Psychiatric vulnerability and the risk for unintended pregnancies, a systematic review and meta-analysis 37 Substance use disorders Pregnancy intention was assessed in 1455 women who used substances. UPs were often reported in this group of women (74–100%)37,40,42. Multi-drug use was reported in one study (Tabi et al.) as aside from opioid use, participants reported the (ab)use of cannabis, cocaine, benzodiazepines, methamphetamine, and alcohol40. Tenkku et al. assessed nicotine dependence, alcohol and drug abuse in 484 women37, of which 74% had UPs. Heil et al. found 86% UPs in 946 pregnant opioid addicted women42. Conduct disorders One study showed higher rates of (lifetime) abortions in women with a history of high CD symptoms at age 15, (≥7 problems based on DSM-III-R) compared to women with low CD symptoms at age 15. After adjusting for multiple social and psychological confounders, the associations between CD symptoms and abortions remained significant41. Eating disorders Assessment of pregnancy intention was performed amongst 927 women with eating disorders in two European studies32,33. In women with anorexia nervosa (AN), OR for UPs were higher than in women without anorexia nervosa, however in women with and without bulimia nervosa (BN), OR for UPs did not differ. Risk of bias of included studies Quality of the included studies is displayed in Table 2.2. The outcome UPs graded with the NIH tool28 resulted in a fair quality for nine out of eleven studies and poor quality in two out of eleven studies. Degree of author agreement was 84% between two reviewers (NS and NR), consensus was reached with a third reviewer (BB). Additional file 2 displays the grading per item in the NIH tool. Risk of bias was high due to cross-sectional analyses of cohort data. Solely one study assessed pregnancy intention in a prospective manner34, one other study assessed abortion in a prospective manner41. In most studies, time from exposure (psychiatric vulnerability) to outcome (UPs) was not measured and/or reported. In addition, UPs were not measured using validated tools. We found that 8 studies primarily focused on UPs or abortions, while three studies included pregnancy intention as secondary outcome or demographic feature35,39,40. Most studies considered relevant confounders, although small sample sizes limited ability to perform multiple regression analyses in some studies36,40. Most studies had a sample size of less than 600 women, while two studies had a larger sample size: Micali et al. included 1961 women and Heil et al. included 946 women33,42. A funnel plot (Figure 2.2) demonstrates the variety in sample sizes and effect sizes per study.
RkJQdWJsaXNoZXIy MTk4NDMw