• Chlamydia
• screening

It is important to know who takes part in screening programmes because this will help to understand and interpret information about their outcomes. The English National Chlamydia Screening Programme (NCSP) was introduced in the first areas in September 2002.1 Riha et al2 make an important contribution to the literature about screening programmes in general and chlamydia screening in particular. They showed that women and men participating in the NCSP have higher levels of sexual risk behaviours and that they are about three times as likely to be infected with Chlamydia trachomatis (chlamydia) infection than their counterparts in the general population, who took part in the UK National Survey of Sexual Attitudes and Lifestyles 2000 (Natsal-2).

Riha et al's study2 shows the importance of defining the denominator population when describing the occurrence of chlamydia infection. Prevalence is the total number of all individuals who have a disease at a particular time divided by the population at risk of having the disease.3 Riha et al can report the estimated population prevalence of chlamydia in sexually active 18–24-year-olds in the UK (about 3%) because they have data from Natsal-2, a random sample that is representative of the general population. They also report chlamydia positivity in the NCSP, which is the percentage of tests with a positive result. After excluding likely duplicate tests, 9–10% of NCSP tests among 18–24-year-olds in 2008 were positive.2 This measure could be described as the chlamydia prevalence in people tested in the NCSP. This group is, however, difficult to define as a population at risk; it includes sexually active adults under 25 years who attended any one of a range of healthcare settings or outreach events where a health professional offered a screening test and the offer was accepted. The intriguing question for understanding a screening programme is what are the processes at each stage of selection that result in the testing of those most likely to be infected with chlamydia?

The finding of higher levels of risk behaviour and higher levels of the target condition among participants in opportunistic screening might come as a surprise. The concern voiced about opportunistic cervical cancer screening, introduced in the UK in the 1960s, was that ‘the women most in need are those least likely to attend of their own accord … it [screening] concentrated upon the women least at risk’.4 Age was the main factor determining risk in that comparison, but rates of cervical cancer screening were also lowest in socioeconomically deprived areas. For chlamydia screening in England, coverage early in the programme has been reported to be higher in deprived areas5 and among those from black minority ethnic groups.2 Given that most adults eligible for chlamydia screening do use primary and community-based healthcare, but only a minority receive a screening test, the decisions of providers must be the main factor that results in differential coverage and selection of infected people. This pattern of selective screening also occurs in the USA, where chlamydia tests are more likely to be done in black and Hispanic women than white women and among those with public insurance compared with those with private insurance.6

Riha et al2 argue that the effectiveness of the NCSP might be higher than is currently assumed because higher risk individuals are being reached. This hypothesis certainly deserves further investigation, but it means that the NCSP needs to collect data about the uptake and outcomes of regular (annual) screening and repeated testing after treatment. Repeated infections after screening and treatment are very common.7–9 This suggests that, even if individuals at high risk are more likely to be screened, the high levels of repeated infection because of lack of treatment of existing sex partners, exposure to infected new sex partners or treatment failure could counterbalance the effects of screening.10

Riha et al's study2 also highlights the need for better completion of the NCSP data collection form. The poor response to the question about condom use at the last sex act (5%) made it difficult to analyse and interpret the data. Improvements in NCSP data collection about individual items and about repeated screening should be able to be introduced in time to repeat the analysis and update the comparison when data from the ongoing Natsal-3 become available and will help to improve mathematical modelling studies.

We should spend more effort to examine the reasons for and impact of differential selection both within and between screening programmes. As Riha et al point out,2 most mathematical modelling studies investigating the effects of chlamydia control assume that screening uptake is random. Their results can now be used to parameterise models that explore how differential screening uptake affects the impacts of one-off and regular screening, partner notification and repeat testing after treatment in reducing chlamydia prevalence. Additional studies using quantitative and qualitative research methods are also needed to determine the factors and processes that influence patient and provider decisions about healthcare seeking and the offering and acceptance of screening tests. Finally, Riha et al's study was carried out during a period when NCSP coverage was low (9% of all 16–24-year-olds in England). It will be interesting and important to see whether differential selection of those at high risk of chlamydia persists at higher levels of coverage.