Nearly half of pregnancies in the United States (and worldwide) are unintended. The health effects of unintended pregnancy have long been based on observing trends in maternal and fetal health outcomes over time when pregnant patients are asked to recall whether they desired a pregnancy at the time of conception. This is because randomized trials evaluating interventions in pregnant individuals are largely unethicalSo, when it comes to pregnancy-related issues, we are almost always left with observational data and the associations that can be made from them as our best available evidence. One way we can inch closer to identifying causal relationships with hard-to-study conditions, however, is with high-quality systematic reviews and meta-analyses.

A recent meta-analysis in the November 1 issue of JAMA, which was devoted to reproductive health, is a good example of how a systematic review can elevate the level of existing evidence. Two factors play key roles in making systematic reviews of observational data high-quality: a systematic search with proper inclusion criteria, and a transparent assessment of study quality using a validated scoring system. This systematic review did a good job of appraising which studies were good enough to be combined in a meta-analysis (remember, garbage in = garbage out). Data from 36 observational studies compared outcomes for unintended (including mistimed) vs. intended pregnancies in 524,522 pregnant persons living in high-resource countries. The meta-analysis found that unintentional pregnancy was associated with increased risks of preterm birth, low birth weight, prenatal and postnatal depression, and interpersonal violence, the latter having the strongest association with an odds ratio of 2.22. While many of the included studies accounted for history of depression, most had limited sociodemographic data and variable adjustments for confounders.

Whether the fact that a pregnancy is unintended is the direct link to an adverse maternal or fetal outcome, or rather, that there are shared associations like socioeconomic status that mediate this relationship, cannot be distinguished from observational data. However, this high-quality meta-analysis provides us with enough certainty to inform clinical practice and health policy such that we can turn our attention to preventing both unintended pregnancies and the adverse health outcomes that follow.

Practice Point: Whatever the factors contributing to an unintended pregnancy, there are increased risks of adverse maternal and fetal outcomes that may be preventable with awareness and appropriate action.

EBM Pearl: When randomized trials are not possible, a systematic review and meta-analysis of high-quality observational data can provide reasonable guidance for clinical care.

For more information, see the topic Risk Factors for Preterm Labor and Premature Birth in DynaMed.