Cervical screening coverage has been falling steadily for a number of years across the whole of the UK. Some 20% of eligible women in the UK have not been screened in the past 5 years: 25% in London and 32% in Hammersmith and Fulham PCT. Response to invitation is even worse: nationally, only 65% of invitations result in screening.
It is hypothesised that having to have a speculum examination is a major obstacle to screening for many women. It is certainly a major component of the cost of screening. Smear taking is said to be about a half of the cost of cervical screening in the UK.Our study and a systematic review have shown that, HPV testing of self-samples had a similar sensitivity and specificity to conventional screening (Szarewski et al 2007, Petignat et al 2007).We have shown that women find cervical self-sampling more acceptable than standard screening (Waller et al 2006, Szarewski et al 2007). A focus group suggested that women from ethnic minorities may be more willing to carryout self-sampling than attend for a smear.
Although cervical screening using cytology has been extremely successfully in reducing the incidence of cervical cancer in many countries, it is not perfect. UK data suggest that approximately 50% of cervical cancer in women aged 20-64 is occurring in women who have been screened following programme guidelines (Sasieni et al 1996 and unpublished audit data). If we are to impact on the cancers occurring in well screened women, we need a more sensitive screening test. There are some who believe that liquid based cytology (LBC) fills this need, but the data supporting the superior sensitivity of LBC for the detection of CIN2+ are far from convincing. Additionally, cytology is currently dependent on human reading and a lapse in concentration or error in judgement can lead to a false negative result. Partly for this reason, cytology is relatively inefficient in that it needs to be repeated three-yearly up to the age of 50. By contrast HPV testing can be automated and could probably be offered at intervals of five or even six years. All European and North American studies directly comparing HPV testing to cytology for the detection of high-grade CIN show that HPV testing is more sensitive, but less specific, than cytology (Cuzick et al 2006). Eventually, HPV vaccination will do away with the need for cervical screening, but since it is unlikely to be offered to women born prior to 1990, improvements to screening could potentially prevent some 40,000 cancers in women born too early to benefit from vaccination.
With this background, we wish to continue to study HPV testing in primary cervical screening. A number of issues need addressing: How should one deal with the poor specificity of HPV testing? Will HPV testing have the anticipated impact on cervical cancer?
Cervical cancer remains an important public health problem in many developing countries, where it is the commonest female cancer and a major cause of death. Cervical cancer is caused by HPV infection and HPV types 16 and 18 are responsible for about 70% of cervical cancer worldwide. Two companies have developed prophylactic vaccines against these two viruses. Both vaccines are based on viral like particles and require three injections (at 0, 1 or 2, and 6 months). Trials show that both are extremely effective at preventing persistent type-specific HPV infection and related high-grade disease and have the potential to substantially reduce cervical (and other HPV-related) cancer. Neither vaccine offers protection to women who have been previously infected.
Since 1992, we have been studying the extent to which the NHS Cervical Screening Programme (NHSCSP) is preventing cervical cancer. Audit not only allows monitoring of the programmes' effectiveness, it identifies areas of good practice and areas where improvement needs to be made. This project has collected data on the screening histories of women with cervical cancer and a sample of controls. Information published from these audits has been valuable from both a scientific and a policy perspective.
Although, we have a project grant to evaluate the NHSCSP, that does not cover all of our activities in this area. Through the programme, we will help to coordinate these activities, develop new methodologies for evaluating service screening, set up small additional studies and collaborate with groups in other countries undertaking similar projects.
With nearly 4000 cases and 8000 controls already collected, we have a large database for addressing a number of questions. These include more in depth analyses of the effects of screening at different ages, on different histological types of cancer and on advanced cervical cancer. By combining the effects of screening on the incidence of cancer of different stages and stage-specific survival, we will also estimate the reduction in mortality associated with regular screening. Other analyses will look at the association between screening uptake in (controls) aged 25-29 and their previous involvement in the screening programme, and the duration of protection offered by screening in well-screened women aged 50-64. We will also calculate attributable-risks so as to study the impact of screening on cervical cancer incidence and mortality in the UK.