Cryotherapy in infants with retinopathy of prematurity. A decision model for treating one or both eyes.

Formal decision model analysis for cryotherapy to reduce blindness from retinopathy of prematurity shows that the decision to treat depends substantially on disease severity, the probability of long-term side effects, and correlations of treatment outcomes and side effects for individuals in whom both eyes are treated. The model indicates that both eyes should be treated regardless of disease severity only when the probability of long-term side effects is zero. As risk of long-term side effects increase, treating neither eye or only the worst eye is preferred for an increasingly large spectrum of disease. With high correlations of long-term side effects, treating the worst eye is preferred for moderate levels of disease severity. For high disease severity, treating both eyes is recommended, even when the probability of long-term side effects is high. Better-informed treatment recommendations must await new data to assess risks of long-term side effects and correlations of treatment outcomes and side effects.