PURPOSE: Mirror-image studies, which compare equal periods of time before and after a new treatment is introduced, may reflect the real-world impact of that treatment. However, most mirror-image studies that have investigated the impact of long-acting injectable antipsychotics (LAIs) were unidirectional in design, for patients switching from oral antipsychotics (OAPs) to LAIs. Therefore, we conducted a bidirectional mirror-image study comparing LAIs and OAPs. METHODS: We included 126 schizophrenia or schizoaffective disorder patients' LAI treatment data from 3 psychiatric hospitals. Patients took OAPs for 6 months or more before initiating LAIs, or the reverse. We obtained data on the number of hospitalizations as a primary outcome, plus the total duration and mean duration of hospitalization as secondary outcomes during the 6 months of the patients' first treatment, and the 6 months after the patients started their second type of treatment. RESULTS: The results indicated that there was no significant difference in any outcomes between LAI and OAP treatment when going from LAIs to OAPs (n = 59). However, when patients started with OAPs and switched to LAIs (n = 67), they were hospitalized a significantly fewer number of times, and the duration of their stays was shorter in the LAI phase than in the OAP phase. When combined with bidirectional data, LAI superiority was still observed. CONCLUSIONS: The findings endorse the relative effectiveness of LAIs over OAPs in the real world, although the inherent flaws of mirror-image studies such as expectation bias and having no parallel comparator should be considered.
PURPOSE: Mirror-image studies, which compare equal periods of time before and after a new treatment is introduced, may reflect the real-world impact of that treatment. However, most mirror-image studies that have investigated the impact of long-acting injectable antipsychotics (LAIs) were unidirectional in design, for patients switching from oral antipsychotics (OAPs) to LAIs. Therefore, we conducted a bidirectional mirror-image study comparing LAIs and OAPs. METHODS: We included 126 schizophrenia or schizoaffective disorderpatients' LAI treatment data from 3 psychiatric hospitals. Patients took OAPs for 6 months or more before initiating LAIs, or the reverse. We obtained data on the number of hospitalizations as a primary outcome, plus the total duration and mean duration of hospitalization as secondary outcomes during the 6 months of the patients' first treatment, and the 6 months after the patients started their second type of treatment. RESULTS: The results indicated that there was no significant difference in any outcomes between LAI and OAP treatment when going from LAIs to OAPs (n = 59). However, when patients started with OAPs and switched to LAIs (n = 67), they were hospitalized a significantly fewer number of times, and the duration of their stays was shorter in the LAI phase than in the OAP phase. When combined with bidirectional data, LAI superiority was still observed. CONCLUSIONS: The findings endorse the relative effectiveness of LAIs over OAPs in the real world, although the inherent flaws of mirror-image studies such as expectation bias and having no parallel comparator should be considered.