Cancer Prevention in Primary Care: Screening for cervical cancer

Cervical cancer: current facts

In England and Wales during 1988, 4467 new cases of invasive carcinoma of the cervix were registered (4940 in the United Kingdom as a whole), making it the eighth commonest cancer in women. Although only 15.5% of cases occur in women under 35, it is the most common cancer in this age group, accounting for 25% of all new cancers. Since the early 1970s there has been a significant increase in the incidence of both carcinoma in situ and invasive cancer in women under 45, particularly in those aged 25- 34, but this may partly reflect detection by screening.

In England and Wales, 18753 women were registered in 1988 as having carcinoma in situ (including grade III cervical intraepithelial neoplasia). Most (85.5%) of these cases were registered in women under 45 (although less than 20% of deaths occur in this age group) (table I). This reflects the higher prevalence of screening among young women before 1988.

In 1992, 1647 women died of cervical cancer in England and Wales (1860 in the United Kingdom as a whole), 95% of them aged 35 and over (table I). Over the past 20 years the mortality in women aged 45 and over has fallen noticeably (fig 1). Moreover, the rise in deaths in women aged 40- 44, observed in the late 1970s and 1980s, has now reversed. For women under 40 the significant increase in mortality observed in the 1970s and 1980s now seems to be stabilising. This reversal of earlier trends may be due to an increase in screening over the past decade.

FIG 1

Age specific death rates for cervical cancer in England and Wales, 1950-904

• Download figure
• Open in new tab
• Download powerpoint

The five year relative survival rate for stage I cervical cancer is 79%, reducing to 7% for stage IV disease. Survival for precancerous lesions is almost 100%, which provides the incentive for screening.

Regional variation in the incidence of and mortality from cervical cancer is considerable in the United Kingdom. The highest risks for cervical cancer are generally in the north of the country and in Wales.

In general, women of low socioeconomic status have higher rates of cervical cancer, with a threefold difference between social classes V and I.

Evidence of effectiveness of screening

Cervical screening has been shown to be effective in several countries, although not by means of randomised controlled trials. In the Nordic countries, nearly complete coverage of the target population by organised cervical screening programmes in Iceland, Finland, Sweden, and parts of Denmark were soon followed by sharp falls in both incidence and mortality (table II, fig 2). Reductions in total incidence and mortality were similar in Iceland and Finland (despite the different screening intervals used) but lower in Sweden (where screening was targeted at a narrower age range), indicating that, in organised screening, achieving a wide coverage of ages is a more important determinant of risk reduction than the frequency of screening.

FIG 2

Trends in annual age adjusted incidence rates of invasive carcinoma of cervix in Nordic countries5

• Download figure
• Open in new tab
• Download powerpoint

Cervical screening in the United Kingdom

A screening programme has been in operation in Britain since 1964. With the exception of a study in the north east of Scotland, screening in the United Kingdom has until recently been largely ineffective. There has been no reduction in the overall incidence of invasive cervical cancer, but a small reduction in mortality is just becoming apparent. However, any observed increases in incidence and mortality might have been much greater in the absence of screening.

Organisation of cervical screening

The reasons for the past failure of the national screening programme in the United Kingdom have been extensively examined. It is generally agreed that the problem has been not so much one of money or skill but of organisation, accountability, and commitment. The current dilemma is not whether screening should or should not be performed but how a service can be organised to greatest effect.

Box 1 shows the factors needed to improve the effectiveness of cervical screening. The target in the Health of the Nation is to reduce the incidence of invasive cervical cancer by at least 20% by the year 2000, from 15 per 100 000 population to no more than 12 per 100 000. If uniform acceptance rates of 80% can be achieved, a reduction in mortality of 65% to 70% can then be expected in the long term.

The introduction in 1988 of a systematic call-recall system and the setting up of an NHS cervical screening programme national coordinating network have brought a greater sense of coherence. A more uniform approach is being formulated for deciding who should be screened and how often, for classifying smears according to the degree of nuclear abnormality, for standardising terminology, for interpreting results and following up abnormalities, and for evaluating the effectiveness of the programme. Steps being taken locally by health authorities, the appointment of a national coordinator, the development of a quality assurance programme, and measures to improve the training of the relevant health professionals should all help to improve the effectiveness of the programme.

Improving population coverage: impact of targets

To be most effective screening must concentrate on a wide age range, with every effort being made to reach those at higher risk of developing invasive cancer - namely, older women of lower social classes. Before 1988 at least two thirds of patients with invasive cancer had never been screened at all and in women over the age of 40 (among whom over 70% of cases of invasive disease occur) the proportion who had never been screened was over 90%. In Aberdeen, where a large percentage of the population have had a smear taken, only 2% of women with invasive cervical cancer had had a negative result on smear testing within the preceeding five years, while 90% had never had a smear tested.

Primary care can exert a major influence on coverage. The 1990 general practitioner contract set targets on which payment for cervical screening depends. Payments are triggered on reaching 50% or 80% coverage of the target population over the preceding 5.5 years, with a differential of 3:1 in favour of the latter. This has led to a considerable increase in screening activity. Between April 1990 and October 1993 the percentage of general practitioners reaching the 80% target increased from 53% to 83%, while the percentage achieving neither target decreased over the same period from 15% to 3%. Coverage of the target population between 1989-90 and 1992-3 increased from 61% to 83%. The range between regional health authorities in England in 1992-3 was 70.3-88.8%, with lowest coverage achieved in the Thames regions. Nine district health authorities in inner London achieved a coverage of less than 60%. There is a strong indication that cervical screening is now beginning to reach those most at risk. Coverage in the age groups 55-59 and 60-64 improved between 1988-9 and 1992-3 by 128% and 225% respectively (from 37% and 25% to 84.3% and 81.3% respectively). At least one survey in 1992 has shown that the social class differential in screening that was apparent in an earlier survey in 1988 has almost disappeared. However, differential uptake is still likely to be a problem in inner city areas, where there is a higher proportion of women in social classes IV and V, a higher population mobility, inaccurate registers, and a lower overall coverage than that achieved in other parts of the country.

The setting of targets has clearly provided a great stimulus to many practices. It remains a matter of concern that it is a disincentive to those for whom any target or the higher target seems impossible to achieve and has resulted in very little interest in either providing or improving the cervical screening programme in these practices.

Cervical screening in primary care: organisational issues

Primary care is central to the overall success of the screening programme. General practitioners are in a unique position to invite women for a smear test, to take smears, to ensure abnormal results are followed up, and to check on reasons for non-acceptance of screening. Box 2 gives the objectives for primary care teams in cervical screening. The main obstacle to the success of cervical screening in primary care relates to organisation.

Need for systematic method of call-recall

The effectiveness of a systematic method of call-recall over an opportunistic approach in encouraging women to have a smear test has been shown in a randomised controlled trial. Other studies have confirmed that a call-recall system increases overall uptake, particularly among older women and those from lower social classes who have never been screened.

There is a wide range of methods of organising cervical screening in primary care. A study has shown that a younger age of general practitioner, a more rural practice, larger practice size, employment of a practice nurse, a belief in the efficacy of cervical screening, and a positive attitude to the time spent on cervical screening were all strong predictors of an organised approach to cervical screening within a practice.

Organisation of cervical screening varies between health authorities and between practices. The call-recall system can be administered either by the local health authority or family health services authority or from within the practice itself. Figure 3 shows a protocol for a practice call system. A coordinated system at a local level could improve the programme substantially. There is strong evidence that most practices are capable of providing well organised systematic programmes and the majority are now achieving high population coverage. When this is not the case, practices may need help and advice from a primary care cervical screening facilitator (box 3).

FIG 3

protocol for call-recall system for cervical screening in general practice*RF 7*

• Download figure
• Open in new tab
• Download powerpoint

Practice registers

A key issue influencing overall uptake and thereby the success of the screening programme is the adequacy of the population register. Studies in inner London and Manchester have found that between 30% and 60% of invitations were sent to wrong addresses. Even in a small, stable general practice population a study found that 2% to 4% of eligible women could not be traced.

Letter of invitation

The style, tone, presentation, and contents of the letter of invitation are important. There is more likelihood of a woman accepting an invitation for screening if the potential benefits of the test are properly explained (box 4). The availability of a female doctor or nurse should also be mentioned. One survey showed that 41% of women would prefer a female doctor to take the smear, and the proportion was higher among older women and women from lower socioeconomic groups.

Some studies have suggested that women who have not had a smear test before are more likely to accept an invitation if they are offered a specific appointment rather than an open invitation. Sending times of appointments is, however, administratively more complex than sending each patient an identical letter, especially if the letters of invitation are dispatched from the family health services authority. Offering a wide range of times when smears can be taken at the surgery should improve response rate.

Reminders by telephone are more effective than standard letters. Personal approaches from practice staff may help overcome anxiety about the test and the disease, but care must be taken not to coerce women against their will.

A leaflet should accompany the invitation letter. This offers the opportunity not only to address women's attitudes and beliefs about cervical screening but also to consider organisational and administrative barriers that can deter women from attending for a smear test.

Non-attenders

Improving population coverage requires to some extent an understanding of the reasons why women do or do not attend for cervical screening. Factors affecting uptake of cervical screening include the influence of age and social class on perceptions of vulnerability and the costs and benefits of screening. Many studies have shown that attendance can be inhibited by a high level of anxiety about the test and fear about cervical cancer, by erroneous beliefs about the relevance of the test, by concurrent family difficulties, and a low priority being accorded to cervical screening.

Studies have shown that flagging the notes of non-attenders and offering them screening on an opportunistic basis increases uptake.

Taking a smear

Of the 4.5 million smears taken in 1991-2 in England, 77.6% were taken in primary care. Evidence suggests that most were taken by practice nurses.

Women's anxiety about cervical screening can be reduced by providing adequate information before taking the smear (box 5).

A key factor determining the effectiveness of cervical screening programmes is the quality of smear taking. Poor smear taking can miss 20% or more of precancerous abnormalities. Ideally, smears should contain endocervical cells, metaplastic cells, endo-cervical mucus, and squamous cells. Adequate smears contain squamous cells and at least two of the remaining three elements. The person who takes the smear has the responsibility of ensuring that the cervix has been adequately sampled. In England in 1991-2, 6.3% of smears taken were inadequate (table III). Box 6 shows some of the reasons why smears may be reported as inadequate. Any infection or atrophy should be treated before repeating the smear. Repeat the smear at a minimum interval of four weeks. If smears are persistently inadequate, assessment by colposcopy should be considered.

Recent guidelines from the Department of Health emphasise the importance of appropriate training for all those who take smears.

Table III also shows the estimated national percentages of adequate smear test results for women of all ages in England in 1991-2. Borderline results and mild, moderate, and severe dyskaryosis occurred more often in the younger age groups, while suspicion of invasive carcinoma and suspicion of glandular neoplasia occurred more often in the older age groups.

Interpreting and managing results of smear tests

A smear may have negative results in that there is no nuclear abnormality, but other incidental comments may also be made about the smear (table IV). When a smear is reported with some abnormal cells, the action required will depend on many factors, including the appearance of any previous smears (table V, box 7).

Results from smear tests can be difficult to interpret. There is a whole spectrum, from completely normal to definitely malignant. The exact risk consequent on each grade is not clear. When recommending further action, referral of all grades of abnormality would lead to considerable overinvestigation and overdiagnosis. A careful balance thus has to be reached, taking into account the benefits and costs both to women and to the health service. Inevitably opinions will differ on how such a balance is arrived at.

The management of mild dyskaryosis presents precisely such a dilemma. Although the consensus is that a single mildly dyskaryotic smear should be managed by a repeat smear test at six months, with referral for colposcopy only if the abnormality persists, there are those who believe that women with such smears should be referred immediately for colposcopy. This is because, although the majority of such smears will revert to normal or persist as mildly dyskaryotic, a small proportion may progress to severe dyskaryosis. A balance obviously has to be achieved between ensuring appropriate management and not subjecting too many women to unnecessary medical procedures. A recent study from Aberdeen concluded that cytological surveillance, although safe, is not an efficient strategy for managing women with mildly abnormal smears. Three quarters of the women in the study subsequently required colposcopy and there was a high default rate among those put on surveillance. However, others have argued that referring all women with one mildly dyskaryotic smear for colposcopy would result in overinvestigation of very many more women than would ever go on to develop invasive disease. Ultimately, it may prove possible to differentiate high risk women within the broader group that has mild dyskaryosis by testing for high risk human papillomavirus types. Such testing could possibly be used as an adjunct to current screening methods. Further research is clearly needed to assess the role of cytological surveillance in mild dyskaryosis and to determine its optimal management. It will also be important to assess the psychological implications for women put on cytological surveillance as opposed to those who proceed to immediate colposcopy.

In practice, mild and moderately dyskaryotic smear test results do not correlate well with the histological diagnosis of mild or moderate dyskaryosis - that is, grade I or II cervical intraepithelial neoplasia. The difference is more in the quantity of cervical intraepithelial neoplasia rather than the quality. Grade III cervical intraepithelial neoplasia in the presence of mild or moderate dyskaryosis generally speaking occupies a smaller proportion of the transformation zone than when it exists in the presence of severe dyskaryosis on the smear. Women with one moderately dyskaryotic or two mildly dyskaryotic smears should be referred for colposcopy (box 7). A fairly large proportion of grade III cervical intraepithelial neoplasia could be expected in such women.

Figure 4 shows a protocol for managing patients with abnormal smear test results.

FIG 4

Protocol for managing patients with abnormal smear results. Whether to take biopsy specimen or treat immediately depends on balance between cytological and colposcopic findings and local protocols

• Download figure
• Open in new tab
• Download powerpoint

How should human papillomavirus infection be managed?

Increasing evidence implicates certain strains of human papillomavirus or wart virus as causing cervical cancer. Different strains of human papillomavirus have been identified, and they vary in their oncogenic potential. Correlation of virus type with the morphology of the cervical lesion shows that human papillomaviruses 16 and 18 are present in over 80% of invasive squamous cancers of the cervix and grade III cervical intraepithelial neoplasia.

Reporting changes due to human papillomavirus in cervical smears poses particular problems, both for women and for those who take smears who have to explain findings (boxes 7 and 8). No cell with evidence of human papillomavirus infection is normal, and no smear in which there is evidence of human papillomavirus infection should be reported as negative whether or not there is a substantial nuclear abnormality. Cells in which there is dyskaryosis in addition to cytoplasmic features of human papillomavirus infection should be reported according to the grade of dyskaryosis, regardless of the cytoplasmic changes.

Giving results to women with abnormal smears

Considerable evidence suggests that women with abnormal results on smear testing are extremely anxious because of concern about investigative procedures such as colposcopy and about the outcome, in particular fears about cancer. In one study 65% of women were worried or alarmed on receipt of a positive smear test result; 27% were shocked, stunned, or devastated.

The way in which a woman is told about an abnormal smear test result will often affect how she will cope with any treatment or future follow up. The most effective and efficient way of conveying this is to ensure that each woman receives her result in writing. A randomised study in the context of an abnormal smear test result concluded that inclusion of an information leaflet with the postal notification of the results led to significantly lower levels of anxiety. The information should be understandable and not alarming. Two studies found that the readability of the written information relating to abnormal results was a key issue in the acceptability of the information provided to women.

Women who receive an abnormal smear test result should be offered an opportunity to speak with their general practitioner to discuss the implications of the results. Waiting and uncertainty are often the most difficult part of the follow up process. If the result is not normal women often assume that it means cancer. It is therefore important to explain terms such as dyskaryosis and preinvasive cancer, and what further steps in investigation, follow up, and treatment may be needed.

Research has shown that anxiety and distress are considerably less in women who have had colposcopy explained to them before their appointment. A randomised study has shown that the high levels of anxiety experienced by women referred for colposcopy were reduced by providing a brief simple booklet about colposcopy with the appointment letter. More detailed information added to knowledge but did not reduce anxiety.

Who needs colposcopy?

Cervical intraepithelial neoplasia may be suspected in patients with abnormal cervical cytology, but diagnosis and treatment will depend on referral of such women for colposcopic assessment. The two main indications for colposcopy are to investigate an abnormal smear test result (table V, box 7) and to assess a clinically suspicious cervix more thoroughly even when smear test results are normal.

When should the cervix be treated?

In order to know when to treat the cervix it is important to have some idea of the rate of progression of abnormalities. Those at high risk of progression must be treated. There is currently inadequate information about the natural history of the lower grades of abnormality. The majority may not progress, but some would lead eventually to invasive disease if not treated at any stage. A balance must thus be reached between potential overdiagnosis and over-treatment and the need to ensure that progression to invasive cancer does not occur. It is therefore not possible to define a treatment policy with any degree of certainty. Although grade I cervical intraepithelial neoplasia carries low risk and grade III carries high risk, grade II cervical intraepithelial neoplasia is more difficult to categorise. On balance, it is currently believed that grades II and III cervical intraepithelial neoplasia should be treated once diagnosed. Grade I cervical intraepithelial neoplasia may be treated or kept under close surveillance (box 9).

Treatment aims at destroying cells in the transformation zone of the cervix. Extremes of heat or cold are equally effective. Some methods of treatment require two visits, while others deal with diagnosis and treatment in one visit, which has obvious advantages for the woman. Box 10 shows current methods of treatment. None of these methods has been evaluated in a randomised study and there is not strong evidence supporting one method over another. General practitioners and practice nurses should be familiar with the type of treatment offered locally.

Follow up of treated patients

There are four reasons for follow up.

• To identify residual disease

• To identify new cervical intraepithelial neoplasia

• To identify new invasive disease

• To reassure both the patient and the clinician.

The risks of the first three occurring are probably no different whether the patient has undergone an ablative or excisional technique.

Box 11 shows the recommendations for the follow up of treated patients.

Running an effective fail safe system

General practitioners have a key role in the fail safe mechanism, contributing to it in several ways (box 12).

When smears are taken by general practice trainees or practice nurses, responsibility lies with the general practitioner principal recorded on the request form. When smears are taken outside primary care and the result is sent to the general practitioner it is important to determine who is looking after follow up and referral as necessary.

Conclusion

Cervical screening is safe, reasonably specific, and has been proved in several countries. It does have disadvantages - a major problem being overdiagnosis, whereby many more women are found with intraepithelial disease than would be expected to develop invasive cancer in their lifetime. It does, however, have the potential to make a significant impact on incidence of and mortality from the disease.

Primary care teams have a vital role in ensuring the success of the cervical screening programme. Numerous studies have looked at the involvement of general practice in cervical screening, identifying many ways in which the service offered by practices can be improved. The setting of targets has provided a stimulus to most practices, a large percentage of which have now obtained the higher target of 80%. Many practices are now running well organised and effective schemes.

A well organised programme depends on screening a sufficiently wide age range, at an interval between three and five years. It requires adequate population registers, an effective call-recall system, a reliable fail safe procedure, and good quality control. If uniform acceptance rates of 80% can be achieved a reduction in mortality of 65% to 70% can then be expected in the long term.