Psychopharmacology : Use of Placebos in Clinical Trials Involving Children and Adolescents

Medications are commonly prescribed to treat child and adolescent psychiatric disorders, despite limited research about the safety and efficacy of their use in this population (1). The lack of data on medication treatment creates a dilemma for the physician. Added to this dilemma are pressures on the physician to treat patients quickly and with the least costly intervention, a circumstance that is often interpreted to mean that medication should be prescribed early in the treatment process. Clearly, additional data are needed to improve our database on the safety and efficacy of psychopharmacologic agents for children and adolescents and to serve as a rational guide for clinical treatment.

At the same time, controversy has developed about the use of placebos in psychopharmacological clinical trials involving children and adolescents as subjects. Some investigators have proposed that the use of placebo treatment is unnecessary and unethical, serving only to deprive subjects of active treatment; a version of this position has been outlined by Rothman and Michels (2). In fact, some have proposed that it would be preferable to use an active placebo, that is, a medication or a medication dosage thought to be ineffective in the treatment of the study condition.

Unfortunately, child and adolescent psychiatric disorders are not well understood, particularly because of the lack of data on the course and the outcome of these disorders. As such, it is impossible to predict changes in symptoms for individuals or groups of individuals within any sample, much as is the case in research on adult psychiatric disorders (3). Therefore, without a placebo control within the study, researchers have no basis for judging whether a given study medication was associated with improvement or worsening of symptoms. Furthermore, subjects in child and adolescent trials often have a high rate of response to placebo.

This paper summarizes the findings of two recent studies—one part of a controlled trial and the other a naturalistic study. The findings highlight the need for clinicians to exercise caution in prescribing medication early in the course of assessment and emphasize the need for placebo in clinical trials.

Lithium for treatment of aggression

One of the authors (RPM) conducted a six-week double-blind and placebo-controlled study of the safety and effectiveness of lithium in decreasing aggression among children and adolescents with conduct disorder. Subjects were patients admitted to a child and adolescent psychiatry inpatient service who had a history of severe aggression and DSM-III-R conduct disorder. Those with diagnoses of mental retardation, affective disorder, substance dependence, or a psychotic disorder were excluded from the study.

The six-week study included a two-week baseline period during which placebo was administered and a four-week treatment period during which lithium was compared with placebo. Only subjects who demonstrated sufficient aggression as inpatients during the baseline period could enter the treatment phase of the study. That is, the criterion for random assignment to the treatment phase was that a subject show a significant level of aggression prospectively while receiving placebo during the washout period.

As Malone and associates (4) previously reported, close to half of the subjects—21 of 44, or 47.7 percent—did not show enough aggression to warrant random assignment to the treatment phase. It was unclear whether the subjects who did not demonstrate aggression were responding to placebo, the hospital milieu, or other factors. However, based on their symptoms in the hospital, it was clear that they were not candidates for lithium treatment. Had they received medication shortly after admission, it would have appeared that they were having a good response to medication when in fact they improved without active medication treatment. These findings suggest the need for allowing enough time for observation between the initial evaluation and the start of medication treatment to avoid needlessly treating children and adolescents with medications whose use in this population is often not well studied.

Stability of major depression

Another study by the first author examined the stability of major depression among child and adolescent inpatients. Sixty-six subjects, consisting of 26 male patients and 40 female patients age eight to 17 years, were followed naturalistically and assessed for major depression. All had been admitted to the hospital with a clinical diagnosis of major depression or after a suicide attempt. Instruments used in the assessments were the major depression module of the Diagnostic Interview for Children and Adolescents, DSM-III-R version (DICA) (5) at one, seven, and 14 days after admission and the full DICA at admission. In addition, depressive symptoms were assessed using the Hamilton Rating Scale for Depression, the Beck Depression Inventory, and the Children's Depression Inventory at days one, three, seven, and 14.

Only 34 subjects, or 51.5 percent, met DICA criteria for major depression at day one after admission; four subjects, or 6.1 percent, at day seven; and seven subjects, or 10.6 percent, at day 14. In all, of the 34 subjects who met DICA criteria for major depression at day 1, only two met the criteria at each subsequent assessment. That is, the diagnosis was not stable. Likewise, the measures of depressive symptoms all decreased significantly over the two-week assessment period.

It is noteworthy that most of the subjects were not being treated with psychotropic medications, including antidepressants, during the assessment period. These children and adolescents initially seemed to meet criteria for major depression, but by and large the symptoms remitted without medication treatment. Again, had they been treated early in the course of the assessment, improvement may have mistakenly been attributed to the effects of medication. Moreover, the high incidence of symptom remission points to the need for placebo controls in studies for such populations.

Discussion and conclusions

The two studies described above illustrate some of the problems facing psychopharmacologic research with children and adolescents and, indeed, with any population. The use of placebos is the subject of much debate in psychiatry in general at this moment. For example, in adult psychiatry, some writers have claimed that antidepressants do not work and that placebos are as good (6,7). On the other hand, some have commented that to give a placebo is unethical when "proven active treatment" such as antidepressants is available (2). Even disorders such as schizophrenia, whether acute or chronic, are subject to a placebo effect (8,9,10) that is not as large as in depression, but is still significant. Recent studies of obsessive-compulsive disorder have shown a similar trend (11,12).

What are we to make of this? In a broad sense, the environment, positive or negative, has an effect on all illnesses. In the course of a disorder such as pneumococcal pneumonia, the effect is relatively unimportant. However, in less precise diagnostic situations where multiple forces may be at work, the effect is more complicated. The therapeutic effect of treatments, including hospitalization, has been discussed elsewhere (4,,13). These effects are important in evaluating the outcome for any intervention because nonspecific factors are at work on all occasions. We would emphasize, however, that where the database is inadequate—as it is in child and adolescent psychopharmacology, particularly for long-term outcome—we have to be somewhat skeptical about categorical statements about the nonuse of placebo.

In the study of lithium described above, if placebo had not been administered during the baseline period and patients had been treated openly with lithium, one would be justified in thinking that lithium was a superior treatment for this disorder. We may have then tried to explain why it did not work with longer-term maintenance. We would have carried out some studies on compliance and would then be forced to study social factors that appear to promote if not perpetuate the disorder. Similarly, in the depression study, the addition of an antidepressant would yield a spectacular result; these drugs would be prescribed for innumerable children, many no doubt would be kept on maintenance therapy.

These findings should be kept in mind in assessing other studies of childhood psychopharmacology. It is also important, in these days of much-discussed evidence-based medicine, that institutional review boards and practitioners, not to mention health maintenance organizations, be aware of the real evidence. With all that has been said, the fact remains that placebo has just as much evidence for efficacy as most drugs for treating child and adolescent psychiatric disorders. Because drugs are so often used to treat this population, we must work to learn more about these treatments. Ultimately, learning about the various drug treatments that are used clinically will require placebo-controlled trials.

Acknowledgment

This work was partly supported grant MH-00979 (RPM) from the U.S. Public Health Service.

Dr. Malone is associate professor of psychiatry and director of child and adolescent psychiatry research at the MCP/ Hahnemenn University School of Medicine and Eastern Pennsylvania Psychiatric Institute in Philadelphia. Dr. Simpson, editor of this column, is professor in the department of psychiatry at the University of Southern California School of Medicine in Los Angeles. Address correspondence to Dr. Malone, Department of Psychiatry, Eastern Pennsylvania Psychiatric Institute, 3200 Henry Avenue, Philadelphia, Pennsylvania 19129 (e-mail, [email protected]).