5 May 2022

We read with interest the study by Brodeur et al^{Reference Brodeur, Courteau, Vanasse, Courteau, Stip and Fleury1} that makes a case for initiating clozapine in all eligible patients, based on the low absolute risk of subsequent poor adherence. Such a finding is reassuring, especially given the repeated finding that clozapine remains under-prescribed despite strong evidence of its superior efficacy. This is a secondary analysis of pharmacy claims data and as such has all the advantages (representativeness, size) and disadvantages (unmeasured covariates, unstandardised diagnoses, the lack of an a priori hypothesis) of any such analysis. There could also be concerns about the use of the MPR (medication possession rate, a measure of pharmacy attendance) as a proxy for treatment adherence, rather than more direct methods.

Other concerns are related to the data and the results themselves. We observe that the sample is heavily skewed towards males (66.3%), which goes against the roughly equal gender ratios that are usually observed in schizophrenia.² Although such and similar skews are often found even in previous studies of treatment-resistant schizophrenia,^{Reference Wesley, Kadra-Scalzo, Pritchard, Shetty, Broadbent and Segev3} the authors have not delved into this in any depth. Explanations for such a skew might relate to differences in comorbidity rates (e.g. substance use), perceived risk (e.g. aggression) or social support, which may influence the decision to commence clozapine. It is unclear whether or how these factors (that could not be studied in this study) may have influenced patients’ treatment adherence, both before and after clozapine initiation. However, most important are the observed rates of adherence before clozapine initiation. Most systematic studies quote an overall treatment adherence of between 40 and 50% for patients with schizophrenia.^{Reference Valenstein, Ganoczy, McCarthy, Myra Kim, Lee and Blow4} With this in mind, the observation that the greatest proportion (75.6%) of subjects in this sample fell into the category of those with ≥80% adherence, as measured by the MPR, suggests that this population was quite different in terms of their treatment adherence. It may be that patients are initiated on clozapine only if the treating clinician is confident of future adherence; it might also be that additional measures (e.g. assertive outreach) are provided to those patients who are treatment resistant, and that these factors may be responsible for their improved adherence. In this light, the finding that even those with relatively poor adherence pre-clozapine also show better treatment adherence is heartening.

Overall, it is difficult to conclude from these results alone that clozapine initiation itself is generally associated with an improvement in treatment adherence – any improvements might instead be attributed to factors unrelated to the medication itself, such as patient characteristics or other interventions that were concurrently provided to these patients. We would hope that future studies of this important topic would also adjust for the influence of these factors. This would, of course, entail the planning of a study with an a priori hypothesis and outcome measurement.^{Reference Velligan, Weiden, Sajatovic, Scott, Carpenter and Ross5}