Assessing the Impact of CATIE Results on Clinical Practice

SAN DIEGO —There has been a long-standing recognition that the lag time between the publication of a potentially practice-altering clinical trial and dissemination of the findings to community-based providers is long—typically years rather than months. It is a reality that has been especially true of psychiatry, where the challenge of applying clinical trial results is compounded by poor compliance and heavy dropout rates among patients with psychotic disorders.

There has been a long-standing recognition that the lag time between the publication of a potentially practice-altering clinical trial and dissemination of the findings to community-based providers is long—typically years rather than months. It is a reality that has been especially true of psychiatry, where the challenge of applying clinical trial results is compounded by poor compliance and heavy dropout rates among patients with psychotic disorders.Researchers believed that the publication of CATIE and other government-funded studies would represent a breakthrough in the extrapolation of clinical trial results to real-world patients. But what is the evidence that such trials as CATIE are having an impact on clinical practice, particularly with regard to those clinicians who regularly treat patients with schizophrenia? A new study conducted by researchers at Massachusetts General Hospital (MGH) and Harvard Medical School suggests that one year after the publication of the initial CATIE results, the long lag time is still in effect. The study also indicates that innovative educational programs may help increase providers’ knowledge of key aspects of important clinic trials.

Drawing on data collected from 10 full-day, live continuing medical education (CME) symposia held in the United States and involving more than 3,300 total participants, the researchers assessed preactivity and postactivity knowledge of a key detail of the CATIE study. Accuracy on the pre-/postactivity question improved at each of the program sites, and overall the proportion of correct answers increased from 24% prior to the CME program to 65% at its conclusion.

Of the nearly three-quarters of program respondents (73%) who reported at the outset of the CME program that they regularly treat patients with schizophrenia, 82% indicated that they were unfamiliar with the CATIE study results or that the results had not led them to change their clinical practice.

The findings by the MGH/Harvard researchers suggest that more work is needed to extend the reach of clinical trial data, according to Timothy J. Petersen, PhD, Associate Director of the Division of Postgraduate Education at MGH’s Department of Psychiatry in Boston, and a coauthor of the study. “We have a long way to go in terms of bridging the gap between key research that’s conducted and dissemination to the ‘front-line troops’—the community-based clinicians,” he said in an interview with NeuroPsychiatry Reviews. “CME is one mechanism to actually bridge that gap, and there are probably others. But our mission is to optimize the delivery of information through CME to reach those front-line people who treat the vast majority of patients.”

The MGH/Harvard study was presented at the 160th Annual Meeting of the American Psychiatric Association, with Dr. Petersen’s colleague, Anthony Weiss, MD, serving as the lead author.

The impetus for the study came, in part, from observations that the CATIE study and other high-profile trials—such as the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study and the Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) trial—were not having the impact they should have on clinical practice.

“Given that we now have these recent seminal trials and there have been a large number of publications in the most prestigious academic journals, we were interested as to the dissemination [of data],” Dr. Petersen said. “It’s one thing to have a study that’s published in a third-tier journal that’s based on a small sample size—many clinicians will not be aware of such findings. But with these high-profile studies, you would hope there would be a more rapid dissemination of findings and knowledge of the findings…. Indeed, that wasn’t the case here.”

The CATIE study was considered by some observers to be a disappointing trial, because it did not demonstrate any significant difference in effectiveness between newer antipsychotic medications and the older, less expensive agent, perphenazine. However, the study is of particular value for secondary outcome data indicating that the biggest challenge in managing patients with schizophrenia may be keeping them on any kind of approved treatment, Dr. Petersen emphasized.

“[For patients] who suffer from psychotic disorders, the rate of treatment noncompliance is extremely high, and even in a trial such as CATIE, where there were very controlled conditions and a lot of follow-up with patients and reminders of appointments, there was a very high level of dropout,” he said. “So those are the findings that we hoped clinicians would use to inform their practices.”

The lack of difference in efficacy between the newer versus older generation of drugs studied in CATIE was partly explained by the difficulty in keeping individuals engaged in treatment—which Dr. Petersen called the “perennial” problem of managing patients with schizophrenia. “The idea was, if we use newer treatments, will we retain more patients in treatment,” he said. “But, in fact, that wasn’t the case with three of the four newer agents.”

In the MGH/Harvard study prior to the CME activity, 24% of respondents correctly answered the pre-activity question, “In the recent CATIE trial evaluating the various treatment options for schizophrenia, what was the primary outcome measure?” The primary outcome measure in the CATIE study was the proportion of patients who discontinue treatment. “It’s technically termed ‘all-cause treatment discontinuation,’” Dr. Petersen said.

ROBUST AND APPLICABLE FINDINGS

In the MGH CME program, participating clinicians were asked to answer questions using an audience-response keypad. Of the 3,333 total participants, 33.2% responded to the pre-activity question and 48.5% to the postactivity question. The researchers included data representing only those participants who responded to both questions (n = 800; 24%). Dr. Petersen noted that this aspect of their findings illustrates one of the limitations of any kind of research involving surveys or live Q&A.

“You will always have a percentage of people who just choose not to participate in the Q&A aspect of the program—probably for a variety of reasons,” he said. “They feel like answering questions is an added burden or they’re eager to get to the lecture, so they don’t want to bother answering questions. There are [many] reasons.”

Despite this methodologic limitation, Dr. Petersen believes that the findings of the MGH/Harvard study are reasonably robust and applicable to further research. “Our sense is that going from about a quarter of participants getting the question right prior and then over 65% getting it correct after—that’s a very positive scenario with ‘pre-/post-’ questions,” he said. “You’d consider that a successful outcome of a teaching exercise.”

He and his colleagues are already applying the findings to the 2007 CME programs certified by MGH, incorporating the data with other pieces of information to modify the curriculum. For example, based on the present study, the researchers have added to some of the CME lectures patient case simulation videos, following which participants are asked how they would assess and treat a patient whose case they just viewed.

“This probably approximates more closely what practitioners will actually do in their office, as opposed to just asking them a question,” Dr. Petersen concluded. “So we’re pilot-testing the use of those kinds of case simulations to more accurately gauge the impact of the programs we’re delivering.”          

—Fred Balzac

Return to table of contents