Rates and predictors of adherence with atypical antipsychotic medication: a follow-up study of adolescent inpatients.

BACKGROUND: The use of atypical antipsychotics is increasing in adolescent populations, but little is known about adherence with these treatments. This study examined postdischarge adherence of adolescents treated as inpatients with either olanzapine or risperidone.
METHODS: Eighty-six (86) adolescent inpatients (43 per treatment) were contacted an average of 10 months (range, 90 days to 18 months) after discharge from the hospital. A structured interview examined compliance, side effects, and subjective impressions regarding the efficacy of treatment. Clinical symptoms rated by clinicians, self-reports from patients, hospital chart diagnosis, treatment group (risperidone/olanzapine), and demographic factors were used to predict compliance, as were the occurrence of side effects and subjective impressions of the efficacy of treatment.
RESULTS: Forty-five percent (45%) of the patients were adherent with their medications, while only 12% discontinued treatment on their own, and 43% stopped medication at the request of their parent or a physician. Medication status and symptom severity did not predict nonadherence, and the only diagnosis related to adherence was substance abuse. The only side effect that predicted nonadherence was rapid weight gain during the hospitalization; this was slightly more common in olanzapine-treated patients. Nonadherent patients' subjective impressions of efficacy were no different from the patients who continued their medications, but the patients who stopped medication on their own were more likely to also have failed to attend recommended psychotherapy as well. IMPLICATIONS: Failure to continue recommended treatment with atypical antipsychotics was most common in cases who did not comply with other aspects of the discharge plan. Efficacy of the medication did not predict nonadherence, as half of the noncompliant subjects stated that the medication had been helpful. Rapid weight gain also predicted nonadherence, but subjective impression of long-term weight gain was not different across adherence or treatment status. These data tentatively suggest that in adolescent patients, tendencies toward general uncooperativeness and substance abuse may be at least as important as the occurrence of common side effects in the determination of medication adherence.