Introduction Mathematical models are frequently used to assess the impact of control interventions for Chlamydia trachomatis and other sexually transmitted infections (STIs). Modelling approaches that stratify the population by the number of sex partners often assume the transmission risk per partner to be constant. However, sexual behaviour data suggests that people with many partners share less sex acts per partner than people with fewer partners. This should lower the risk of transmission per partner for highly sexually active individuals and could have important epidemiological consequences for STI transmission.

Methods We devised a new epidemiological model that we fitted to chlamydia prevalence data from Natsal-2 and CSF, two population-based probability sample surveys of sexual behaviour in Britain and France.

Results Compared to a standard model where the transmission risk per partner is constant, a model with realistic numbers of sex acts per partner provided a better fit to the data. The improved model provided evidence for strong assortative mixing (ε = 0.83; 95% CI 0.46–0.96) among individuals with different numbers of sex partners. The basic reproduction number (R₀) exceeds the threshold of one for all individuals with one or more new heterosexual partners in the last year, and saturates around three for higher number of partners.

Conclusion Our results suggest that all chlamydia infected individuals with one or more new heterosexual partners per year contribute significantly to ongoing transmission, underlining that control interventions should be aimed towards all sexually active young adults.

Disclosure of interest statement CLA received funding through an Ambizione grant from the Swiss National Science Foundation (project 136737). SA is funded by an ATIP-Avenir from CNRS and IN- SERM. SA and MC are also supported by the CNRS and the IRD. We declare no competing interests.