Pregnant women with a documented psychiatric disorder or substance abuse problem have a significantly elevated risk of having poor birth outcomes, according to results of a population-based study conducted in California.1 Among infants born there in 1995, those whose mothers had a diagnosed psychiatric or substance-related disorder had roughly 2-4 times the odds of other babies of having a low or very low birth weight, and of being delivered preterm. As the analysts observe, the findings are important because psychiatric disorders are not uncommon among pregnant women, and earlier studies of their relationship to poor birth outcomes have examined symptoms or stress rather than documented diagnoses.

Using a data set that links statewide birth and infant death records with maternal and infant hospital discharge records, the analysts gathered information on more than 521,000 women who delivered in 1995 and on their liveborn, singleton infants. Data included women's demographic characteristics, use of prenatal care, birth outcomes, and diagnoses of psychiatric disorders (for example, mood, psychotic, eating, sleep, sexual or gender identity, and adjustment disorders) and substance-related disorders.

In all, fewer than 3% of women had a documented diagnosis--0.4% had a psychiatric diagnosis, 1.0% a diagnosis of substance abuse and 1.4% both. Those with any documented disorder were more likely than those with none to be black or white, to be covered by Medi-Cal (California's Medicaid program), to be single, to have had more than three prior deliveries and to have received inadequate prenatal care (as measured by a standard index). The vast majority of women with a diagnosis (92%) had their condition identified at the time they were hospitalized for delivery; 14% received their diagnosis while pregnant, and 6% had a disorder diagnosed both prenatally and at delivery.

Some 15-21% of women with a diagnosis delivered before 37 weeks' gestation (preterm), compared with 9% of those with no diagnosis. Similarly, whereas 5% of women with no diagnosis bore an infant who was low-birth-weight (less than 2,500 g), the proportion was 10-18% among those with a diagnosis; for very low birth weight (less than 1,500 g), the proportion was 1% for those with no documented disorder and 3% for women with a psychiatric or substance-related disorder or both.

The analysts used logistic regression to examine the risk of poor outcomes while controlling for the effects of marital status, ethnicity and adequacy of prenatal care. These calculations revealed that compared with women who had no diagnosis, those with a psychiatric disorder had twice the odds of bearing a low-birth-weight infant (odds ratio, 2.0), those with a substance-related disorder had almost four times the odds (3.7) and those with both types of diagnoses had three times the odds (3.0). Each category of diagnosis was associated with about a tripling of the risk of very low birth weight: Odds ratios were 2.8-3.0. Women with a psychiatric disorder had a 60% greater risk of preterm delivery than those with no documented disorder (odds ratio, 1.6), while women with substance-related or dual diagnoses had roughly doubled risks (2.4 and 2.3, respectively).

Additional analyses--one including a larger number of potentially confounding variables and one that was restricted to women who had not given birth before, to control for the possible confounding effect of a history of preterm delivery--produced essentially similar results. When the analyses were limited to women whose diagnoses had been made before delivery, the findings remained unchanged.

The analysts comment that the findings "underscore the importance of improved detection of psychiatric and substance use disorders" among pregnant women. Once such disorders are identified, they point out, "increased monitoring...could enhance timely interventions and improve birth outcomes." --D. Hollander


1. Kelly RH et al., Psychiatric and substance use disorders as risk factors for low birth weight and preterm delivery, Obstetrics & Gynecology, 2002, 100(2):297-304.