Introduction. This is a study of the process of diagnosis in family medicine (FM) in four practice populations from the Netherlands, Malta, Serbia and Japan. Diagnostic odds ratios (ORs) for common reasons for encounter (RfEs) and episode titles are used to study the process of diagnosis in international FM and to test the assumption that data can be aggregated across different age bands, practices and years of observation. Methodology. Participating family doctors (FDs) recorded details of all their patient contacts in an episode of care (EoC) structure using the International Classification of Primary Care (ICPC). RfEs presented by the patient and the diagnostic labels (EoC titles) recorded for each encounter were classified with ICPC. The relationships between RfEs and episode titles were expressed as ORs using Bayesian probability analysis to calculate the posterior (post-test) odds of an episode title given an RfE, at the start of a new EoC. Results. The distributions of diagnostic ORs from the four population databases are tabled across age groups, years of observation and practices. Conclusions. There is a lot of congruence in diagnostic process and concepts between populations, across age groups, years of observation and FD practices, despite differences in the strength of such diagnostic associations. There is particularly little variability of diagnostic ORs across years of observation and between individual FD practices. Given our findings, it makes sense to aggregate diagnostic data from different FD practices and years of observation. Our findings support the existence of common core diagnostic concepts in international FM.