Age difference between heterosexual partners in Britain: Implications for the spread of Chlamydia trachomatis

Evaluating distributional regression strategies for modelling self-reported sexual age-mixing

The age dynamics of sexual partnership formation determine patterns of sexually transmitted disease transmission and have long been a focus of researchers studying human immunodeficiency virus. Data on self-reported sexual partner age distributions are available from a variety of sources. We sought to explore statistical models that accurately predict the distribution of sexual partner ages over age and sex. We identified which probability distributions and outcome specifications best captured variation in partner age and quantified the benefits of modelling these data using distributional regression. We found that distributional regression with a sinh-arcsinh distribution replicated observed partner age distributions most accurately across three geographically diverse data sets. This framework can be extended with well-known hierarchical modelling tools and can help improve estimates of sexual age-mixing dynamics.

Discrepancies between observed data and predictions from mathematical modelling of the impact of screening interventions on Chlamydia trachomatis prevalence

Mathematical modelling studies of C. trachomatis transmission predict that interventions to screen and treat chlamydia infection will reduce prevalence to a greater degree than that observed in empirical population-based studies. We investigated two factors that might explain this discrepancy: partial immunity after natural infection clearance and differential screening coverage according to infection risk. We used four variants of a compartmental model for heterosexual C. trachomatis transmission, parameterized using data from England about sexual behaviour, C. trachomatis testing, diagnosis and prevalence, and Markov Chain Monte Carlo methods for statistical inference. In our baseline scenario, a model in which partial immunity follows natural infection clearance and the proportion of tests done in chlamydia-infected people decreases over time fitted the data best. The model predicts that partial immunity reduced susceptibility to reinfection by 68% (95% Bayesian credible interval 46-87%). The estimated screening rate was 4.3 (2.2-6.6) times higher for infected than for uninfected women in 2000, decreasing to 2.1 (1.4-2.9) in 2011. Despite incorporation of these factors, the model still predicted a marked decline in C. trachomatis prevalence. To reduce the gap between modelling and data, advances are needed in knowledge about factors influencing the coverage of chlamydia screening, the immunology of C. trachomatis and changes in C. trachomatis prevalence at the population level.