Aims Medication count and appropriateness are often used as outcome measures to evaluate the effectiveness of deprescribing interventions. The aim of this study was to evaluate changes in prescribing, potentially inappropriate prescriptions (PIP) and use of low value medicines during the SPPiRE trial. Methods We retrospectively analysed trial prescription data from 51 general practices with 404 participants aged ≥65 years and prescribed ≥15 repeat medicines. A dataset was created with 7,051 ATC coded medicines at baseline. Outcomes were the most commonly prescribed and potentially inappropriately prescribed drug groups, the most frequently stopped or started drug groups and the number of changes per person between baseline and follow-up. Results There were 7,051 medicines prescribed to 404 participants at baseline. The most commonly prescribed drug group were proton pump inhibitors (82% participants) and statins (77%). There was a median of 17 medicines (IQR 15-19) at baseline and 16 (IQR 14-19) at follow-up. PIP represented 17.1% of prescriptions at baseline and 15.7% (n=6,777) at follow-up. There were reductions in the prescription of most drug groups with the largest reduction in antiplatelet prescriptions. Considering medication discontinuations, initiations and switches, there was a median of five medication changes per person (range 0-30, IQR 3-9) by follow-up. There were 95 low value prescriptions at baseline reducing to 78 at follow-up. Conclusions The number of medication changes per person was not reflected by summarising medication count at two time points, highlighting the need for repeated measurements of prescribing outcomes especially for populations with high degrees of polypharmacy.
