The effect of clinical patient information on the diagnosis of and decision to treat retinopathy of prematurity

Abstract

Purpose : The internationally recognized retinopathy of prematurity (ROP) classification and treatment guidelines are based on morphologic disease variables found at examination for the condition. However, information about clinical variables associated with the development of ROP may influence the examiners’ grading and treatment decision. This cross-sectional agreement study evaluated the effect of clinical patient information on ROP experts’ diagnosis of ROP and decision to treat it.

Methods : Wide-field fundus photographs of eyes of 52 premature infants of <32 weeks’ gestational age or <1500 grams’ birth weight were presented on two consecutive days to seven recognized ROP experts for ROP diagnosis (stage, plus disease, aggressive posterior ROP) and the necessity for treatment. On assessment day 1 the experts were blinded to all clinical patient data whereas on assessment day 2 they were provided with information about the patients’ gestational age and birth weight. The complete study group was analyzed statistically followed by subgroup analysis of infants with a very low birth weight (1^st quartile, 580 grams) and infants with a heavier birth weight (4^th quartile, ≥850 grams). The McNemar test and the Wilcoxon signed-rank-test were used for statistical analysis.

Results : A statistically significant shift in the experts’ rating towards a less aggressive ROP grading stage (p=0.006) and a less frequent decision for intervention (p=0.021) could be observed after providing clinical patient information. The subgroup analysis showed that this was more true for infants with heavier birth weights (ROP stage p=0.001, treatment decision p=0.013), than for those with very low birth weights (ROP stage p=0.774, treatment decision p=0.703). No statistically significant shift could be found for plus disease (p=0.366) and aggressive posterior ROP (p=0.136).

Conclusions : The results of this study suggest that knowledge of clinical patient information such as birth weight and gestational age has an effect on the grading of ROP disease and the decision for treatment. ROP staging and the decision for treatment seem to be set at a lower level for babies with heavier birth weights. These findings may have implications for further refinements of the ROP guidelines and recommendations for clinical practice.

This is an abstract that was submitted for the 2016 ARVO Annual Meeting, held in Seattle, Wash., May 1-5, 2016.