OBJECTIVES: To compare the risk of psychiatric hospitalization among adults with schizophrenia using aripiprazole with an ingestible sensor (AS), to those on oral atypical antipsychotic (OAA) treatment in the real world.

METHODS: A real-world data (RWD) platform, NeuroBlu was used to develop a retrospective cohort of adults with schizophrenia, using US MindLinc electronic health data. This RWD cohort was compared to adults with schizophrenia from a completed single-arm phase IIIb trial using AS. Risk of psychiatric hospitalization was the primary clinical trial endpoint. Inclusion criteria for the RWD cohort closely mirrored the AS trial, and the index date for the RWD cohort occurred 6 months post-initiation of an OAA. Propensity scores were used to match the AS and RWD cohorts on a 1:1 basis and to balance covariates between cohorts. Covariates were selected using stepwise regression and clinical judgement, and standardized mean differences (SMDs) were used for covariate balance diagnostics. Retrospective hospitalizations within 6 and 3 months pre-index were ascertained for both cohorts. The primary outcome was the number needed to treat (NNT) to prevent a hospitalization within 3 months post-index. RESULTS: Cohort matching with six variables (baseline CGI-S, duration of psychiatric hospitalizations, age, sex, race, and baseline OAA dose) produced 95 matched pairs out of 113 possible pairs. These covariates were well balanced (SMD < 0.1) across cohorts (mean/max SMD 0.041/0.08). There were 10 (10.5%) hospitalizations pre-index, and 0 hospitalizations post-index for the trial cohort. In the RWD cohort, there were 29 (30.5%) and 11 hospitalizations (11.6%) pre/post-index respectively. The post-index difference was statistically significant (p<0.05) and the NNT for AS was 9 (95% CI 6-20).

CONCLUSIONS: The results of this analysis are consistent with findings from the AS trial; it might suggest that treatment with AS for adults with schizophrenia indicates a reduction in psychiatric hospitalization.