Prenatal crack cocaine exposure causes permanent, severe brain damage in children.

In September 1985, Ira J. Chasnoff, a pediatrician at Northwestern University Medical School, published a brief paper in the New England Journal of Medicine describing his observations of 23 cocaine-using women during pregnancy. The study was small and carefully hedged. Chasnoff and his colleagues noted signs of fetal distress in the cocaine-exposed group, including lower birth weight and signs of neurobehavioral disturbance in the first days of life, but were explicit about the preliminary nature of their findings. The paper called for further research. What it sparked was not further research.

Within months, the study had been transmuted by journalism and public health communication into something its authors had not written. Newsweek and Time ran cover stories. Television news programs showed footage of infants in neonatal intensive care units, trembling and difficult to console. The phrase 'crack baby' entered the language in 1986, carrying a specific and vivid image: a child permanently damaged by its mother's choices before birth, entering the world already broken and unlikely to recover. By 1987 and 1988, when the crack cocaine epidemic was at its most visible, the narrative had been incorporated into DARE curricula, anti-drug public service announcements, and school health instruction across the country. Pediatricians began forecasting a generation of developmentally disabled children that would overwhelm the educational system by the mid-1990s.

The scientific picture was more complicated, and it was becoming more complicated with every new study published. Researchers attempting to replicate Chasnoff's early findings ran immediately into a methodological problem: cocaine use during pregnancy was almost never isolated. Women who used crack cocaine while pregnant also, in high proportions, used alcohol, tobacco, and other drugs; lived in conditions of severe poverty; received limited or no prenatal care; and experienced levels of chronic stress that had their own documented developmental consequences for children. Disentangling the effect of cocaine exposure from the effect of everything else that accompanied it was difficult, and in many of the early studies, it was not adequately attempted.

A 1990 survey of research in neonatology journals found that studies reporting harmful effects of prenatal cocaine exposure were significantly more likely to be published than studies finding no effect, a pattern that suggested the published literature was providing a systematically distorted picture of the evidence. Researchers who tried to publish null findings encountered resistance from journals that considered them uninteresting.

The landmark reassessment came in March 2001, when Deborah A. Frank and colleagues at Boston University published a systematic review in JAMA examining 36 studies of early childhood outcomes following prenatal cocaine exposure. The review found that among children up to age six, there was no convincing evidence that prenatal cocaine exposure was associated with developmental toxicity that differed in severity, scope, or kind from the effects of other risk factors. The deficits that had been attributed to cocaine were, when researchers controlled for poverty, prenatal alcohol and tobacco exposure, and the quality of the child's postnatal environment, no longer specifically associated with cocaine at all.

By then, Chasnoff himself, whose early paper had helped ignite the panic, had publicly lamented the trajectory of the story. In interviews through the 1990s, he noted that the media had taken a preliminary study and constructed a certainty that the data did not support, and that children labeled 'crack babies' were entering schools and foster care systems with expectations of failure written into their records before anyone had assessed what they could actually do.

The children born into the crack epidemic of the late 1980s were, in many cases, born into poverty, instability, and environments that posed genuine developmental risks. The scientific consensus that emerged by the early 2000s was not that these risks were imaginary. It was that cocaine exposure was not their primary source. Poverty, it turned out, was at least as reliable a predictor of developmental outcome as any prenatal substance exposure. That was not the message that had been taught.