Point-of-care testing versus standard practice for chlamydia: a new approach to assessing the public health effect of rapid testing and treatment at local level

Background Chlamydia trachomatis is the most commonly diagnosed bacterial sexually transmitted infection in Britain. Present standards specify treatment within 14 days of testing positive; point-of-care testing (POCT) can eliminate this delay and potentially reduce loss to follow-up; its greater convenience might increase testing. 90-min nucleic acid amplifi cation tests are the best available POCTs for chlamydia, with alternatives under development. However, cost-eff ectiveness depends on cost-per-test, sensitivity and specifi city, and the eff ect of POCT on transmission. Methods We developed a user-friendly web-based method, based on a transmission-dynamic model for chlamydia, to assess the epidemiological impact and cost-eff ectiveness of introducing POCT in diff erent local settings. The model uses behavioural and prevalence data from the National Survey of Sexual Attitudes and Lifestyles, and Public Health England surveillance data; these data inform on local-level variation, which is represented by sampling parameter values from within their ranges of uncertainty and selecting parameter sets that reproduce local coverage and diagnosis rates. The user can select diff erent local settings, and vary sensitivity and specifi city for the tests, specify costs (fi xed and unit costs, including staff time), and then assess the eff ect of introducing POCT in diff erent clinical services by comparison with a situation with no POCT. In the model, presumptive treatment is represented, which we expect to be reduced with the introduction of POCT because test results would be rapidly available to guide treatment. Findings Changes in numbers of infections and diagnoses occurring under diff erent scenarios (including conventional testing) were estimated, with uncertainty ranges, allowing calculation of total costs, and cost per infection (and serious sequelae) averted, while accommodating the considerable variation in chlamydia testing coverage, positivity, and diagnosis rates. Potential changes in sexual behaviour between test and treatment could determine the relative contribution of increased treatment rates and reduced treatment delay to the reduction in prevalence as a consequence of POCT. Interpretation The eff ect of POCT was dependent on both the test performance characteristics and the assumptions about the implementation of the test across local services. Exploration of many uncertainties surrounding chlamydia epidemiology and screening implementation is possible with this model. This method can complement local and national knowledge, and contribute to local-level management of chlamydia infection.