NICU care is expensive, and outcomes for survivors may be less than might be hoped for. Permanent neurologic disability is a serious burden, and surrogate decision-makers may wish to consider this possibility in the benefit/burden calculations they face anew each day. We wondered how accurate medical caretakers were in their daily predictions of residual neurologic disability for NICU patients.

Methods: We surveyed RN and MD caretakers of ventilated newborns in our NICU for 48 weeks, asking on each day whether they thought the infants in their care would 1) survive or die before discharge; and2) if they survived, whether severe permanent neurologic disability would be present. We identified three groups of NICU survivors: those who at some time during their NICU stay had received predictions of a) death; b) severe residual neurologic morbidity; or c) neither death nor severe morbidity. We performed modified Denver Developmental Scale (DDS) assessments of neurologic status on infants in each population.

Results: 13 infants were predicted to have severely morbid outcomes -- 6 (46%) of these were significantly delayed on the DDS at 12-18 months adjusted age. 27 infants were predicted NOT to have severely morbid outcomes -- only 1 (4%) was significantly delayed (p < 0.001 vs infants with predicted severe morbidity). 8 infants were predicted to die during their NICU course, but survived to discharge -- 3 (37%) were delayed at 12-18 months. 40 infants were never predicted to die -- 6 (15%) were delayed. The sensitivity of predictions of severe neurologic morbidity was very high, (6/7= 86%), but the positive predictive value of predictions of severe delay was not (6/13 = 46%).

Conclusions: 1) The large majority of NICU survivors predicted neither to die nor be severely delayed are apparently neurologically normal, at least through their first year. 2) The large majority of NICU survivors with residual neurologic morbidity are accurately identified during their NICU stay. 3) Nevertheless, predictions of severe neurologic morbidity for infants surviving NICU care are incorrect approximately half the time.