Our results show that the prevalence of schizophrenia increased with age in males and females aged ≤ 18 years. Pediatric schizophrenia is diagnosed through clinical assessment using the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition [14]. Two or more of the following characteristic symptoms must be present for a significant duration over one month (or less if successfully treated): delusions, hallucinations, disorganized speech, grossly disorganized or catatonic behavior, and negative symptoms. Therefore, patients must be old enough to communicate for an accurate diagnosis, and careful consideration is necessary regarding the prevalence observed in the 0–4-year age group in this study.

Aripiprazole and risperidone, FDA-approved for pediatric schizophrenia, were the main drugs prescribed for schizophrenia in children. Although these drugs have been tested for treating pediatric schizophrenia in RCTs [6, 8], in Japan, they remain unapproved for pediatric schizophrenia and are only approved for treating autism spectrum disorder (ASD)-associated irritability in children. Rapoport et al. reported that childhood-onset schizophrenia was usually preceded by and comorbid with ASD/pervasive developmental disorder in 30–50% of cases [15]. Although aripiprazole and risperidone are unapproved for schizophrenia in Japan, they may be prescribed for pediatric schizophrenia because it is comorbid with ASDs. Moreover, their approval for treating schizophrenia must also be considered in Japan.

The proportion of prescriptions for aripiprazole increased significantly from 2015 to 2022, whereas that for risperidone decreased significantly during the same period. Kumar et al. reported the short-term treatment efficacy and side-effect profiles of aripiprazole and risperidone for schizophrenia [16]. They found similar efficacy, while adverse events were more frequent with risperidone than with aripiprazole when assessed using the Udvalg for Kliniske Undersogelser side effect rating scale. Similarly, Coleman et al. reported that aripiprazole was superior to risperidone, considering its adverse effects [17]. Another study reported that risperidone induced more hyperprolactinemia side effects than did aripiprazole [18, 19]. Hyperprolactinemia is associated with menstrual disturbance in women [20]. Therefore, concerns about risperidone-induced hyperprolactinemia may have contributed to the marked reduction in the proportion of prescriptions for risperidone among females aged 13–18 years. Because a previous study has demonstrated similar efficacy between aripiprazole and risperidone [16], one reason for the increased trend toward aripiprazole and the decreased trend toward risperidone could be safety considerations.

We found that the prescription rates of aripiprazole and risperidone were higher in patients with schizophrenia, regardless of age, sex, and developmental disability. Aripiprazole and risperidone may be prescribed for reasons beyond efficacy and safety owing to their availability in several dosage forms, including tablets, orally disintegrating tablets, liquids, and powders. Therefore, a liquid or powder formulation can be administered to children, enabling a fine-tuned formulation design. Switching to tablets is also possible as patients age. These factors may have contributed to higher proportions of prescriptions for risperidone and aripiprazole than for other drugs.

The proportion of blonanserin prescriptions was low in our study. However, blonanserin was only approved for treating pediatric schizophrenia in Japan in 2021, resulting in the inadequate assessment of the drug within our study period; therefore, the prescribing trends must be monitored in the future.

This study has some limitations. First, regarding patients with both schizophrenia and developmental disorders, whether aripiprazole and risperidone were prescribed off-label specifically for schizophrenia or both schizophrenia and comorbid ASD is unclear. Second, comparing the proportion of antipsychotic prescriptions with that of each antipsychotic prescription may reveal concomitant cases. However, the MDV analyzer cannot distinguish concomitant from switched cases when multiple medications are prescribed within the same period. Third, the MDV analyzer does not provide information on the prescribing practices of general practitioners. Finally, this study was limited to outpatients, and patients with severe symptoms requiring hospitalization might have been excluded, possibly introducing a selection bias. Nevertheless, to our knowledge, this is the first study to clarify antipsychotic prescription trends for children in Japan. Therefore, our findings may contribute to the standardization of antipsychotic treatments for pediatric schizophrenia.