We are pleased about the interest in our report shown by Dr Padhi and colleagues and the opportunity to respond to the issues they have raised.

We did mention the types of comorbidities in our cohort, included to establish the similarity of eligible infants for whom there were insufficient outcome data to those for whom we had this information, trying thereby to reduce a potential source of bias. The writers do not state how more details would have informed our results—a single adverse outcome—or conclusion.

They have assumed, without basis, that our ophthalmologist authors possessed “different levels of ROP expertise.” The extent to which each of us performed the examinations reflects the allocation of duties within our department. As our endpoint (the need for treatment according to prevailing criteria for our environment) was not based on interpretation of subtle findings, the suggestion that photographic documentation, unavailable over much of the 10-year period of this study, and a masked observer were necessary seems better suited to prospective studies exploring other questions.

All examinations contributing to our results occurred at least 4 weeks after birth. In a few instances, examinations were done earlier owing to temporary unfamiliarity with our protocol by newly arrived junior neonatologists. Data from these examinations were not used and do not affect our results. We agree that “severe” can be variably interpreted; “treatable” was a substitute term for the degree of disease for which intervention was indicated under guidelines informed by well-respected prior studies.

We cannot explain the absence of outcome information for 4 of the 9 infants with mild ROP, despite repeated diligent searches. It is for each reader to determine the weight to be given to this circumstance; we could not, in fairness, have omitted this disclosure. For this and other reasons, we clearly indicate that our findings should be more rigidly tested by larger, hopefully prospective, studies.

Dr Padhi and colleagues describe their own experience regarding our study question. Other than observing that theirs may be a different environment with respect to ROP, a point that can be noted in our report, we decline further comment as we could not identify a citation in which we might review their methods and data beyond the brief précis in their communication.