Background Pelvic inflammatory disease (PID) results from the ascending spread of microorganisms, including Chlamydia trachomatis, to the upper genital tract. The timing of ascending infection is unknown. Screening can prevent PID either by identifying and treating infections before they progress (direct effect) and/or reducing chlamydia transmission (indirect effect). We did this study to examine the contributions of direct and indirect effects of a screening intervention, using different assumptions about the timing of progression from chlamydia infection to PID.

Methods We developed a compartmental model of chlamydia transmission in a heterosexual population of 16–25 year olds with two sexual activity classes. The model explicitly incorporates the progression from chlamydia to clinical PID. Behavioural parameters are informed by a British population-based. We studied the effects of chlamydia screening introduced at low levels but with coverage increasing to 40% after ten years. We estimated the numbers of PID cases prevented and the proportions prevented by direct and indirect effects.

Results At baseline, the cumulative probability of developing PID by age 25 years was 3.1%. After five years, screening prevented a total of 187 PID cases per 100,000 women. Most PID cases were initially prevented by interruption of progression to PID (direct effect). The indirect effect produced a small net increase in PID cases early on, which was outweighed by the effect of reduced chlamydia transmission after 2.2 years. The later that progression to PID occurs, the greater the contribution of the direct effect.

Conclusion The ratio of direct to indirect effects depends on the timing of progression from chlamydia infection to PID. Mathematical modelling has helped to understand the mechanisms of chlamydia screening programmes by showing that there are separate roles for direct and indirect PID prevention and potential harms of screening, which could not have been observed by empirical studies.