Exploring a potential healthy survivor bias in retinopathy of prematurity

To assess whether a healthy survivor bias contributes to the association between Black race and less severe retinopathy of prematurity (ROP).

This was a single-center, retrospective cohort study evaluating all infants eligible for ROP screening between January 1, 2012 and May 1, 2023. Race/ethnicity was considered as a combined variable: non-Hispanic Black, Hispanic or Latino, and non-Hispanic White. The outcome measures were type 1 ROP, death before 50 weeks' postmenstrual age (PMA), or the composite outcome of death or type 1 ROP. Multivariable logistic regression models were used to evaluate the association between race/ethnicity and each outcome measure, adjusting for birth weight, gestational age, sex, and area deprivation index.

Of 1,087 infants, 785 (72.2%) were non-Hispanic Black, and 48 (4.42%) received ROP treatment. Comparing non-Hispanic Black infants with non-Hispanic White infants, there were no significant differences in ROP treatment (OR = 1.54; 95% CI, 0.49-4.82), death before 50 weeks' postmenstrual age (OR = 1.60; 95% CI, 0.82-3.11), or composite outcome (OR = 1.63; 95% CI, 0.87-3.06). Comparing non-Hispanic Black infants with Hispanic/Latino infants, there were also no significant differences in ROP treatment (OR = 2.09; 95% CI, 0.92-4.81), death before 50 weeks' postmenstrual age (OR = 0.94; 95% CI, 0.52-1.72), or composite outcome (OR = 1.24; 95% CI, 0.72-2.14).

Data from a single-center, majority Black population indicate that a healthy survivor effect among Black infants was not detected. Further studies across diverse settings are needed, however, given the consequences of missing this potential bias in studies assessing race/ethnicity and ROP.