The Diabetic Retinopathy Study-Reply

Reply To the Editor. —The investigators in the Diabetic Retinopathy Study (DRS) appreciate Professor Phillips' interest in our study and we share his concern about observer and patient bias in any serious attempt at evaluating therapeutic efficacy. The existence of both of those biases has been demonstrated often, yet emphasized inadequately. Any clinical investigator who assumes, without efforts to verify, that these biases are not present may be seriously deceiving himself. The ideal defense against these biases, if it can be made to work, is a "double masked" (we prefer this term to "double-blind," particularly in ophthalmology) study in which neither the patient nor the observer knows how or whether the patient was treated. The inclusion of a sham procedure was considered for the DRS, but is neither ethically justifiable, because retrobulbar injection would be required for some patients, nor feasible, since diminution of peripheral field or night vision or