Background The risk of Pelvic Inflammatory Disease (PID) and Ectopic Pregnancy (EP) from Chlamydia are crucial in estimating the cost-effectiveness of screening, but they remain poorly understood.

Methods We use evidence from RCTs of screening and controlled observational studies to estimate the risk of PID following Chlamydia and the probability PID would be prevented by annual testing. The studies are synthesised using a model that allows for the possibility that the rate of developing PID is higher in the period soon after infection. We examine the role of Chlamydia and PID in EP using prospective evidence from the Lund study, evidence on the incidence and cumulative incidence of PID and EP in England, and retrospective evidence from case control studies. We assess the consistency of the data under different sets of assumptions about the severity of undiagnosed and non-hospital referred PID.

Results If the risk of PID due to Chlamydia is constant over time then the probability that an untreated Chlamydia episode causes clinical PID is estimated to be about 15%, and there is approximately a 60% chance that annual testing would prevent an associated PID in a woman who becomes infected. If the PID rate is assumed to be higher for 1–3 months the respective figures are 16% and 50%. We estimate that between a third and a half of EPs are caused by PID. Of these, around a third are due to Chlamydia though estimates are highly uncertain. Our comparison of different data sources suggests that undiagnosed PID carries some risk of EP.

Conclusions Our findings support a public health strategy that (a) identifies women with Chlamydia as soon after infection as possible i.e. to get tested on change of sexual partner; (b) has a low threshold for diagnosing and treating women with pelvic pain or suspected PID.