A natural history model of Chlamydia Trachomatis infection: incidence and prevalence of preventable complications

The National Chlamydia Screening Programme has now been rolled out to all Primary Care Trusts in England. However, the extent of the health gain that can be achieved by early detection and treatment of chlamydial infection remains unclear. Estimates of the numbers of complications that could be prevented in two recent UK models were based on entirely different sources of data, and neither met the requirement that they should be both internally consistent and consistent with all the data. This project will scrutinise the full range of prospective and retrospective evidence, consider potential biases in it, and reach a consensus view on its interpretation. New statistical modelling techniques will be deployed to combine information from disparate sources, adjusting for bias and using all relevant studies and routine surveys, in order to make the most accurate possible estimates of the complications of chlamydia and the potential gains from screening. While oriented to CT policy in the UK, the research will introduce an approach and scientific results that could be applied in other industrialised countries, and to other public health conditions.

The effectiveness and cost-effectiveness of screening for Chlamydia trachomatis in the UK has been assessed in recent modeling studies. Cost-effectiveness is highly sensitive to assumptions regarding the frequency of major preventable complications, pelvic inflammatory disease, ectopic pregnancy, infertility, and neonatal conjunctivitis and pneumonia. There is however no consensus on these key parameters - or even on which evidence sources should be used to inform them. The proposed research aims to obtain a consensus interpretation of the existing evidence, and internally consistent and externally validated estimates of the incidence and prevalence of Chlamydia trachomatis infection and its complications. This will be based on systematic analysis of all available evidence identified under a strict protocol for inclusion, scrutiny of studies to identify potential sources of bias, and quantification of potential biases by epidemiologists with expertise in the field. Statistical modelling via Bayesian multi-parameter evidence synthesis will be used model the natural history of CT, incorporating all sources of data and uncertainty. The model will be thoroughly assessed for consistency with the data. The results of the model will be available for future cost-effectiveness analyses of Chlamydia trachomatis screening in the UK and elsewhere, and for dynamic disease transmission modelling.