Letter

To the Editor: We contributed an article to the May issue's special section on interpreting the implications of CATIE (Clinical Antipsychotic Trials of Intervention Effectiveness) ( 1 ). As the section was being prepared for publication, the editor asked Richard Frank for a commentary on our article ( 2 ), which was also included in the CATIE special section. We are taking this opportunity to respond to Frank's commentary.

We agree that caution should be exercised in the implementation of formulary policies and that full consideration must be given to the concerns of advocates and consumers. We also think that mental health services researchers and economists are responsible to taxpayers and other purchasers of health care for offering a balanced consideration of the value of public expenditures. We think four issues may deserve further clarification.

First, Frank suggested that our appraisal of what is known may have been based too exclusively on the CATIE trial and that it may be premature to come to any conclusion about the cost-effectiveness of second-generation antipsychotics. However, the many studies we reviewed extend considerably beyond CATIE, which was only the most recent of a series of independent studies that have called into question the superiority of second-generation antipsychotics over older drugs. This perspective is perhaps most clearly reflected in the conclusion of the Texas Medication Algorithm Project authors, a distinguished group of schizophrenia experts, that second-generation antipsychotics should no longer be considered uniformly preferable to older drugs in the treatment of chronic schizophrenia ( 3 ). There are not likely to be additional large studies of these issues, so a broad consideration of current knowledge seems warranted.

Second, Frank expresses concern that the high drop-out rate from CATIE at 18-months seriously limits what can be concluded about these drugs. However, CATIE, like several other large trials that we cited, was substantially longer and had higher follow-up rates, at equivalent time points, than most previous trials and thus is more informative than previous studies, many of which lasted only six weeks ( 4 ) or stopped collecting data after the first medication change.

Third, Frank implies that we proposed placing strict restrictions on the use of second-generation antipsychotics. In fact, we suggested that a range of policy options be considered, most of which were not especially strict, and we sought to identify those that would be likely to be too precipitous or disruptive and thus should be avoided. Moreover, current practice is quite restrictive in that it relies predominantly on the five or six most expensive antipsychotics rather than on the full range of about 20 antipsychotic drugs currently available. One recent study showed that older drugs can offer new hope to patients who have failed to benefit from a second-generation agent ( 5 ). We believe that choices can be widened and costs lowered at the same time.

Fourth, we quoted a state Medicaid official who commented on the influence of coalitions between left-leaning advocacy groups and conservative businesses to highlight the political difficulties involved in considering cost-containment initiatives that would keep up with emerging research findings. Neither advocates nor businesses are focused on the cost concerns to the taxpaying public, and we think that mental health researchers should not ignore these concerns.

We agree that caution is warranted whenever health care policy is considered, but the need for caution should not, in itself, be a reason for inaction. The substantial public investment in government-funded studies such as CATIE can be justified only if new findings are carefully considered in developing policies that might better serve the public interest.