BACKGROUND: Structured medication reviews were introduced in 2020 to address polypharmacy in patients most at risk of medicines-related harm. AIM: To evaluate the impact of SMRs on prescribing in primary care. DESIGN AND SETTING: Retrospective observational cohort study of electronic health records from patients aged ≥65 years, prescribed one or more medications and fulfilling the specific eligibility criteria for a SMR, registered at practices contributing data to the Oxford Clinical Informatics Digital Hub (ORCHID), between 1 st April 2020 and 30 th September 2022. METHODS: The association between SMRs and prescription changes was examined by matching individuals who received an SMR to individuals who did not receive an SMR, according to age, sex and primary care practice using cumulative density sampling. Analyses were undertaken using adjusted logistic regression. RESULTS: Of 635,698 eligible patients, 82,285 (12.94%, 95% confidence interval [CI] 12.86%-13.02%) received at least one SMR during the study observation period. In those prescribed potentially inappropriate drug combinations prior to an SMR, between 12.5% and 40.0% were corrected up to three months later. In matched analyses, SMRs were most strongly associated with an increase in new prescriptions of ACE inhibitors (adjusted odds ratio [aOR] 1.56, 95%CI 1.35-1.81), statins (aOR 1.78, 95%CI 1.57-2.02), and antidepressants (aOR 1.45 95%CI 1.28-1.63). SMRs were also associated with stopping these drug classes in those previously prescribed treatment. CONCLUSIONS: SMRs were associated with starting new medications and stopping existing prescriptions compared to usual care. Further work is needed to understand if these changes improved patient outcome.