Well-deserved excitement over new discoveries about the neurobiology of mood disorders has tended to overshadow the real therapeutic advances in these same conditions. While we impatiently await the translational connections from basic science findings to clinical applications, as we indicated in Breaking Ground, Breaking Through: The Strategic Plan for Mood Disorders Research of the National Institute of Mental Health+(1), there is still much to be done in the arena of clinical research +(2). We have not developed a consistent set of treatments for many facets of bipolar disorder, nor have we fully understood and communicated how much of a fiscal and medical burden this disorder represents. The paucity of controlled clinical trials in geriatric mood disorders, particularly with newer antidepressant agents, is striking. In the conduct of clinical trials for acute episodes, we still are inconsistent with regard to how long a clinical trial should be and whether the endpoint of interest should simply be to achieve response or a full remission of symptoms. Finally, the success of placebo administration in research in mood disorders, especially in the last decade, makes the question of when placebos should be used and for how long in controlled clinical investigations a more difficult one to answer +(3).

These comments are stimulated by the four reports in this issue that address important current problems in clinical research in mood disorders. The constant and appropriate call for longer studies among patients with bipolar disorder is partially answered by Tohen and colleagues, who examined the longer-term safety and efficacy of two agents now commonly used to treat acute mania and for longer-term prophylaxis, olanzapine and divalproex. Their results are useful, pointing to the sustained efficacy of these agents after an episode of acute mania. Perhaps a more important aspect of this report is its focus on remission, not just response, in addition to its focus on relapse. The acknowledgment that bipolar disorder, like most recurrent mood disorders, is to be viewed as a long-term chronic recurring condition requiring long-term interventions is refreshing. It must also be noted, however, how difficult a problem treatment adherence (dropout) is in the treatment of this disorder and the extent to which this problem reduces the robustness of secondary analyses in this investigation. Indeed, we must find ways to improve the level of treatment adherence among patients with bipolar disorder, both in clinical trials and in routine clinical practice. All treatment arms of such studies need equal enhancement to increase completion rates, especially if complete recovery from the episode is the goal. Notwithstanding the other limitations of this study that are clearly acknowledged by its authors, the report adds substantially to our empirical evidence base about two commonly prescribed agents with antimanic properties and a high degree of safety.

Kim and Holloway present an interesting set of analyses on the burden and benefits of placebos in antidepressant clinical trials. Their conclusions are based on a simulated model of an 8-week placebo-controlled, double-blind, randomized trial. They argue that a depressed patient who agrees to participate in a randomized trial can expect an almost 25% lower chance of treatment response than participation in an individualized treatment approach. Kim and Holloway, however, made a number of somewhat controversial assumptions. They offer a relatively weak basis for establishing the probable efficacy of individualized treatment (they acknowledge that the quality of data for individualized treatment involves clinical judgment and that it is hard to estimate treatment response), do not acknowledge that the aim of treatment is remission or recovery, not response, and do not acknowledge that in individualized treatment in the community, a second prescription is filled less than two-thirds of the time, making their claim of a 50% response rate to an initial strategy in individualized treatment somewhat suspect +(2). Nevertheless, they raise the bar of responsibility to adequately inform patients of the risks and benefits of participating in placebo-controlled treatment trials.

In the third report, Schneider and colleagues report on an 8-week controlled trial in late-life depression with a selective serotonin reuptake inhibitor (sertraline) and placebo. Similar to the first report comparing two active agents, this collaborative investigation pointed out our reliance on multisite studies conducted in many cases in academic settings with pharmaceutical companies as partners. While the need for placebo-controlled trials for this particular indication in late-life depression is obvious, few investigations have been conducted. The findings of the study, while positive, highlight a series of dilemmas. Such trials require large study groups (of more than 750 subjects). Greater severity (anxiety and endogenous subtype) was associated with a more robust difference between treatment with sertraline and placebo. Eight weeks may be too short a time for studies in late-life depression to achieve adequate numbers of patients showing response—let alone remission. Finally, more sophisticated biostatistical procedures +(4) are now available for testing the role of moderators and mediators in randomized clinical trials.

The final report in this series is our study (Peele et al.), which was devoted to an evaluation of insurance expenditures from a large national database of employer-sponsored health insurance. Our focus on bipolar disorder confirmed our suspicion that bipolar disorder is the most expensive behavioral health care diagnosis for insured patients and their insurance plans. These costs are reflective of high inpatient admission rates and the high frequency of comorbid psychiatric and medical conditions. The limitations of the results are that the study was based on 1996 data and did not include pharmacy costs. While it is possible that the managed care approach has been able to reduce expenditures for bipolar disorder, we find that unlikely. This study underscores the gap between our current knowledge about how to manage a serious disorder with a comprehensive treatment strategy embracing specialized ambulatory services and current treatment practices.

The management of mood disorders requires the recognition that these conditions are frequently lifelong and the willingness to treat these disorders in a fashion similar to that of other chronic medical conditions (e.g., ischemic heart disease, diabetes, and asthma). This remains the crucial challenge to all who must assume accountability for the well-being of those who have depression and manic-depressive illness.

Address reprint requests to Dr. Kupfer, Western Psychiatric Institute and Clinic, Department of Psychiatry, University of Pittsburgh School of Medicine, Pittsburgh, PA 15213; kupferdj@msx.upmc.edu (e-mail).