malignancy overall (per 100 person years of follow-up) was 0.26; for patients with intestinal metaplasia the risk was 0.40 (95% confidence interval (CI) 0.26 to 0.59) and for those without intestinal metaplasia the risk was 0.06 (95% CI 0 to 0.32). In other words, if intestinal metaplasia was absent in biopsy specimens, the risk of oesophageal malignancy was not significantly higher than that in the normal population. Further, we found 93% concordance among the 2969 patients between intestinal metaplasia status on the first biopsy and that on any subsequent biopsy (unpublished data). Intestinal metaplasia seems to be either there from the start or absent. Given that these data are derived from ‘‘real world practice’’, where probably few biopsy specimens per patient were taken and hence sampling error might have occurred, it is reassuring to find that the risk of malignancy in patients in whom intestinal metaplasia was not shown was low. These patients may have a ‘‘patchy’’ distribution of intestinal metaplasia in the segment of columnar mucosa that is biologically distinct from those where the distribution of intestinal metaplasia is uniform, but this is purely speculative. Patients with Barrett’s oesophagus are at low risk of oesophageal adenocarcinoma; refinement of surveillance programmes is needed to focus resources on those most likely to benefit from surveillance—perhaps concentrating on those patients in whom intestinal metaplasia is evident at initial endoscopy is one way to do this?