Our study has demonstrated differences in chlamydia and HIV testing rates between the two study areas. Less than half of residents in the London-based study area (Brent) were registered with general practices that had tested for chlamydia and HIV, in contrast to over 90% of the non-London study area (Avon). Indeed, we found that the specialist GUM clinic in Brent conducted more chlamydia and HIV testing than all the Brent general practices combined. The historical emphasis in the UK on GUM clinics providing HIV testing services and care may explain why most testing in Brent is conducted in the GUM clinic. Greater GUM provision in London, along with a higher number of full time equivalent GPs per 1000 population in Avon (0.64 in Avon compared to 0.47 in Brent) might partly explain the more widespread testing in Avon, and the differences in general practice to GUM clinic ratios between Brent and Avon. More work is needed to understand and address these geographical differences.
Characteristics of GPs have previously been reported to be associated with levels of sexual health service provision. In our study, practices that had tested tended to have younger GPs than practices that had not tested, at least in Brent. Younger GPs may be more likely to have had education about STIs and HIV in their vocational training and consequently feel more comfortable discussing sexual health issues. This finding has important implications for the provision of effective sexual health services in primary care, and particularly the challenge of general practice participation in the NCSP, which has been identified as a key requirement for achieving the goal of controlling Chlamydia trachomatis infection . While an association between younger practitioner age and chlamydia testing has also been observed in a study in Australia, we did not reproduce that study's association between chlamydia diagnoses and female practitioners . Our findings were closer to an earlier UK study which suggested that the number of GPs per unit population was more important .
In both study areas, most chlamydia testing in general practice was carried out in female patients aged over 25 years, who are known to be at lower risk , while men received only a small proportion of chlamydia and HIV tests. These findings are consistent with a study using a large national UK primary care database, which concluded that chlamydia testing in general practice disproportionately targets women aged over 24 years and there are extremely low testing rates in men . Whilst more men who have sex with men attend GUM clinics than general practice, this would not be large enough to account for the differences in testing ratios in women compared to men. We did not, however, have data to examine this formally. The disparity is more likely to reflect the routine testing of male patients for chlamydia in GUM clinics. It was worrying that there was some evidence that areas of Brent with higher teenage conception rates were more likely to have general practices that had not taken either chlamydia or HIV tests, while more deprived areas were found to have practices that had never conducted any tests for HIV. It seems therefore that STI testing in general practice continues to be limited, highly variable and poorly targeted eight years after the publication of England's National Strategy for Sexual Health and HIV .
A major strength of our study was the comparison of data from both general practices and GUM clinics for two contrasting populations. While linkage of information about tests and test results to area level factors such as deprivation and teenage pregnancy enabled exploration of testing in relation to hypothesised population-level indicators of sexual ill-health. However, we were not able to explore individual-level demographics in this study such as patient's ethnicity or patient-level measures of deprivation. Another possible limitation of our study is that the Brent general practices included in this analysis were those that submit samples to the Northwick Park Laboratory and so exclude seven practices in south Brent that send samples to a different laboratory. Patterns of chlamydia and HIV testing in these practices may differ, but we anticipate that general practice:GUM clinic testing ratios are likely to be similar since patients resident in this area of Brent are also closer to another GUM clinic.
Future chlamydia testing patterns will be heavily influenced by the widening implementation of opportunistic chlamydia screening in England, which is targeted at sexually active women and men under 25 years  and was a core proposal of the National Strategy for Sexual Health and HIV . Whilst it has been recommended that chlamydia screening be integrated with other sexual health services, offering HIV testing at the same time as chlamydia testing in general practice might be challenging. Recently published HIV testing guidelines propose a strategy of active HIV testing in general practice, particularly at the time of GP registration in areas of high prevalence, in addition to targeting to "indicator diseases" (in which HIV is more common) in all populations . This strategy could be considered in Brent, which has a higher prevalence of HIV than Avon, and where a much higher proportion of GPs are not yet undertaking HIV testing at all. The development of interventions to support primary care practitioners in delivering STI and HIV testing should take into account the observation that rates of testing might be low in areas with high levels of need for sexual health care. We are completing further research to develop a web-based tool to support sexual health care for people presenting to general practices in Brent and Avon, which has been informed by the findings of this study.