Background The number of tests performed as part of the National Chlamydia Screening Programme (NCSP) has increased since the start of the programme in 2003 and the positivity has declined. We explored the extent to which available data can be used to adjust for changes in who is being screened in order to estimate any changes in the population prevalence up to 2010.

Methods Analyses of positivity trends were conducted using available data for opportunistic asymptomatic tests (screens) from the NCSP national dataset for 2005 to 2010 from areas that implemented screening throughout this time period. Age, sex, ethnicity, sexual behaviour, regional and venue of screen weights for the English population of 15–24-year-olds were derived (from national sources where available) and applied to the dataset.

Results From 2005 to 2010 there was an increase in screens among men. There were no major changes in characteristics known to be associated with infection (year of age, sexual behavioural variables). Available data on sexual behavioural variables and ethnicity decreased over time. There were some changes in venue use over time. Weighting for 5-year age group, sex, <2 sexual partners in past 12 months, ethnicity and region lowered positivity in each year but slightly increased the decline in positivity from an average decline of 13% per year (from 11% in 2005 to 6% in 2010) to an average of 14% per year (from 10% to 4%). Additional standardisation by screening venue did not reduce the overall observed decline in positivity during this period. Differences in positivity between venues remained, but were slightly reduced, after weighting for differences in known characteristics of screened clients.

Conclusions The observed decline in positivity over time among screens was not accounted for by weighting for known characteristics of those screened or changes in testing venues. Together with the consistency of declining positivity in all sub-categories this suggests that a true decline in population prevalence may have occurred. Further analyses of the potential effects of data limitations and using regression techniques with additional variables (eg, deprivation) are in progress to better understand the relationship between screen positivity and population prevalence at different levels of screening uptake in England.