Close

Changes in chlamydia prevalence and duration of infection estimated from testing and diagnosis rates in England: a model-based analysis using surveillance data, 2000-15.

Abstract

The data included numbers of tests and diagnoses in men and women aged 15-19 years and 20-24 years in England each year from 2000 to 2015. We estimated reductions in prevalence and average infection duration in both sexes once screening was fully implemented. From 2008 to 2010, estimated posterior median prevalence reductions in people aged 15-24 years were 0·68 percentage points (95% credible interval 0·26-1·40; minimum) and 0·66 percentage points (0·25-1·37; maximum) for men and 0·77 percentage points (0·45-1·27) for women (minimum and maximum estimates were the same for women). Over the same time period, mean duration of infection reduced by 75 days (95% credible interval 17-255; minimum) and 74 days (95% credible interval 17-247; maximum) in men and 30 days (22-40) in women. Since 2010, some of the progress made by the NCSP has been reversed, alongside a reduction in testing.

National Institute for Health Research, Medical Research Council.

Chlamydia screening programmes have been implemented in several countries, but the effects of screening on incidence, prevalence, and reproductive sequelae remain unclear. In England, despite increases in testing with the rollout of the National Chlamydia Screening Programme (NCSP; 2003-08), prevalence estimated in 10-yearly population-based surveys was similar before (1999-2001) and after (2010-12) the programme. However, the precision of these previous estimates was limited by the low numbers of infections. We aimed to establish annual, rather than 10-yearly, estimates of chlamydia prevalence and infection duration.

In this model-based analysis, we used previously published minimum and maximum estimates and Public Health England data for chlamydia test coverage and diagnoses in men and women aged 15-24 years in England, before, during, and after the scale-up of national chlamydia screening. We used a mechanistic model, which accounted for symptomatic chlamydia testing and asymptomatic screening, to estimate changes in prevalence and average duration of infections for each year. We describe estimates derived from the maximum and minimum numbers of tests and diagnoses as maximum and minimum estimates, regardless of their relative magnitude.

Our analysis provides the first evidence for a reduction in chlamydia prevalence in England concurrent with large-scale population testing. It also shows a consistent decline in the average duration of infections, which is a measure of screening effectiveness that is unaffected by behavioural changes.

MIDAS Network Members

Citation: