Chapter 2 38 Table 2.2 - Quality assessment of included studies. Outcome Risk of bias assessment NIH tool Inconsistency Indirectness Imprecision Publication bias Overall assessment Unintended pregnancy Studies had poor to fair quality assessment I2 tests showed significant heterogeneity between studies in the random effects model Although the outcome abortion is used as a proxy for UPs in two of the included studies, the outcome of interest (UPs) is investigated directly in all the other studies As displayed in the random effects model (Figure 2.3), the absolute prevalence of UPs amongst women with psychiatric vulnerability is higher than national averages even at the lower border of the 95% CI Most studies were sponsored by national health programs, solely one study was sponsored by industry38, namely Pfizer Health Research fund Observational studies start at low certainty, with all 5 domains this does not change (downgrade for inconsistency, upgrade for large effect)
RkJQdWJsaXNoZXIy MTk4NDMw