"If this were a medication, it would be big news," says Yale psychologist and psychiatry professor Kathleen M. Carroll, PhD.

"This" is an innovative series of modules called Computer-based Training in Cognitive Behavioral Therapy (CBT4CBT) that offers a version of cognitive-behavioral therapy for substance abuse via computer.

Created by Carroll and colleagues Samuel A. Ball, PhD, and Steve Martino, PhD, the project grew out of their observations that many clinicians who treat substance abuse weren't providing CBT at levels of adherence or skill close to the level provided in the trials that established its efficacy. With funding from the National Institute on Drug Abuse, Carroll has since discovered that computerized interventions offer more than just standardized treatment: When supplemented with traditional counseling, computer-based training in CBT can also improve outcomes.

In a study published in 2008 in the American Journal of Psychiatry (Vol. 165, No. 7), Carroll and colleagues randomly assigned 77 outpatients who met dependence criteria and who had used drugs or alcohol in the last month to receive the standard treatment of weekly individual and group sessions of general drug counseling or the standard treatment plus access to the computerized CBT4CBT program.

Eight weeks later, they found that 53 percent of the weekly urine specimens submitted by the treatment-as-usual group tested positive for drugs compared with 34 percent of the specimens submitted by those who also used the CBT4CBT program. The participants using computers also had longer periods of abstinence during treatment.

"I think this is the first time a computer-assisted intervention has shown outcomes on biologically verified indicators," says Carroll, noting that past research has relied on potentially more biased reports by clinicians or patients themselves. Although she cautions that this was a single, small study, she says, "that's impressive."

Supplementing face-to-face treatment

Carroll's path to developing the computerized intervention began in frustration: While researching CBT's effectiveness in substance use treatment settings, she found that many clinicians were not providing CBT with enough skill or intensity.

In a study of real-world practice in addiction treatment settings, published in the Journal of Substance Abuse Treatment (Vol. 35, No. 4), her research group found that many clinicians reported that they were doing CBT. However, an independent analysis of treatment sessions found that CBT strategies showed up in just 3 percent of them.

In a paper published in the Journal of Consulting and Clinical Psychology (Vol. 73), Carroll found that only with intensive training and supervision could the predominantly masters'-level substance abuse treatment providers in her study achieve minimal proficiency.

"Workshops are expensive, and there's huge turnover among clinicians in the addiction treatment system," says Carroll. "And supervision helps, but it isn't yet reimbursable and is rarely provided in a manner and with enough regularity necessary to ensure competent delivery of evidence-based therapies."

Carroll found a solution in computer-based treatments. Randomized trials by other researchers have already indicated that computer-assisted therapy for depression is both effective and cost-effective. What little research there has been on using computer-assisted therapy for addictions, however, has focused primarily on smoking. It was time to extend this promising approach to the challenging population of people addicted to alcohol, cocaine, opioids and marijuana, Carroll concluded.

Computer-based treatment could be especially helpful in breaking down barriers to care for those with substance abuse problems, she believes.

"For many reasons, including stigma and lack of access, only a small percentage of people with addiction seek treatment," she explains. In addition, she says, a computerized program could provide extra support in a field where clinicians often have very little time to spend with individual patients.

Carroll and her colleagues have now developed a seven-module CBT4CBT software program based on NIDA's treatment manual for CBT (http://www.drugabuse.gov/txmanuals/cbt/cbt1.html), which Carroll authored. The modules feature film clips depicting people struggling with real-life situations where drug use is likely, with opportunities for patients to reflect on what they might do in the situations. In addition to these movies, the modules include interactive graphics, games and assignments; practice exercises; and a narrator who guides users through the self-paced process of using the modules.

Each module focuses on a specific skill, such as recognizing and avoiding risky situations, dealing with cravings and negotiating tough decisions. There's also a module on HIV prevention.

"We tried to make it as user-friendly as possible to make it work for our patient group," says Carroll, who works primarily with low-income drug users. "No experience with computers is required, and there's hardly any text to read."

Enhancing outcomes

In an editorial accompanying Carroll's American Journal of Psychiatry study, John Greist, MD, of Healthcare Technology Systems calls her program "a paradigm-shifting approach that closes the gap to cost-effective, real-world delivery of an effective cognitive-behavioral therapy program for substance dependence."

Carroll's research participants have a more basic view: They love the program, says Carroll.

What's more, she adds, the effects last. In a follow-up study published in Drug and Alcohol Dependence (in press), the treatment-as-usual group improved but didn't get any better once the trial ended. The computer-assisted-training group, however, continued to make some improvements even after they stopped using the program.

Carroll has also collaborated with cognitive neuroscientists to conduct neuroimaging of 20 of her patients who participated in the CBT4CBT trial.

In a study published in 2008 in Biological Psychiatry (Vol. 64, No. 11), the researchers found relationships between indicators of brain activity during a cognitive control task and measures of treatment outcome. Participants' percentage of drug-free urine tests correlated with striatal activation, for instance. And longer periods of abstinence correlated with activation of the ventromedial prefrontal cortex, left posterior cingulate cortex and right striatum. The results, say the authors, provide insight into the mechanisms of the treatment.

"This is a very small sample," says Carroll, explaining that many patients weren't eligible for the fMRI study because they had metal objects in their bodies. "With this kind of sample size, it's absolutely impossible to tell if this is treatment-related, but the findings are intriguing."

Another benefit of computer-assisted treatment may be its cost-effectiveness. Even in the least favorable case, the additional program costs were fairly modest.

Next Carroll hopes to test the computer program with other patient groups. She wants to explore whether the program could be used effectively in methadone maintenance, primary care and other settings and evaluate each of the modules independently. And it will be important to test whether the program could work on its own rather than as a supplement to traditional treatment before it is widely released, she adds.

"The program is a skills-building machine. It doesn't build a relationship, ask you how you're doing or really care," she says. "It would be very interesting to compare CBT delivered by a human versus CBT delivered by a machine."

To watch Carroll present her findings at the National Institutes of Health Behavioral and Social Sciences Research Seminar Series, visit http://videocast.nih.gov/launch.asp?14653.

Rebecca A. Clay is a writer in Washington, D.C.