. Andy Grove, past president of Intel, and considered one of the leading business innovators of the last century, notes that “better is the enemy of good enough.”

The National Institute on Drug Abuse (NIDA) has made both dissemination and sustained adoption of evidence-based practices a goal for over a decade, funding dissemination research and designing a clinical trials network to test the effectiveness and sustainability of treatments for substance abuse in real-world settings (www.nida.nih.gov/ctn/, last accessed May 25, 2008). In addition, the National Advisory Mental Health Council's Clinical Treatment and Services Research Workgroup report to the National Institute of Mental Health (NIMH) recommended that NIMH begin to support more research to facilitate a better understanding of how best to disseminate effective psychological treatments (www.nimh.nih.gov/about/organization/dsir/services-research-and-epidemiology-branch/dissemination-and-implementation-research-program.html, last accessed June 2, 2008). NIMH has implemented a funding program for the adoption of evidence-based practice in state mental health services in an effort to bridge mental health services research with real-world practice. Statewide and national initiatives to implement evidence-based practices are becoming more common and are providing treatment researchers with funding and opportunities to disseminate treatments with demonstrated efficacy into more applied settings. Grant proposals require a statement as to how the findings will be translated. The impact of these efforts is not known but demonstrate movement in the right direction.

Effectiveness of self-help programs

As a method to ensure that patients are provided an adequate “dose” of CBT, many have advocated the use of manuals (including self-help books). In general, the use of self-help programs for depression has been associated with a low to moderate effect size (standardized mean difference = − 1.36; 95% confidence interval [CI] = − 1.76 to −0.96; [15]). Self-help interventions appear to be somewhat more effective for anxiety disorders. One recent review found that the average effect sizes (Cohen's d) comparing self-help interventions to control groups for target symptoms were 0.62 at post-treatment and 0.51 at follow-up [16]. In this same analysis, the effect sizes for self-help materials were much lower for self-help than therapist-delivered interventions. The use of self-help materials only achieves similar results to guided self-help in improvement in eating disorder symptoms [17].

Computer therapy or computer-assisted therapy

Many CBT techniques lend themselves to manualized treatment and in turn, manualized treatments can easily be provided by computers, through CD-ROMs or on the Internet. Such programs are capable of achieving a wide reach and have been advocated as a way to ensure that patients receive at least a minimal dose of treatment. For instance, MOODGYM is a free on-line program available to any user, anywhere (www.moodgym.anu.edu.au last accessed May 8, 2008). The user assesses their mood and is provided a series of four sessions covering the basic CBT practices. It has been used by over a 100,000 individuals worldwide. Use of the program is associated with significant reductions in depressive symptoms [18], [19]. People who used the public version of the program improved as much as participants enrolled on the website for a randomized controlled trial [19].

A computerized version of a manualized program, whether delivered on a hand-held computer or on the Internet, should be similar to, if not better than, the same program delivered in a manual or a book. Some studies suggest that this is the case [20]. The number of computerized programs has exploded in recent years and a number of programs are available to treat a variety of problems. Marks and colleagues have recently evaluated the effects of a computer-based CBT (CCBT) program across a large number of problems [21]. They identified 103 randomized controlled studies of computer-aided systems. Together, they found that patients often improved more with CCBT than with contrasting approaches (typically waiting-list controls or usual care) with an average reduction of 50% in usual therapist time. In fact, the effectiveness of some of the more sophisticated systems have raised issues as to the need for any or more than minimal therapist contact for certain levels of symptoms or problems. On the other hand, another meta-analysis found low to moderate effects for Internet-based programs, for depression with moderate effect sizes for anxiety, partly accounted for more therapist contact [22]. The conflicting research indicates that more information is needed to determine the efficacy of these programs for different conditions and levels of severity.

Excitingly, such programs have potential to reach huge numbers of people. Few Internet-based CBT programs take full advantage of the computers or the Internet. The next wave of Internet-based programs are likely to involve much more interactivity including improved graphic design, personalization, connectivity, multiple communication modalities (I-messaging, hand-held devices, games, avatars, virtual realities and social networks). Although researchers have begun to combine existing treatments [e.g. Internet self-help with therapist feedback and in vivo group exposure [23] or Internet CBT with telephone calls [24]], the use of multimedia combined with other modalities such as face-to-face, care management, electronic medical record databases, and telemedicine as well as models for using multimedia, multimodal interventions are not being developed and examined. For instance, we are involved in a project designed to determine the effect of a program that combines on-line patient CBT/psychoeducation, support groups and medication management by psychiatrists and care management using algorithms, all integrated with and coordinated through electronic medical records for treating depression and bipolar disorder in a large HMO. The results of this study should be available in 2009.

Therapist use of manualized therapies

Some of the obstacles to CBT dissemination include the negative or suspicious attitudes of administrators and clinicians towards the new treatment and the labor-intensive training that may be required [25]. Clinicians complain that following structured interventions is “boring” and fails to acknowledge the complexity of providing treatments to people with many issues, demands and needs. One study found that rigidly following structured interventions had a worse outcome than employing some flexibility [26], indicating that adequate training and familiarity is an essential component of CBT dissemination. While most therapists report using some self-help materials, forms and programs with patients, Addis [10] found that therapists were concerned that the manual would affect the therapeutic relationship, not adequately address patient needs, reflect poorly on their competence (why does this therapist need to use a manual), make their work less satisfying and restrict their clinical innovation. Many of these same issues apply to evidence-based therapies, in general.

There are few studies that compare manualized to non-manualized treatment, where both are provided by the same therapist or in the same setting to similar populations. This remains an important issue for dissemination. If therapists can provide an adequate dose of CBT with a manual, it might require less skill and training. In theory, the therapists could use the manual to provide the core elements and focus on other issues using their expertise. One study on the treatment of panic disorder found that a six-session CBT program delivered by a therapist with augmentation of a hand-held computer was as effective as a 12-session therapist-delivered CBT and more effective than a six-session therapist-delivered CBT [27]. This suggests that a computer program, essentially providing an interactive manual, is effective and efficient at least for panic disorder.

However, some CBT training programs (e.g. at the Beck Institute) focus on having the therapists become very proficient at adapting CBT to the patient without relying on manuals (although forms are used). Such an approach would seem to be more appealing to therapists but runs the risk of non-adherence to the intervention. In the absence of data on this critical issue, it would seem that the therapist should use a manualized treatment, or part of it (but which part), to ensure that the patient receives a reasonable dose of the manualized therapy. In fact, a better approach would be to focus on outcome, as we will discuss later.

Therapist training

CBT research has a long history of developing training programs for therapists involved in clinical trials and monitoring treatment fidelity. Therapist training for clinical efficacy trials generally consists of selection of therapists who are committed to the type of treatment they will implement in the trial, an intensive didactic seminar or workshop that includes review of the treatment manual with extensive role-playing and practice, and successful completion of at least one closely supervised training case. Many trials also include monitoring of treatment fidelity, for instance, through review of video or audiotaped sessions. In general, these strategies appear to be successful, in that therapist adherence or skill generally improves during training (or at least achieves acceptable levels; [28]).

However, surprisingly little work has been conducted on how best to train therapists in CBT in real-world settings [29]. Miller & Mount reported that after a 2-day motivational interviewing (MI) training seminar, counselors reported large changes in practice, whereas observational ratings suggested more modest changes in practice behavior [30]. Morgenstern and colleagues trained 29 counselors drawn from community drug abuse treatment clinics to deliver CBT [31]. However, no significant differences were found in substance use outcomes when 252 substance abusers were randomly assigned to treatment as usual, high-standardization CBT or low-standardization CBT as delivered by those clinicians [32].

In an interesting study designed to determine the effectiveness of different training practices with 78 clinicians, most of whom treated patients with substance abuse, Sholomskas et al. [29] found that 1) a training seminar plus supervised casework was superior to 2) a manual-only condition in teaching CBT skills and in treatment adherence, whereas the effectiveness of 3) the manual and web-based training was found to be between the 2) manual-only and 1) the manual and the seminar plus supervision conditions. The conventional wisdom is that workshops alone are not sufficient to help even experienced therapists achieve competence and that follow-up supervision training is necessary. Although their effectiveness is unproven, CBT training programs, offered in a variety of formats, lengths and prices are widely available. Training in CBT is offered in most clinical training programs; experience with doing CBT is now a core competency of psychiatric training programs. Access to training in CBT does not seem to be a rate-limiting step in its dissemination.

Stepped-care approaches

A number of individuals have considered the advantages of combining self-help programs with more intensive treatment using stepped-care models. The basic notion is that many individuals will improve over time or on their own with minimal assistance and that easy access to such programs may both reduce the need to provide more expensive face-to-face programs while increasing reach of effective programs. Stepped-care has become an essential feature of most treatment guidelines in psychiatry. Research demonstrates the cost effectiveness of stepped-care depression treatment among individuals with diabetes [33], and smoking cessation for outpatients with depression [34], as well as the efficacy of stepped-care treatment for obesity [35]. However, research on stepped-care treatment of mental health disorders such as depression appears to be in the early phases, with most articles focused on theory rather than empirical data [36], [37].

In our own work, we combine populated-based universal, targeted and indicated models (e.g. [38]). Such interventions can also be combined with stepped-care techniques, as seen in Fig. 1. Such problems are particularly applicable to captive patient populations, such as HMOs, or public programs focused on providing effective interventions to large populations at relatively little cost.

Fig. 1.  Stepped-care approach for universal, indicated and targeted interventions for eating disorders. Key: +, no longer meets high-risk criteria (if high risk) or subclinical/clinical criteria; as appropriate; nc, no change; −, risk or clinical status is worse; CBT, cognitive–behavior therapy. *Estimated from [45]. **Adapted from NICE guideline for bulimia nervosa ([46], p. 16). ***Estimated from [47], [48]. ****Estimated from [49].

Two cases of dissemination

Some CBT based programs have been carefully disseminated with attention paid to many of the issues discussed above. In this section, we present two examples, the Staying Free program, a smoking cessation program for inpatients, and the IAP program in Britain.

Staying Free

Staying Free is an evidence-based program designed to help inpatients smoking before admission to stop smoking by the time of discharge. The program was designed following social learning theory and CBT. The program was shown to be efficacious in a number of controlled trials [39], [40] and has been shown to be equally effective when disseminated in a variety of real world settings [41]. The intervention is cost-effective [42] and has been recommended as a treatment in evidence-based reviews [43].

The intervention combines principles of social cognitive learning theory with nicotine addiction and relapse prevention models. After receiving a standardized message from their physicians, patients meet with a tobacco cessation counselor for about 1 hour during their hospitalization. They are shown a 16-minute videotape, given a CBT-oriented workbook with an accompanying audiotape, and counseled on how to cope with any high-risk-to-relapse situations. Patients who report significant withdrawal symptoms or high rates of tobacco dependence are offered pharmacotherapy before discharge. Patients receive 10-min nurse-initiated standardized phone contacts, optimally at 48 h, 7 days, 21 days and 90 days after hospital discharge (for program details, see [44]).

Patient manuals and materials were developed by the American Heart Association. A 2-day standardized training program was also developed. In addition, a method of determining fidelity of treatment was designed: recordings of actual therapists’ sessions are reviewed against the defined standards. With extensive data on effectiveness, cost and feasibility, it was possible to influence the Joint Commission of Accreditation of Healthcare Organizations to develop a standard requiring hospitals to assess the pre-admission tobacco use status of all patients to provide counseling. The metric is the number of patients admitted for myocardial infarction, congestive heart failure or respiratory disease who were offered any smoking cessation program. We have been involved at the next level, finding inexpensive methods of improving hospital adherence to these guidelines. We have evaluated the effectiveness of a simple strategy to increase adherence to the guidelines. The steps are listed in Table 2. We randomized all of the larger California hospitals via matched regions to this intervention or control to determine its effectiveness. Follow-up data is now being collected.

Table 2.  Staying Free hospital dissemination model.

Step 1. Identify all hospitals in a geographic region

Step 2. Collect the most recent results from the JCAHO performance measures for these hospitals on smoking cessation

Step 3. Prepare a letter reporting the hospital's performance on smoking cessation advice/counseling measures and notify hospitals that we will be publishing the data in their local paper within a year. Send the letter to the hospital CEO, Medical Director and/or Chairman of the Board, and one or two other relevant individuals (e.g. Director of Health Promotion or Quality Improvement)

Step 4. Provide regional face-to-face training if desired (at cost)

Step 5. Send a 6-month reminder letter

Step 6. Organize (and publish hospital performance data for smoking cessation advice/counseling in the local media about 1 year after the initial letter is sent

JCAHO, Joint Commission on Accreditation of Healthcare Organizations.

Improving access to psychological therapies (IAP) in the UK

The most impressive effort for disseminating CBT interventions is the recent adoption of CBT as the treatment of choice for mild–moderate depression in the UK. The essentials of this extraordinary development can be seen in Table 3. This historical event may have a profound impact on expanding the availability of CBT to populations.

Table 3.  United Kingdom cognitive–behavior therapy (CBT) program to improve access to psychological therapies.

Main goals

1) Primary care trust (PCT) coverage: At least 20 PCTs to implement the Improving Access to Psychological Therapies (IAP) program by 2008/9. About 40 properly trained therapists/250,000 people. Working in a team consisting of director, supervisors, low- and high-intensity therapists

2) Building a skilled workforce: Training programs to deliver 3600 therapists by 2010/11 with an appropriate skill mix (low and high intensity) and supervision arrangements

3) Extending access to evidence-based care: 900,000 more people accessing treatment, with the expectation that half of those who complete the program will recover completely, with 25,000 fewer sick days by 2010/11

4) Effectiveness: At least 90% of patients will have pre-post treatment outcome data. Assess reductions in symptoms, functional impairment, acceptability and quality, equity of access, population coverage

Treatment

1) Every person receives a “patient-centered assessment,” using a stepped-care approach

2) Mild–moderate depression: Watchful waiting, guided self-help (which can be delivered over the telephone, or brief face-to-face psychological interventions (up to seven sessions) or include guided use of computerized CBT

3) Moderate-severe: up to 20 sessions; including medication

4) Routine outcome monitoring

Cost

£33 million in 2008/09

£70 million in 2009/10

£70 million in 2020/11

Outcomes

Given the many still unresolved issues in dissemination, the limitations of evidence-based guidelines and algorithms for comorbid, complex conditions, the many practice constraints and therapist preferences, we believe that the single most important factor leading to successful dissemination will be the use of standard outcomes. Even with 40 years of research on CBT, many important questions related to dissemination remain unanswered, as suggested above. In short, it is not known if CBT techniques adapted to a particular client's needs by a skilled therapist achieve a significantly better result that a therapist following a structured routine. It does seem that most experienced therapists prefer to select from a variety of techniques rather than to follow a regimented program. Assessing the competency of therapists to apply proven techniques and the fidelity of treatment are difficult in most treatment settings. Most of the research has been conducted using samples exhibiting isolated (rather than comorbid) disorders, whereas most therapists help individuals with multiple comorbid problems. The principles and practice of supplementary manuals, books, Internet programs, self-help groups, on-line groups, etc. is also more clinical than empirical.

One solution to these problems is to focus on the outcome: patients should be expected to achieve a certain outcome by a certain time in therapy depending on their symptoms, severity and other factors. Therapists could do as they wish, within standard ethical and professional guidelines, as long as the patient improves. If therapists consistently fail to meet accepted outcomes, they might then be expected to follow a more standardized treatment. If they then adhere to the accepted protocol, a suboptimal outcome would more clearly be attributed to the intervention rather the interventionist. Thus we aim to prioritize the patient rather than the use of CBT. We believe that CBT can be a means to reaching adequate outcomes, particularly if a therapist's current strategies are ineffective.

Leadership

With this plethora of evidence, models, training programs and need, we are at a stage when CBT can have a wide impact on mental health problems. The future of the widespread implementation and dissemination of CBT techniques will rest with leadership rather than scientific knowledge. The development of the UK IAP program occurred because of a few individuals deeply committed to improving the care of mild–moderate depression and anxiety using evidence-based techniques. Such leaders are unlikely to receive much financial benefit from their activities and must do so in part out of commitment to the greater good and at some risk of alienating their own peers comfortable in traditional practices. We are honoring Gunnar Gotestam today, who is one such leader. He has played a critical role in evaluating and disseminating evidence-based treatments throughout Norway and his work has influenced many others. He has worked tirelessly to improve treatments, training, public understanding and access to the evidence-based treatment and knowledge. For this, those of us in the scientific community, and the many people who lead better lives because of his efforts, owe him much.

Acknowledgements

Dr Chang is supported by the NIMH postdoctoral research grant 5 T32 NIH MH019938.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.