Follow-up data their use in evidence-based decision-making.

Experience with certain perinatal interventions, such as supplemental oxygen and dexamethasone, leads to the conclusion that follow-up data are needed to be well informed about the safety of certain perinatal interventions. Experience with indomethacin suggests that follow-up data also are regarded by some clinicians as a necessary aspect of evidence about effectiveness. Ideally, clinical trials of perinatal interventions might involve collection of data about neonatal predictors of outcome (such as a neuroimaging study and a standardized neurologic assessment); several developmental and neurologic assessments before school entry; a comprehensive evaluation of the child's cognitive function, behavioral competencies, and academic performance at 7 to 8 years of age; serial detailed assessments of the family psychosocial functioning; and an inventory of resources available for the child. Many clinical trials have not included follow-up after the neonatal period, and in such cases information about the effect of the intervention on participants' HRQL is incomplete. The approach taken in several recent trials, in which the outcome of interest is neurodevelopmental outcome at 18 months, attempts to strike a balance between a theoretical ideal (a comprehensive, longitudinal follow-up through school age) and a follow-up regimen that is not prohibitively expensive. Such trials include follow-up during the first 1 to 3 years of life, when major disabilities can be identified reliably, thereby providing moderately informative data about participants' eventual quality of life, related to the presence or absence of major disability. If, however, there is reason to suspect that the intervention has effects on the developing brain, follow-up after school entry may provide additional evidence pertinent to the risks and benefits of the intervention.