Compliance with Good Clinical Practice (GCP) guidelines should ensure the safety of trial participants and the reliability of trial results. Over the last decade, increasing emphasis has been placed on the role of costly on-site monitoring and source data verification as processes to demonstrate that GCP is being followed, despite a lack of empirical evidence that these are effective. To assess whether findings from on-site monitoring of a recent international multi-centre clinical trial could have been identified using central data review and other centralised monitoring techniques. Findings documented in a sample of site monitoring reports, and Programme Management Board Executive (PMBe) reports, from the Microbicides Development Programme (MDP) 301 trial - a randomised placebo-controlled trial of a microbicide gel to prevent vaginally acquired HIV infection conducted in four countries in East and Southern Africa - were extracted and individually assessed to determine whether they could have been detected in the trial database or through other central means. Four site visit reports contained 268 monitoring findings from a review of 104 participant files covering 324 study visits. Of the 268 findings, 76 (28.4%) were also identified in the study database. Central checks, had these been in place (such as central receipt and review of back-translated documents, enrolment and testing logs, informed consent, and more complex database queries), could have identified a further 179 (66.8%); 13 (4.9%) other findings (all minor) could have been identified through a review of the participant folder at site. The four PMBe reports reviewed included six major and three critical findings from a review of over 1000 participant files: only two of these (both major) were assessed as unlikely to be identified using central monitoring techniques. The study data used were not collected with this retrospective review in mind. It suggests that prospective work is needed to compare monitoring practices in real time. While there may be some categories of findings that it is not possible to identify centrally, the very large majority of findings reviewed in this analysis could be identified using central monitoring strategies. These data suggest that with better central and targeted on-site monitoring, it should be possible to identify and address most protocol and procedural compliance issues without performing intensive and costly routine on-site data monitoring.