Prognostic neurodevelopmental testing of preterm infants: do we need to change the paradigm?

Longitudinal follow-up with assessment of developmental status at about 2 years of age is routine for high-risk newborns. The results of these assessments can be used for many purposes, including helping physicians, parents, and teachers plan educational or developmental interventions. These assessments also provide outcome measures for clinical research studies. Outcome results may also serve as a source of information for clinicians when counseling parents regarding provision of care for extreme preterm infants. Consideration should be given to use of different outcome metrics based on the purpose for testing. Categorization of composite cognitive, motor and neurosensory findings to define levels of impairment should be limited to research. Planning for individual interventions is better guided by descriptive findings. Current tools for assessing neurodevelopmental status at 2 years of age have important limitations. First, outcomes at early ages do not always predict function later in life. They are, at best, an estimate of longer-term outcomes, with important individual variation. For infants without severe neurologic injury, postnatal environmental factors play a predominant role in determining long-term cognitive and academic outcomes. Further investigations should assess quality of life and other considerations that are important for parents when making decisions about neonatal intensive care unit care for their infant.

Predictive utility of the Bayley Infant Neurodevelopmental Screener (BINS) risk status classifications: clinical interpretation and application.

Predictive validity and clinical implications of the increasingly popular Bayley Infant Neurodevelopmental Screener (BINS) risk status classifications have not been previously reported. In this longitudinal follow-up study, the BINS was administered to high-risk infants at 6, 12, and 24 months of age, and the McCarthy Scales at 3 years of age. Ninety-two children were evaluated at 6 and 36 months, 105 at 12 and 36 months, and 118 at 24 and 36 months; 190, 125, and 140 infants were included in the comparisons at 6 to 12, 6 to 24, and 12 to 24 months. BINS risk status was classified as low, moderate, or high; or as a binary variable, LOWRISK/HIGHRISK. The three BINS items groups were moderately correlated. Consistency was most variable in the moderate-risk group. BINS risk was predictive of 36-month function in 18 out of 18 comparisons. Odds ratios, ranging from 2.76 to 54.70, were significant in 15 out of 18 logistic models. An early high-risk classification was associated with increased probability of later developmental morbidity. The BINS offers an alternative to detailed assessment in high-volume clinical applications and has good concurrent and predictive validity.

Conceptual issues in developmental screening and assessment.

Effective screening requires an understanding of underlying conceptual issues and their relationship to pragmatic concerns. Pragmatic concerns include the concepts that there are many underlying reasons for an "abnormal" screening result; that sensitivity and specificity should be combined with relative risk when considering developmental outcome; and that patterns of congruence among motor, language, cognitive, and adaptive/personal social areas of development should be considered. Important conceptual issues include the following: there is continuity of underlying processes or functions in development; canalized behaviors might give the appearance of discontinuity; integrated functions are more predictive of later developmental levels than are individual functions; the "window" of assessment and the developmental emergence of a specific function will affect screening results; one must consider biologic and environmental risks and their specific effects; and different types of neural structures and their relationship to environmental input help to explain why screening results vary over time.

Perinatal asphyxia: effects of biologic and environmental risks.

The literature on biologic risks, such as asphyxia, and later outcome contains many confounding issues. Earlier data may not be applicable today because of high mortality rates and significant changes in medical practice. Data from a more contemporary cohort indicate an association between arrays of biologic variables and outcome, although early developmental/behavioral measures are more predictive. Some variables had a 'sleeper effect,' which did not surface until 36 months. Suggestions for better delineation of relationships between nonoptimal risk variables and outcome are provided.

Attention deficits in school-aged children and adolescents. Current issues and practice.

Attention deficits in school-aged children and adolescents pose unique diagnostic and management challenges for the practitioner. While this symptom reflects a heterogeneous spectrum of disorders with varying etiologies, clinical manifestation, treatment needs, and outcomes, there are sufficient common attributes to enable identification and treatment. Accurate diagnosis relies on synthesis of information from history, rating scales, direct observations, and psychometric testing. Close attention must be paid to potential underlying or associated medical, processing, emotional, and psychosocial problems. Multimodal treatment is essential with close collaboration between the child, parents, and professionals. While there is a risk of negative outcomes if the symptoms are left unattended, many of these children also harbor strengths that need to be identified and nurtured.

Data analysis techniques in behavioral pediatrics: some practical advice.

A review of the last three volumes of the Journal of Developmental and Behavioral Pediatrics revealed that 27% of the articles used correlations, 17% chi 2, 16% t tests, 15% analysis of variance or covariance, and 13% involved regression analysis. Multivariate techniques, discriminant or factor analysis, logistic regression, and sensitivity/specificity were used in less than 5% of articles; the average number of statistical techniques per article was 2.5. These statistical techniques are described in general, conceptual terms in regard to appropriate and inappropriate usage. Clinical examples are provided. Familiarity with these techniques allows the reader "statistical access" to almost 90% of the articles in behavioral pediatrics and other areas of the medical literature.

Outcome for preschoolers of very low birthweight: sociocultural and environmental influences.

The impact of the environment on the outcome of children of very low birthweight at preschool age was assessed. Embracing an ecologic perspective, the study incorporated several family and sociocultural variables. The role of the environment was significant in explaining the cognitive, language, and temperamental status of these children at age 3 years. Although the high-risk cohort performed less well than a matched comparison group, maternal IQ and family sociocultural status essentially obscured differences on two of the three outcome measures. In addition, all three high-risk groups--children of very low birthweight with and without concomitant medical complications, and full-term children with perinatal asphyxia--obtained scores on all outcome measures within one standard deviation of the mean of the scores of the comparison group. Favorable early environmental circumstances appear to compensate frequently for difficult prenatal and perinatal insults. Biologic and environmental factors interact and appear differentially to influence emerging cognitive, language, and behavioral functions.

Outcome studies of low birth weight infants published in the last decade: a metaanalysis.

We conducted a metaanalysis and methods review of 80 studies, published in the last decade, that explored the outcome of low birth weight infants; 27% involved infants whose birth weights were less than or equal to 2500 gm (low birth weight), 44% less than or equal to 1500 gm (very low birth weight), and 29% less than or equal to 1000 gm (extremely low birth weight). Problems found in these studies were grouped into three categories: subject and methods issues, environmental factors, and outcome measurement. The combined average intelligence quotient/developmental quotient (IQ/DQ) of all low birth weight groups was 97.77 (SD 6.19); for control subjects the mean IQ/DQ was 103.78 (SD 8.16). This difference was statistically significant but perhaps not clinically significant. No differences in mean IQ/DQ scores were found among the low birth weight, very low birth weight, and extremely low birth weight subgroups. Statistically significant differences among all groups and control subjects were found when categoric data were analyzed, as were differences among the three subgroups; however, the variety of outcome criteria makes interpretation of the categoric analyses difficult.

Correlation of asphyxia and other risk factors with outcome: a contemporary view.

The relationships between 14 maternal/prenatal, 24 perinatal and 12 asphyxia-related variables and outcome at 40 weeks, nine, 18 and 36 months were evaluated in a sample of 608 children enrolled in the multi-center National Heart, Lung and Blood Institute Collaborative study. Correlations were weak generally, although they tended to be stronger for perinatal variables and neurological outcome at 36 months. When the maternal/prenatal and perinatal variables (medical/biological), the SES-Composite Index (environmental/psychosocial) and the Early Neuropsychologic Optimality Rating Scale-9 (behavioral/developmental) were used as optimality scales, prediction was enhanced; however, it appeared that the environmental/psychosocial and behavioral/developmental scales were more predictive of 36-month outcome than were medical/biological variable groupings.

Impaired mental capabilities and hand function in myelomeningocele patients.

Hand function and intelligence were evaluated in 82 spina bifida cystica patients, aged 9 to 20 years. The patients required significantly more time to complete hand function tests and had significantly lower IQs than the test standardization populations. Inverse relationships were found between hand function times and IQ scores; patients with lower IQs (especially those with IQ scores less than 80) took significantly longer to complete the hand function tests. Patients with three or more shunt operations for correction of hydrocephalus required more time to complete the hand function tests and had lower IQs than patients with no shunt operations. Patients with higher level spinal cord deficits (particularly those with thoracic level lesions) required more time to complete the hand functions tests and had significantly lower IQ scores than patients with lower level lesions.

The Early Neuropsychologic Optimality Rating Scale (ENORS-9): a new developmental follow-up technique.

The development and application of the Early Neuropsychologic Optimality Rating Scale (ENORS-9) is described. This 24-item technique, scored in an optimal/nonoptimal fashion, enables assessment of posture, tone, and movement, as well as developmental delays. The ENORS-9 was compared with the Bayley Mental and Psychomotor Developmental indexes and a neurologic examination administered at 9 months, in terms of sensitivity and specificity of cognitive, motor, and neurologic outcome at 36 months. Using an ENORS-9 cut-off score of 85% for cognitive and motor, and 75% for neurologic outcome, sensitivity values were more than three times greater than more traditional assessment techniques. This instrument meets many needs in developmental follow-up because of straightforward scoring, decreased personnel requirements and cost, and flexibility in regard to missing data and cut-off scores. The implications of these findings are discussed.

Who goes and who stays: subject loss in a multicenter, longitudinal follow-up study.

Subject loss in a cohort of 645 infants and their families, enrolled in a multicenter clinical trial, was described. Medical/biologic, socioeconomic and social support (Environmental Quality Index [EQI] composite), and developmental outcome data were obtained. Dropout was evaluated by comparing infants who withdrew at any time throughout the study to those who remained, as well as at four different time periods between 40 weeks conceptional age and 36 months. A dropout profile was also developed. Overall dropout was predicted by EQI score, clinical center, gestational age category, and 18-month developmental status. In addition to these variables, the dropout profile included race, but not 18-month developmental status. Medical/biologic variables, environmental quality, and race and center were most influential in the first, second and third, and fourth time periods, respectively. The identification of factors which are associated with increased subject loss can help researchers project needed sample sizes in future studies.

Factors affecting neurobehavioral responses of preterm infants at term conceptional age.

A revised Prechtl Neurological Examination was administered to 510 singleton infants at term conceptional age to assess the influence of gestational age, race, and sex on neurobehavioral responses in a "typical" newborn population. The infants were born at 1 of 5 centers and were subjects of the NIH Collaborative Study on Antenatal Steroid Therapy. Of these babies, 392 were born at less than 37 weeks gestational age, whereas 118 infants were born at greater than or equal to 37 weeks. Of a total of 50 items of interest, 12 differed significantly based on race, 14 items differed in terms of gestational age at birth, and 5 items were influenced by the sex of the baby. These data suggest that singleton preterm infants born at differing gestational ages should not be considered to comprise a homogeneous sample. Further, caution should be exercised when considering babies of heterogeneous race groupings, and the babies' sex has minimal impact on neurobehavioral responses at term conceptional age.