Editor, A recent discussion amongst us to improve the methods of scheduling operations led us to re-analyse the original data presented in the recent model.1 This revealed small errors in four of the 153 data points published in our article. These were due to incorrectly labelled scheduled surgical list times (i.e. a half-day list mislabelled as full day and vice versa; two errors) and incorrectly applied gap times (two errors). Indeed, the original article stated correctly that gap time was calculated as 10% of the total estimated list time, but later incorrectly in the text this was calculated to be 10% of the scheduled list time. Because gap times are known to be very modest,2 the effect of either choice is in fact trivial. More importantly, there is a correction to be made to the method of calculating a ‘pooled’ standard deviation (SD) for a theatre list (a series of cases, each with a known mean and SD for its individual duration). When planning a number of surgical operations (x) to fit into a scheduled list time, a formula for SD was used in the article that in fact yields the pooled SD for an estimated mean (i.e. the standard error): However, the correct formula for the pooled SD for the sum of the series of times is different: The effect of applying this correction (Equation B) is to make the SD for the lists somewhat larger than quoted, by an amount related to the number of operations in the lists analysed. Figure 1 shows the general relationship between the two equations for different numbers of operations (x). When x is greater than 1, the effect is that Equation B results in estimated probabilities somewhat less extreme than had been produced by Equation A. For the list data used in the original article, Fig. 2 shows how the probabilities calculated using Equation B differ from those originally presented.Fig. 1: No captions available.Fig. 2: No captions available.In turn, these somewhat larger list SD influence the calculation of the probability that a list will finish late. (Incidentally, these probabilities were estimated from the normal distribution, not the t-distribution, as erroneously stated in the original article.) Figure 3 shows a recalculation of Fig. 2 of the original article: the effect is that the prediction bars (±SD) are generally a little wider. Thus, the effect of larger list SD is to make the estimate of the finish time slightly more uncertain.Fig. 3: No captions available.The difference is relatively minor when the two figures are compared. In fact, for the data set used in the original article, these changes do not alter the conclusion; namely that for high estimated probabilities of over-run, most lists do over-run, and for low probabilities of over-run, most lists under-run. This is shown in Fig. 4, which is a recalculation of Fig. 4a of the original article.Fig. 4: No captions available.These improvements are now embedded in a new Excel spreadsheet (available at: link to website, https://links.lww.com/EJA/A32). This now includes several options for estimating the total gap time. In summary, the overall mathematical effect of larger SD for lists is to make the estimate of end time generally more uncertain, but the main conclusion of the original article is unaltered: it is possible to predict list scheduling in a manner that is better than non-quantitative methods or guesswork. Acknowledgements Assistance with the article: none declared. Financial support and sponsorship: none declared. Conflicts of interest: none declared. Presentation: none declared.