ABSTRACT
BACKGROUND: There is a need to identify, as early as possible, infants who are at risk for long-term neurological morbidity. METHODS: To predict neurodevelopment outcome of preterm infants <30 weeks' gestation in a population of 100 infants, we used several neonatal and neurobehavioral tests, including cranial ultrasonography, the Prechtl neurological test, quality of spontaneous general movements, and quality of sleep-wake organization. RESULTS: The Prechtl test at corrected term age and findings on cranial sonograms both had high specificity, but the Prechtl test had better overall positive predictive power for normal neurological and developmental outcomes at 2 years' corrected age. Developmental changes in sleep and the amount of indeterminate sleep did not correlate with outcome. Scoring general movement quality did not predict outcome and did not augment the positive predictive power of the Prechtl test. CONCLUSIONS: The Prechtl test at corrected term age (independent of the other tests) is the best positive predictor of normal neurological outcome and Bayley test results at 2 years' corrected age.
Subject(s)
Developmental Disabilities/diagnosis , Infant, Premature , Double-Blind Method , Echoencephalography , Female , Humans , Infant, Newborn , Male , Predictive Value of Tests , Prognosis , Prospective Studies , Sensitivity and SpecificityABSTRACT
Three statistical methods are described and illustrated with working examples. Using these methods it is possible to include all subjects involved in cross-sectional studies in calculations of results, even if the data are incomplete at the end of the investigation. The use of censored data is taken from survival analysis. It is shown that combination of the Kaplan-Meier estimate, the log-rank test and the proportional hazards model gives more satisfactory results in nutritional and growth studies.