Sir,

We thank Dr Yeo1 for the kind comments and would like to answer the specific points raised:

  1. 1

    The lists were training lists and trainees started on the straight forward cases and avoided, for example, white cataracts, pseudo-exfoliation, and vitrectomized eyes initially and we believe that this is the standard practice.

  2. 2

    The trainers were not blinded and there was no clear tendency to take over early, please note that completed number of cases in the 6-month block ranged from 151 to 194. In fact, the program was designed with reverse module training to ensure that this was minimized as repeatedly taking over cases both reduces confidence and interferes with learning, which we were keen to avoid.

There is a tension between the six sigma approach, adopted here, to try and reduce the risk of an adverse event to more than six standard deviations from the mean (or one in a million) with evidence-based medicine (which is two standard deviations from the mean (or one in twenty) and requires a control arm). The advantage of the six sigma approach is that it is potentially much more responsive to change, easier to implement, and has the added advantage of setting a higher standard to aim for. We think that the role of a randomized control trial is best indicated when there is equipoise between the competing options, which we do not think is the situation here. So, while the authors are supporters of evidence-based medicine and actively undertake clinical trials, here the improvements were achieved by implementing what we thought would be a much better system, in keeping with the six sigma approach, followed by evaluation. It should be noted that the trainee surgeons were not selected but were simply the next three in-line on the basis of their appointment dates to the program. Although the numbers are small, what we have shown is proof of principle that it is possible to train a cataract surgeon without an initial high-risk phase for the patient and we would now consider it inappropriate to revert, even for a control arm in a trial.

We see the challenge now is to make these results the norm and not the exception.