OBJECTIVES: Medication non-adherence is common in the treatment of patients with severe mental illness. Different formulations have been developed in an effort to improve medication adherence. The aim of this study is to explore whether there is a differential impact on treatment discontinuation between two different formulations of olanzapine: orodispersible (OD) or standard oral tablets (SOT) for the treatment of non-adherent patients with schizophrenia or bipolar disorder.  METHODS: This post-hoc analysis included 266 non-adherent patients diagnosed either with schizophrenia or bipolar disorder from an observational study (n=927) that measured the proportion of patients who discontinued treatment for any reason with olanzapine OD or SOT formulations over a 1-year period. Non-adherence was defined as having a baseline rating from 0 to 4 in the Medication Adherence Rating Scale (MARS). Treatment discontinuation was defined as discontinuing or adding a new antipsychotic to the index medication. A Kaplan Meier estimation of time to medication discontinuation was calculated. A Cox regression model adjusting for covariates was fitted to study the effect of baseline treatment on time to discontinuation.  RESULTS:  Patients treated with OD (n=177) vs. SOT (n=89) were more severe as measured by the Clinical Global Impression scale (CGI) (4.63 [SD 1.03] vs. 4.0 [SD 1.16], p < 0.0001) at baseline. During the 1-year follow up period the Kaplan Meier graph showed that patients treated with OD were less likely to discontinue treatment (11% vs. 27%, p<0.01). The Cox regression showed that patients taking OD had a significantly lower risk of discontinuing their baseline medication compared to patients taking SOT (hazard ratio: 0.35; 95% CI: 0.15-0.80). CONCLUSIONS: Treatment discontinuation was low with both olanzapine formulations; however the use of the orodispersible formulation in non-adherent patients with schizophrenia or bipolar disorder was associated with a significantly lower treatment discontinuation rate over a 1-year period.