Pericyazine 2.5mg/5ml oral solution

BACKGROUND Data analysis can be used to identify signals suggestive of variation in treatment choice or clinical outcome. Analyses to date have generally focused on an hypothesis-driven approach. OBJECTIVE Develop hypothesis-blind data driven approaches to identify outlier prescribing behaviour. METHODS Here we report an innovative hypothesis-blind approach (calculating chemical-class proportions for every chemical substance prescribed in each Clinical Commissioning Group and ranking chemicals by (a) their kurtosis and (b) a ratio between inter-centile differences) applied to England’s national prescribing data, and demonstrate how this identified unusual prescribing of two antipsychotics. RESULTS We identified that, while promazine and pericyazine are barely used by most clinicians, they make up a substantial proportion of all antipsychotic prescribing in two small geographic regions in England. CONCLUSIONS Data-driven approaches can be effective at identifying unusual clinical choices. More widespread adoption of such approaches, combined with clinician and decision-maker engagement could lead to better optimised patient care. CLINICALTRIAL n/a

Abstract licence: CC BY

Abstract Background Data analysis can be used to identify signals suggestive of variation in treatment choice or clinical outcome. Analyses to date have generally focused on an hypothesis-driven approach. Methods Here we report an innovative hypothesis-blind approach (calculating chemical-class proportions for every chemical substance prescribed in each Clinical Commissioning Group and ranking chemicals by (a) their kurtosis and (b) a ratio between inter-centile differences) applied to England’s national prescribing data, and demonstrate how this identified unusual prescribing of two antipsychotics. Outcomes We identified that, while promazine and pericyazine are barely used by most clinicians, they make up a substantial proportion of all antipsychotic prescribing in two small geographic regions in England. Interpretation Data-driven approaches can be effective at identifying unusual clinical choices. More widespread adoption of such approaches, combined with clinician and decision-maker engagement could lead to better optimised patient care. Funding NIHR Biomedical Research Centre, Oxford; Health Foundation; National Institute for Health Research (NIHR) School of Primary Care Research and Research for Patient Benefit Research in context Evidence before this study Identifying variation in clinical activity typically employs a traditional approach whereby measures are prospectively defined, and adherence then assessed in data. We are aware of no prior work using data science techniques hypothesis-blind to systematically identify outliers for any given treatment choice or clinical outcome (numerators) as a proportion of automatically generated denominators. Added value of this study Here we report an innovative hypothesis-blind approach applied to England’s national prescribing data, to identify chemical substances with substantial prescribing patterns between organisations. As illustrative examples we show that promazine and pericyazine, while rarely used by most clinicians, made up a substantial proportion of all antipsychotic prescribing in two small geographic regions in England. Implications of all the available evidence The choice of antipsychotics between English regions could be further investigated using qualitative methods to explore the implications for patient care. More broadly, data-driven approaches can be effective at identifying unusual clinical choices. More widespread adoption of such approaches, combined with clinician and decision-maker engagement could lead to better optimised patient care.

Abstract licence: CC BY