Representativeness of enrollment into a perinatal research repository

To describe enrollment and compare demographic and clinical characteristics of neonates enrolled in the perinatal research repository (PRR) with eligible, but non-enrolled neonates.

Characteristics from enrolled infants were compared with eligible, but non-enrolled infants using t-tests, Wilcoxon rank tests and χ(2) tests.

Between 1 January 2010 and 31 December 2011, 622 infants admitted to the neonatal intensive care unit (NICU), born < 37 weeks were screened for the PRR were eligible for enrollment. Of these, 233 enrolled in the study. There were no significant differences between enrolled and non-enrolled infants with regard to race, birth weight, gestational age and Apgar scores. Enrolled patients had a significantly longer length of stay [mean (standard deviation) = 55.0 (58.9) versus 38.0 (45.8) days; p < 0.01]. Mortality was significantly greater in the non-enrolled group (11.7 versus 3.0%; p < 0.01). Both groups had a similar total number of diagnoses [mean (standard deviation) = 5.7 (3.1) versus 5.8 (3.2)]; however, among survivors enrolled infants had slightly lower prevalence of hypoxic ischemic encephalopathy (0.0 versus 2.9%; p < 0.01) and seizures (1.8 versus 6.4%; p < 0.01). There were no significant differences in enrollment by prevalence of other morbidities.

These results suggest that a specimen collecting repository can enroll a sufficiently representative sample of eligible patients.