Clarifications: open access follow up for inflammatory bowel disease

Editor—In our authors' reply to the letter by Barber and Thompson, we stated that further analysis of our data, using their recommended t test on bootstrapped data, had confirmed the results of the non-parametric tests used in our original article.^1–3 In other words, open access follow up for irritable bowel disease led to cost savings in secondary care. We have since discovered that a calculation error occurred in our bootstrapping exercise. The further analysis did not, in fact, confirm the statistical significance reported for secondary care, although it did confirm the original result that differences in overall costs (primary, secondary plus patient-borne costs) were not significant. We accept that bootstrapping is the preferred approach and that, on this basis, the study did not demonstrate resource savings in secondary care.

It is unfortunate that Bland's letter, published on bmj.com (bmj.com/cgi/eletters/320/7251/1730#8480) and in the journal,⁴ draws attention to a phrase in our authors' reply that was due to an unsanctioned editorial change and distorted our meaning.² We did not claim that the savings in secondary care were great, and we drew attention to this promptly in our electronic reply to Bland.⁵ This reply was not published in the journal with Bland's letter.