Clinical Outcomes and Patient Profiles in the UK Medical Cannabis Registry: A k-Means Clustering Analysis.

There is a paucity of high-quality evidence on the clinical efficacy of cannabis-based medicinal products (CBMPs). The objective of this study was to perform trajectory k-means clustering of health-related quality of life (HRQoL) outcomes in patients prescribed CBMPs to identify distinct response patterns and baseline predictors of treatment outcomes over 24 months. A cohort study of patients enrolled in the UK Medical Cannabis Registry with any qualifying indication was performed. Participants completed patient-reported outcome measures including EuroQol 5-Dimension 5-Level (EQ-5D-5L), Generalized Anxiety Disorder-7 (GAD-7), and Single-Item Sleep Quality Scale (SQS), at baseline, 1, 3, 6, 12, 18, and 24 months. Longitudinal k-means clustering was performed on EQ-5D-5L index values where the optimal number of clusters was selected via the gap statistic. Univariable and multivariable logistic regression analyses identified predictors of cluster membership. The 8945 patients were included in the analysis, from which 10 distinct trajectory clusters were identified, with eight demonstrating HRQoL improvements representing 77.72% of the cohort (n = 6952). Over 70% of participants reported improved EQ-5D-5L index values at each timepoint, whilst 54.21% (n = 4849) and 44.07% (n = 3942) achieved clinically significant improvements in GAD-7 and SQS at 24 months, respectively. Adverse events were reported by 13.65% (n = 1221) of patients, predominantly rated as mild (n = 4732; 42.31%) or moderate (n = 4860; 43.46%). Baseline patient characteristics, particularly treatment indication, severe anxiety, poor sleep quality, female sex, and cannabis-naïve status, were stronger predictors of favorable treatment response than product-specific factors.