Of relevance in this issue are the patients who did not receive surgery, but should have done. This would no doubt include the 14 patients, of 1084, who were still waiting 6 months after being listed for surgery.¹ Of the 85 patients who were removed from the list, 38 became too ill to risk surgery and 3 died; because their condition might have deteriorated or they might have died for reasons relating to their surgical condition, it can be argued that they too should have been included in the waiting time analysis.

Patients who were removed from the list either because their condition improved or because they decided not to have surgery speak more to the issue of list inflation. These patients should not have been included in the estimate of waiting times.

Even though all patients were included, the median waits were 6 (95% confidence interval [CI] 5–6) weeks for retrospective analysis and 7 (95% CI 6–7) weeks for prospective analysis. Because the confidence intervals overlapped, there appears to be no statistically significant difference. Medians, rather than means, are often preferred in measuring waiting times because of the tendency for the distribution to be skewed, with a long tail to the right of the distribution, with the result that most patients receive service in less time than the mean wait.

The ability to compare the results of prospective and retrospective methods of estimating waits adds a valuable dimension to the debate. The fact that the median waits calculated by the 2 methods were not significantly different supports arguments that retrospective methods of estimation are valid.