The article by Cleary et al entitled “Methadone and perinatal outcomes: a retrospective cohort study” published in the American Journal of Obstetrics and Gynecology in February 2011 provides retrospective cohort data for consideration regarding methadone’s risk:benefit ratio. However, several study shortcomings suggest inferential flaws and limit confidence in their conclusions. First, the methadone-exposed group was categorized based on “use at delivery.” This simplistic classification ignores both the positive impact of prepregnancy methadone stabilization (eg, fetal protection from maternal-addiction behaviors) and the negative impact of late-pregnancy entry into methadone treatment (eg, fetal harm from illicit drug exposure during organogenesis and from repeated episodes of withdrawal) on maternal/fetal/neonatal outcomes. Not accounting for early vs late treatment entry yields uninformative congenital abnormality rate data. Further, treatment elements significantly associated with maternal/fetal/neonatal well-being, such as behavioral counseling, case management, and obstetric care, are not provided. Not controlling for adequacy of opioid dependence treatment and/or aspects of care that affect fetal well-being undermines the data’s clinical utility. Second, comparing “methadone-exposed” to “unexposed” women without adequate control of important concomitant factors (eg, gestational illicit drug and alcohol use, and psychiatric comorbidity) that are known to be independently related to poor maternal/fetal/neonatal outcomes likely produced biased estimates of between-group differences and inference errors. The teratogenic potential of maternal medications must be considered in the context of the fetal environment, developmental state susceptibility, genetic predisposition, and fetal programming. To associate teratogenic propensity to a medication while ignoring the myriad factors that independently impact fetal development in this high-risk patient group is inaccurate and potentially harmful. Third, comparing the methadone-exposed and nonmethadone-exposed groups is further flawed, because methadone exposure is confounded with publicly funded obstetrical care rates (99.8% vs 74.3%, respectively). (See Table 2 for statistical problems associated with this comparison [eg, odds ratio magnitude and its associated SE].) Concluding that “Methadone exposure is associated with an increased risk of adverse perinatal outcomes, even when known adverse sociodemographic factors have been accounted for,” is likely to be inaccurate, given these serious limitations. Thus, these results should not be viewed as providing “population-based estimates of the risk of adverse perinatal outcomes” associated with gestational methadone exposure. Moreover, unquestioning acceptance of these results may unduly impede women’s access to, and obstetrician’s ability to provide, appropriate treatment for this high-risk obstetric population.