Using Negative Predictive Values: In Reply

Abstract

In Reply.—We appreciate Dr Camp's comments and her emphasis of the importance of the high negative predictive value of the Bayley Mental Developmental Index (MDI). We agree that a normal MDI of >85 at 20 months' corrected age in an extremely low birth weight (ELBW) child is highly predictive of having a normal IQ of >85 at 8 years of age as measured by the Mental Processing Composite,1 and that this may be a reassuring finding for both parents and caregivers.Our reason for emphasizing the poor positive predictive value of the Bayley MDI was because published reports of developmental outcomes at 18 to 20 months' corrected age are used to counsel parents of extremely immature and low birth weight infants about the potential future development of their children and to make critical decisions concerning the provision or withholding of intensive care in the delivery room. It is important that both caregivers and parents be aware of the limitations of outcomes reported at 18 to 20 months' corrected age when making critical “life-and-death” decisions in the delivery room.Concerning the 5 children in the nonimpaired group who were “false-negatives” (ie, whose MDI at 20 months dropped from ≥85 to an IQ below 85 at 8 years), none of these 5 children were diagnosed with a neurologic abnormality at 8 years of age, and we know of no insult occurring between 20 months and 8 years that may explain their drop in IQ. Furthermore, in none of the 5 children did the IQ drop to the subnormal range (<70). Their mean MDI at 20 months was 94.6 (range: 86–108) and mean IQ at 8 years was 80.8 (range: 80–83). The drop in IQ in these 5 children may reflect measurement error, although past studies of IQ changes across childhood have revealed variations of this magnitude in a minority of children.2