Gschließer and associates, in their article in this issue of the American Journal of Ophthalmology entitled “Inter-expert and Intra-expert Agreement on the Diagnosis and Treatment of Retinopathy of Prematurity,” question the efficacy of photographic screening for retinopathy of prematurity (ROP) based on inter- and intrascreening variability by experienced readers for ROP. This issue has been addressed by several studies looking at the grading of photographic images vs a bedside examination. It has been found in several studies that photographic screening can detect treatment- or referral-warranted ROP at a rate that is safe and comparable to live screening. The interpretation of an isolated image at one point in a progressive, but time-sensitive, disease such as ROP does not mean that treatment would not be provided in a timely fashion. The really important question this paper raises is, can we reduce subjectivity for ROP screening? The areas that are most likely to benefit from such a change are the determination of zone 1 disease and the determination of plus disease, both of which have treatment implications. The variability between “experts” or even averagely trained screeners is expected, given the use of clinical data, whether it is photographic or from bedside examinations. This type of variability has been true even in the bedside examinations of the CRYO-ROP study. Recently it has been shown by Chiang that doctors are poor at determining zone 1 based on live examinations, but very good at identifying the center of the macula on photographs, critical to determining the radius of zone 1. In addition, the determination of plus disease has been shown by this paper and others to be variable. Without the benefit of a photograph, which allows the image to be studied, the subjectivity of a live examination may be even greater with the doctor’s personal interpretation and then written comments; subjectivity is unavoidable.

Much of medicine is the doctor’s using objective information such as laboratory values and imaging to determine clinical management. The question is how subjectivity can be reduced for ROP screening. The authors point out that these observations suggest that software programs that help get objective data in regard to findings such as the accurate measurement of zone 1 and changes in vascular tortuosity indicating plus disease would reduce the subjectivity of ROP screening. Much of medicine is moving toward the use of telemedicine techniques for monitoring of retinal and other ocular disease. A physician who can balance the demographic data and photographic examination to determine treatment will still be needed, but we know now that even non-physician readers can provide readings that determine if a physician examination is needed. Physician adjudication of photographic images to determine the need for bedside examination or treatment will continue to be needed and adds an additional safeguard for remote ROP care. The future of ROP screening is to use photography and smart software to provide appropriate monitoring of babies at risk for blindness from ROP. Elimination of subjectivity in ROP and all of medicine may not be possible, but objective data with physician clinical oversight may be the best way to reduce subjectivity and provide a high quality of care. The authors are to be congratulated for bringing these issues to our attention.