PurposeRoutine data are needed to monitor ethnic health inequalities. The proportion of hospital discharge records with ethnicity information has been improving in Scotland. The aim of this paper is to assess whether routine data can provide valid comparisons of admission rates by ethnic group.Design/methodology/approachRoutine hospital admissions data in four NHS Boards were analysed by ethnic group and sex to compare incidence rate ratios (IRRs) for acute myocardial infarction (AMI) and coronary heart disease (CHD). A previous study linking health and census ethnicity information for 2001‐2003 provided the comparison standard.FindingsThere was a similar risk of AMI for South Asian compared to non‐South Asian people in 2009‐2011 and 2001‐2003. South Asian people and Pakistani women had higher risk of CHD than White Scottish people. The Other White group had higher and the White Irish lower risk of AMI admission in comparison to 2001‐2003 data.Research limitations/implicationsThe comparison used a different age range, did not include community deaths, covered a part of Scotland rather than the whole, and may have been affected by changes to denominators, which were based on the UK census 2001.Originality/valueThe similar IRRs for AMI from census linkage in 2001‐2003 and NHS data from 2009‐2011 suggest routine ethnicity data are valid in some NHS Boards. Analyses can reveal previously unknown variations to justify health improvement action. To maximise the precision of analyses, data completeness needs to be increased and sustained.