SEARCH

SEARCH BY CITATION

Keywords:

  • Cervical cancer;
  • patient compliance;
  • screening programme

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgement
  13. References

Please cite this paper as: Moss E, Pearmain P, Askew S, Owen G, Reynolds T, Prabakar I, Douce G, Parkes J, Menon V, Todd R, Redman C. Implementing the national invasive cervical cancer audit: a local perspective. BJOG 2010;117:1411–1416.

Objective  To monitor the effectiveness of the cervical screening programme and identify suboptimal management in order to improve patient care.

Design  Retrospective study.

Setting  A university hospital serving a population of 1 million people.

Population  All women diagnosed with a cervical cancer between 2003 and 2006.

Methods  Analysis of data from invasive cervical cancer reviews.

Main outcome measure  Categorisation of cervical cancer cases according to the Invasive Cervical Cancer Audit classification.

Results  Eighty-seven women were diagnosed with cervical cancer during the 3-year study period. The ‘lapsed attender’ group accounted for the greatest number of cases (30%), followed by screen detected (26%), interval cancers (13%), never attended (12%), lost to follow-up (10%) and never invited (9%). Women who had never attended for cytology presented with higher stage disease, stage-II or above, compared with the screen-detected cases: 60% were stage II or above, compared with 13.0%, Chi-square = 0.018. The most frequently identified screening programme problem was patient compliance, which was determined to be the principle contributing factor in 39 cases (45%) and a secondary factor in a further ten cases.

Conclusions  The categorisation of cervical cancer cases has the potential of yielding invaluable information for improving programme effectiveness. Patient compliance is the greatest challenge to the screening programme, and the need for regular screening and adherence to follow-up regimens needs to be reinforced in order to maximise the efficacy of the national screening programme.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgement
  13. References

The fall in the incidence of cervical cancer in the UK has been attributed to the national screening programme (NHSCSP),1 and it is estimated that at least 80% of expected cervical cancer deaths have been prevented over the past 20 years.2 Despite the continued fall in the incidence of cervical cancer, there are still over 2800 new cases diagnosed each year in the UK,3 and there is growing concern that the fall in the uptake of screening,4 particularly by the younger age groups, will contribute adversely to this rate. In 2006–07, 4.01 million women aged between 25 and 64 were invited for screening, a fall of 1.2% compared with 2005–6; however, the number of women in this age group attending for screening during the period 2006–07 was 3.17 million, a fall of 5.6% compared with the previous year. There is also evidence that the frequency of screening is falling in the 25–64-year-old population: 82% of women had an adequate result within the previous 5 years in 1997 compared with 79.5% in 2007.5

Detailed analysis of cervical cancer cases that do develop despite the presence of such a comprehensive national screening programme can yield information on the effectiveness or inadequacies of the programme. In 2006 the NHSCSP published a document entitled ‘Audit of invasive cervical cancers’, which contained detailed information on the aims, requirements and methods for evaluation needed to perform a review of cervical cancer cases.6 Classification of cancers into one of six categories enables an analysis of the epidemiology of cases and an evaluation of the effectiveness of the screening programme within subpopulations of women.

A detailed investigation has been performed on all cervical cancer cases diagnosed at the University Hospital of North Staffordshire (UHNS) since 2003. The aim of this study was to review all the cases and categorisations to assess the effectiveness of the screening programme within the North Staffordshire population.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgement
  13. References

All cases of cervical cancer diagnosed in the triennium 2003–06 were classified by the UHNS into one of the six nationally described categories:6 screen detected, interval cancer, lapsed attender, never invited, never attended and lost to follow-up. The process consisted of a review of screening history, gynaecological management, cytology and histology slides. All available original slides and reports were examined, and absent slides were traced and if necessary requested from other medical institutions. Liquid-based cytology was not in use at UHNS during the study period, and 3-yearly invitations were consistently being issued during the study period for all age groups. The cytology slides were reviewed by a cytology screener, a senior biomedical scientist in cytology and two consultant pathologists with a special interest in gynaecological oncology. The histological review was performed by two consultant pathologists. A local decision was taken, outside of the core national protocol, to obtain an independent external second opinion on specimens where a lack of consensus was found among the panel reviewing either the cytology or histology specimens.

The core review team consisted of a consultant gynaecological oncologist, a consultant pathologist and a senior biomedical scientist in cytology. The principle factor contributing to the development of a cancer was determined by the reviewing team. In all cases, except where no screening history was identified or if the woman was significantly ‘at fault’ (i.e. lapsed), and there were no other screening programme problems identified, once the review was completed the woman was invited to attend an outpatient appointment to discuss the outcome with the consultant in charge of their care.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgement
  13. References

Eighty-seven women were diagnosed with invasive cervical cancer during the 3-year study period (Table 1). The median age of the study group was 43 years, with a range of 24–91 years. There was a significant difference in age and stage of disease, with a greater number of stage-I tumours diagnosed in younger women, median age 39.5 years (range 24–69 years), as compared with stage-II and above, median age 53 years (range 26–91 years), Mann–Whitney = 0.002. There was no association between histological subtype, adenocarcinoma and squamous cell carcinoma, and stage at presentation, stage I compared with stage II and above, Yates’ corrected chi-square = 0.116.

Table 1.   Stage at diagnosis and tumour type, n = 87
Stage at diagnosisSquamous cell carcinomaAdenocarcinomaOther*
  1. *Other: 1B1, adenosquamous and neuroendocrine; 2B, adenosquamous.

1A12050
1A2130
1B12062
1B2200
2A210
2B1701
3400
4210
Total68163

Screening history review

Seventy-five women (86%) had had at least one smear taken at some point in the past or at the time of diagnosis. There was no association between the number of smears performed and the age of the patient, Spearman’s rank correlation = 0.197. In 40 of the 87 cases (46%) one or more smears had been reported as abnormal or inadequate. The time interval between last cytology and the diagnosis of malignancy was 3 years or less in 40 cases, 5 years or less in 52 cases and over 10 years in seven cases.

Cytological slide review

There were non-consensus reviews for four smears, two negative and two inadequate smears on original review, that were subsequently referred for a second, independent review. Of the 75 women who had had a previous smear, 67 had a satisfactory review of all the available cytological slides. In six cases the smears had been performed over 10 years before, and were therefore not available; in two cases the slides could not be located and therefore the review was classed as inadequate. Cytological review revealed an under-call in 23 cases, an overcall in one case and inadequate cytology in four cases. The undercall smears included: seven negative smears, three of which were deemed highgrade on review; four inadequate smears, two were deemed high-grade on review; and in three cases there was a difference in grade of abnormality, but this was only clinically significant in one case. Cytological undercall was deemed to be the principle-contributing factor in 15 cases (17%): three adenocarcinomas and 12 squamous cell carcinomas. No differences were identified on review of the histology specimens, and no punch biopsies were performed in the cohort.

Gynaecological management review

Twelve patients (14%) had previously been referred for colposcopy. In four patients who had had colposcopy previously, a problem with the colposcopic management was identified; however, colposcopic management was judged to be the principle factor in only three cases, one adenocarcinoma and two squamous cell carcinomas.

Classification

The lapsed-attender group was the largest group, with 26 of 87 cases (30%) (Figure 1). Six of the eight cases in the never invited category were over the age of 80 years at diagnosis, and would never have received an invitation for screening. Nine women were lost to follow up, four women had not attended for repeat cytology, four women had failed to attend for colposcopy following abnormal cytology and one woman had not attended for repeat cytology following a colposcopy.

image

Figure 1.  (A) Classification of cases, n = 87. (B) Classification of cases with the exclusion of the never-invited category, n = 79.

Download figure to PowerPoint

Breaking down the cases into the audit categories, a significantly greater proportion of the never-attended group were diagnosed with advanced disease (stage II and above) as compared with the screen-detected group, six out of ten (60%) patients compared with three out of 23 (13%), Yates’ corrected chi-square = 0.018, whereas no difference was identified between the lapsed-attender, six out of 26, and screen-detected groups, Yates’ corrected chi-square = 0.592 (Table 2). The median age at diagnosis differed between the categories with never-attended cases being diagnosed at a later age compared with the screen-detected and lapsed-attender groups, but this result was not statistically significant (Figure 2).

Table 2.   Classification of cases and stage at diagnosis, n = 87
Stage at diagnosisScreen detectedInterval cancerLapsed attenderNever invitedNever attendedLost to follow-upTotal
12062004959
234464021
30002204
40120003
Total231126810987
image

Figure 2.  Age at diagnosis in the screen-detected (n = 23), lapsed-attender (n = 26) and never-attended categories (n = 10).

Download figure to PowerPoint

Screening programme problems

One single screening programme problem was identified in 45 (52%) cases, and two concurrent problems were identified in 12 cases (14%). The most frequently identified problem was patient compliance, which was identified as the principle contributing factor in 39 cases (45%). In cases where two problems were identified, compliance was the secondary factor in ten cases (83%). Of the 18 cases associated with cytological undercall or colposcopic management, only three cases (16.7%) were stage II or above at diagnosis, compared with 12 out of 39 (30.8%) of the cases associated with patient compliance, and 13 out of 30 (43.3%) of the cases where no screening problem was identified.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgement
  13. References

The dramatic fall in the number of cervical cancer cases in the UK is testament to the effectiveness of the national screening programme; however, the fact that over 2800 cases are still diagnosed each year in the UK implies that improvements could be made within the programme in order to reduce the incidence still further. This study reviews the classification of cervical cancer cases by the UHNS prior to the introduction of the NHSCSP Audit of Invasive Cervical Cancers document.6 The original aim of the classification exercise was to investigate all cervical cancer cases diagnosed in the North Staffordshire population in an objective and unbiased manner in order to assess the effectiveness of the national screening programme. Several studies have performed similar audits;7,8 however, to the best of the authors’ knowledge this is the first study to comprehensively review individual cases in such an in-depth manner, and in particular to perform a pathological review of all previous cytology and histology.

There is a debate as to whether the analysis of cervical screening failures should be confined to frankly invasive cancer (stage IB1+), as the majority of microinvasive cancers are screen detected and could be regarded as a successful outcome of screening.9 In this study we have included all cervical cancers in order to investigate the epidemiological associations of all disease stages, and have shown that screen-detected cases were not confined to microinvasive disease.10

The population of women in our study diagnosed with a cervical cancer was demographically similar to the national population with regards to peak age at diagnosis and histological diagnosis.3 The compliance rate with the screening programme in our population of cervical cancer cases, however, was substantially lower when compared with the North Staffordshire and national screening populations: only 59% of women having had an adequate smear within the last 5 years versus 80 and 79.2%, respectively, for 2006–07.4 Poor compliance with national screening programmes has been consistently demonstrated as a risk factor for cervical cancer in many different countries.11–18 In our study, patient compliance was the single greatest contributing factor to the development of cervical cancer, being identified as the principle factor in 45% of cases and a contributing factor in a further 11.4%. Factors associated with non-compliance with cervical screening are well recognised, in particular anxiety and embarrassment,19 and it appears that the level of knowledge and understanding of women is not improving with time.20 Many strategies have been employed to overcome the obstacle of compliance at the screening and referral level, including reminder letters and phone calls, detailed information leaflets and pre-colposcopy discussion sessions.21,22 There are standard call, recall and reminder procedures within the programme; however, because compliance has been identified as such a strong contributing factor the introduction of additional reminders, psychological support or media campaigns may potentially have an additional effect on the uptake of screening.

In our population, lapsed attendance was associated with the greatest number of cancers in women aged between 30 and 39 years, which corresponds with the NHSCSP data showing reduced coverage in younger age groups.4 When looking at the distribution of stage at presentation between the lapsed-attender and the never-attended groups, our results also indicate that sporadic compliance with the screening programme is associated with a more favourable outcome compared with no screening at all. Our finding that women with screen-detected cancers are significantly more likely to be diagnosed at an earlier stage compared with women who had never attended for screening has been shown in other studies.8,23–25 Van der Aa reported that 82% of screen-detected cancers were stage I compared with 62% that were not screen detected.16 Also, the median age of diagnosis in our study was greater in women who had never attended compared with the screen-detected group, a finding consistent with the published literature.17,24 The ‘never attended’ are a small group of women who may never be persuaded to attend for screening; however, the identification of such patients to their general practitioner and time spent exploring individual patient’s concerns or misconceptions may reduce this group to a minimum.

An inherent false-negative rate associated with conventional Papanicolaou cytological screening is acknowledged, and is estimated to be in the order of 20–25%.25 Even severe abnormalities, including invasive carcinomas, carcinoma in situ and adenocarcinoma in situ, can be misdiagnosed as normal because of screening, interpretation or sampling errors.26 In our study the percentage of cases associated with false-negative cytology was 17%, which is below the rate determined in a meta-analysis by Spence, who concluded that approximately 29.3% (95% confidence interval 21.2–40.4%) of cases were associated with false-negative Pap smears.27 The meta-analysis also investigated the contribution of poor follow-up management on the development of cervical cancer, and determined a rate of 11.9% (95% confidence interval of 9.0–15.6%), which is markedly higher than our rate of 3.4%. All three cases in our study associated with poor follow-up management were stage I at diagnosis, and this suggests that although failures may occur in the clinical management within the screening programme they are unlikely to be associated with increased patient mortality because of the early stage of diagnosis.

This study gives a clear picture of the epidemiological factors associated with the development of cervical cancer in the North Staffordshire population. The principle contributing factors may not be the same across the UK, and may vary depending on geographical location, and ethnic and social class mix of the population, as coverage of the screening programme has been shown to vary by up to 9% in different regions the UK.5 We have highlighted the detailed information it is possible to glean from analysing the screening history and management of women diagnosed with a cervical cancer, and how that information can be used to inform local decisions regarding the screening programme.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgement
  13. References

Cervical cancer case review has the potential of yielding invaluable information for improving programme effectiveness. Patient compliance is the greatest challenge to the screening programme, and the need for regular screening and adherence to follow-up regimens needs to be reinforced in order to maximise the efficacy of the national screening programme.

Contribution to authorship

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgement
  13. References

CWR and EM designed the study. GO, SA, IP and PP participated in the data collection. GD, JP, RT and VM contributed to the review of the data. CWR, EM and TR analysed the results. EM drafted the manuscript and all authors approved the final version.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosure of interests
  9. Contribution to authorship
  10. Details of ethics approval
  11. Funding
  12. Acknowledgement
  13. References