Randomized controlled trials (RCTs) have strong internal validity but often have limited external validity. Observational studies have good generalizability and an increasing role in key healthcare decision making. We compared incidence rates of intracranial and major gastrointestinal bleeds in the low-dose aspirin arm (N = 9126) of the COMPASS double-blind RCT (conducted at 602 centres in 33 countries) with those from an observational cohort of preventative low-dose aspirin users (N = 54,140) in a primary care database representative of the UK general population – The IQVIA Medical Research Data UK (IMRD-UK). In our observational study analysis, we restricted follow-up to 2 years to be comparable with the duration of the COMPASS trial. Among low-dose aspirin users, incidence rates per 1000 person-years (95% confidence intervals [CIs]) in the IMRD-UK cohort and COMPASS trial participants, respectively, were 0.6 (0.5–0.8) vs. 1.4 (0.9–2.1) for intracranial bleeds, and 3.5 (3.1–3.8) vs. 3.7 (2.9–4.8) for major gastrointestinal bleeds. These broadly comparable bleeding rates among COMPASS trial participants and an observational cohort of low-dose aspirin users in IMRD-UK support the use of the latter for generating robust therapeutic evidence, and indicate that the rates from the COMPASS trial are broadly consistent with realistic population-based rates.