Self-collection of samples for group B streptococcus testing during pregnancy: a systematic review and meta-analysis

BackgroundSample self-collection for reproductive tract infection diagnosis has been found to offer greater convenience, privacy, autonomy, and expanded access to testing in non-pregnant adults. This review aimed to determine whether sample self-collection is as accurate as provider-collection for detection of group B streptococcus colonisation in pregnancy and whether a strategy of self-collection compared to provider-collection might improve maternal and neonatal health outcomes.MethodsWe searched CINAHL Plus, Medline, EMBASE, Maternity and Infant Care Database, Cochrane Central Register of Controlled Trials, and Cochrane Database of Systematic Reviews in June 2022. Eligible studies compared self-collected and provider-collected samples taken from the same participants or participants randomised to either self-collection or provider-collection for reproductive tract infection testing using the same test and testing method in pregnant individuals. We included trials and observational studies. Reviewers assessed risk of bias using the QUADAS-2 checklist and independently extracted data. Sensitivity and specificity for group B streptococcus colonisation of self-collected compared to provider-collected samples were pooled using a bivariate, random-effects, meta-analytic model. This review was registered with PROSPERO (CRD42023396573).ResultsThe search identified 5909 references, of which eleven diagnostic accuracy group B streptococcus studies were included (n = 3269 participants). No studies assessed the effects of self-collection in pregnancy on health outcomes. All studies had high or unclear risk of bias. Pooled sensitivities of self-collected samples for group B streptococcus detection were 82% (95% CI: 66-91%; I2 = 68.85%) in four trials (n = 1226) and 91% (95% CI: 83-96%; I2 = 37.38%) in seven non-randomised studies (n = 2043). Pooled specificities were 99% (95% CI: 98-99%; I2 = 12.08%) and 97% (95% CI: 94-99%; I2 = 72.50%), respectively.ConclusionsSelf-collected samples for group B streptococcus detection in pregnancy had high specificity compared to provider-collection, but lower sensitivity, particularly for included trials. Studies investigating the effect of self-collection on health outcomes, and further higher quality trials comparing accuracy of self-collection to provider-collection, are required.