Comparative Effectiveness of Antipsychotic Medications for Children and Adolescents

As a member of the American Psychiatric Association, I get a daily feed of news about psychiatry and mental health. The story today is "Study indicates newer antipsychotics may be no more effective than older medications for pediatric psychotic symptoms."

The National Institute of Mental Health, sponsor of the study (called TEOSS - "Treatment of Early Onset Schizophrenia Study"), summarizes it as follows:

The TEOSS team randomly assigned the children to eight weeks of either olanzapine (Zyprexa) or risperidone (Risperdal) — both new generation atypical antipsychotics — or to the older conventional antipsychotic molindone (Moban) plus benztropine, a medication often used to reduce side effects like uncontrolled shaking or tremor that can be associated with molindone. The children were monitored throughout the study by an NIMH oversight board to ensure their safety.

Response rates after eight weeks of treatment were comparable among the three medications — 50 percent of the children taking molindone improved, 46 percent taking risperidone improved, and 34 percent taking olanzapine improved. Children taking olanzapine or risperidone improved within the first two weeks, while the children on molindone improved within three weeks.

The treatment groups did differ in side effects. The children taking olanzapine gained about 13 pounds (6 kilograms) during the trial on average, while children taking risperidone gained about 8 pounds (3.6 kilograms), and those taking molindone did not gain weight. The olanzapine group also showed increases in cholesterol levels and other metabolic disruptions that may have become dangerous. The outcome prompted the safety review board to end the olanzapine arm of the study in 2006.

What struck me most in the first descriptions of the study and its implications was how careful the researchers were to avoid talking about health care costs. This is consistent with my sense that serious attention to the cost of care is still a forbidden academic activity outside of health policy circles:

"Atypical antipsychotics are commonly used to treat kids with [early onset schizophrenia], but these results question the wisdom of that approach," said lead author Linmarie Sikich, M.D., of the University of North Carolina at Chapel Hill. "They also remind us that we need to develop safer, more effective medications to treat these children, given the limited effectiveness of both the atypical and the conventional medications."

Dr. Jon McClellan of the University of Washington, a co-author of the new study and of the current guidelines for treating childhood schizophrenia, said in a telephone interview that older schizophrenia drugs should now be considered as an alternative in some cases. “Some of the children in this study gained 15 pounds or more in eight weeks,” Dr. McClellan said. “That’s as much as adults gain in a year on these medications. Children are especially susceptible to these side effects, and this has broad implications across the board, for the use of these agents to treat any disorder.”

If I were a state Medicaid director, or responsible for a pharmacy benefit management program, rather than pussyfooting around the fact of finite resources and the implications of this comparative effectiveness study, I would go to the community of child psychiatrists, pediatricians, and advocacy groups, and say "In light of the TEOSS findings [and other studies], is there a good reason why we shouldn't institute a fail first policy that favors the most effective, least costly medication option, and plan for other uses of the money we save, whether for other child services or to reduce overall costs to everyone?"