Research and analysis

NHS Cervical Screening Programme Audit of invasive cervical cancer: national report 1 April 2016 to 31 March 2019

Published 30 November 2023

Applies to England

1. Executive summary

The NHS Cervical Screening Programme in England provides high-quality cervical screening to a target population of just under 16 million individuals (women and people with a cervix) who are eligible for screening. The cervical screening programme is highly effective in preventing cervical cancer and still more effective in preventing death from the disease. This report focuses on analysis of 6,369 cases of cervical cancer diagnosed between April 2016 and March 2019.  The audit is carried out by the Screening Quality Assurance Service (SQAS) collecting and linking data from hospital trusts and the national cervical screening call and recall computer system. 

1.1 Human papillomavirus (HPV) triage

This report evaluates the impact of introducing human papillomavirus (HPV) triage of borderline or low-grade cytology into the screening programme. We find it has led to a timelier diagnosis by improving the time between the abnormal test and diagnosis. It has increased the proportion of individuals being diagnosed with early stage cervical cancer (stage IA) which has a much better prognosis and requires less aggressive treatment than for late stage cancer.  The proportion of people who did not go on to develop cancer who are asked to return early for repeat testing following a low-grade abnormality halved from 19% to 10% by 2013 and is now three-fold lower (6%) suggesting HPV triage and test of cure protocols have significantly reduced the number of women needing to be on early recall. This has the benefit of reducing the anxiety people experience with early repeat testing.

1.2 Screening history

The audit assesses screening histories of individuals diagnosed with cervical cancer and compares them to those of people without cancer. Among people diagnosed with cancer, the proportion who are up to date with screening has decreased over time, and a larger proportion have no screening history. The increase in people with no screening history is mostly observed among those diagnosed with early stage cancers (stage IA and IB) and is most noticeable in people under the age of 35.

Compared to people who attend screening every 3 years, the likelihood of being diagnosed with cervical cancer among those who have no screening history has increased over time at all ages and stages. Since screening has improved over time (through better quality assurance and more sensitive screening tests) the risk of cervical cancer among those previously screened as recommended has become lower increasing the odds among those who have no screening history.

These audit results show that the high priority being given to encouraging individuals to attend cervical screening routinely is evidence-based as it reduces the chance of developing cervical cancer.

1.3 Deprivation

For the first time, we report on deprivation data from the audit. Young individuals diagnosed with cervical cancer were more likely to be in the most deprived quintile compared with individuals over age 65 at diagnosis.

Individuals aged 25 to 64 in the lowest quintile of deprivation were more likely to never have been screened and those in the least deprived quintiles more likely to have been screened as recommended prior to diagnosis. These audit results show the importance of initiatives to increase awareness of cervical screening and support participation in deprived populations.

1.4 Age at diagnosis

Fourteen percent (14.4%) of all cervical cancers are diagnosed in people aged 65 or over. There has been an increase in the proportion of cancers diagnosed as stage II or worse in this age group from 64% between 2007 and 2008 to 78% between 2018 and 2019. This is because less people with cancer diagnosed when they were aged 65 to 79 (70%) had a screening test between the ages of 50 to 64, when compared with people of the same age without cancer (93%).

The audit results show that there has been no increase in the total number of people being diagnosed with cervical cancer aged 25 to 34, following the change in age at which people are first invited from age 20 to age 25 in 2008 to 2009. In the previous report we found the number of individuals diagnosed with cervical cancer aged 25 to 26 had tripled compared to the period where women were invited at age 20. In this report the number of individuals diagnosed with cervical cancer aged 25 to 26 was double compared to those diagnosed between 2007 and 2010. This is in line with expectations that most women in this age group will have been offered vaccination against human papilloma virus (HPV) prior to age 20. It is still the case that most people in the 25 to 26 year old age group are being diagnosed on or around their first invitation for screening making a diagnosis of early stage (FIGO 1A and 1B) cervical cancer (rather than late stage cancer) more likely.

In this report there were 130 people diagnosed under the age of 25. Of the 130, only 25 (19.2%) were diagnosed before they would have been invited for screening for the first time.

1.5 Incidence

For the first time the peak incidence of cervical cancer is observed in the 30 to 34 year old age group (967 cases or 15.2%), followed by cases in people aged 25 to 29 (856 or 13.4%). The most common histological type is squamous carcinoma (65%), followed by adenocarcinoma (18%) and adeno-squamous carcinoma (2%). 

1.6 Staging and morphology

The proportion of cases with missing cancer stage information has more than halved since the first audit report was published in 2011 (from 21% to 5.5% in this report). We estimate that had all the cervical cancers been staged, 33% of cancers would be stage IA, 31% stage IB, 27% stage II or III and 9% stage IV.

2. The audit

The NHS Cervical Screening Programme (NHS CSP) aims to reduce the incidence of, and mortality from, invasive cervical cancer. It does this by regularly screening all individuals at risk, so that abnormalities that might otherwise develop into invasive cancer can be identified and treated.

Unfortunately, despite the effectiveness of population-based screening, people continue to develop cervical cancer. (1) Purely monitoring changes in cervical cancer incidence and mortality rates can help to determine whether the programme is achieving its objectives. However, this alone does not give a complete picture because it cannot assess whether changes to the programme (for example, improvements to the screening technologies employed, changes to the age range over which people are called for testing, or to the frequency of screening at different ages) are affecting the incidence of cervical cancer. Nor does it indicate the effectiveness of the screening programme under optimal conditions, for example if everyone attended when invited or the benefit of screening individuals who have never had a test.  Audit of all cases of cervical cancer along with those people’s interactions with the screening programme and comparing them with people who do not develop cervical cancer is therefore essential.

The main purpose of the NHS CSP audit of invasive cervical cancer (hereafter, ‘the audit’) is to:

• monitor the effectiveness of the cervical screening programme by comparing the screening histories of people (with a cervix) who develop cervical cancer with those (with a cervix) who do not
• identify areas of good practice and indicate where improvements might be made to support evidence-based policy and practice
• support the continuous learning and development of health professionals involved in the programme

For the information to be reported consistently, all parties in the cervical screening programme follow the same national protocol for auditing cases of invasive cervical cancer. This has recently been updated in 2021 (2) so data in this report were collected under the previous protocol (1).

All diagnosed cases of cervical cancer in England are included in the audit. Data are collected by financial year, which runs from 1 April to 31 March. There are currently 5 published national reports covering cases diagnosed from April 2007 to March 2016.

This, the sixth report, contains data for the 3 year period from 1 April 2016 to 31 March 2019 (financial years 2016 to 2017, 2017 to 2018, 2018 to 2019). Several tables shown in previous reports are not included or have been changed and thus it is not possible to directly map all the tables in this report to the first four national audit publications. This is due to a reconfiguration of regions in recent years. 

This report prioritises timeliness over data completeness. Hence, we include cases for which not all details have yet been recorded within the audit, and readers should take note of the caveats attached to this approach.

Summary details of the audit methodology are available to read alongside this report, and the full national audit guidance is also available.

3. Background

Cervical cancer is a malignant neoplasm of the cervix uteri.  In 2018, 2,668 cases were registered in England, with an age standardised incidence rate of 9.6 per 100,000 people. (3)

There is consistent evidence from across the world that high-risk human papillomavirus (HR-HPV) infection is a necessary, but on its own not sufficient, cause of cervical cancer. (4) Cofactors that appear to increase the risk of developing cervical cancer in HPV-infected people include:

• the use of oral contraceptives
• smoking
• high number of pregnancies
• previous exposure to other sexually transmitted diseases, such as chlamydia trachomatis and herpes virus type 2.

Cervical screening is not a test for cancer, but a means of preventing it.  In the time period covered by this report, the NHS CSP used cervical cytology with HPV triage testing to detect potential abnormalities or the presence of the HPV virus which could develop into cancer if left untreated. People identified with these results are referred for a detailed assessment of their cervix, and potential treatment, at a specialist colposcopy clinic.  Subsequent to time period covered in this report the NHS CSP have implemented HR-HPV primary screening where HR-HPV testing is used as the first screening test with cytology used as a triage if the HPV test is positive for HR-HPV. The next audit report will cover this programme change.   

All individuals with a cervix between the ages of 25 and 64 are eligible for free NHS cervical screening. Cervical screening is not offered to people who have no cervix, or to those who have made an informed choice to opt out of the programme.

Table A. Current cervical screening intervals

Age group (years)	Frequency of screening
24.5	First invitation received
25 to 49	Every 3 years
50 to 64	Every 5 years
65+	Invitation as required for people who have had recent abnormal tests. Individuals who have not had an adequate screening test reported since age 50 may be screened on request

4. Audit findings

This section analyses and discusses the audit’s main findings.  Detailed data tables are presented in Appendix B: Data Tables.

5. Invasive cervical cancer

Over the period April 2007 to March 2019, 26,867 cases of invasive cervical cancer were included in the audit. Their screening histories are compared to 52,738 individuals without cervical cancer. When a person is diagnosed with cancer in the UK information about them is automatically included in their national cancer registry. It is data from cancer registries that tells us how many people are diagnosed with cancer. Table B provides a broad assessment of the audit’s coverage (see also Appendix A: data completeness Table A-1), comparing the number of cases of invasive cervical cancer included in each audit year (corresponding to the financial year) with the number reported nationally in the national cancer registry in each calendar year. Although some cases included in the audit are not included in cancer registry data, and vice versa, the number of cancers reported to registries since 2007 is around 12% greater than the number included here. However, the number of cancers reported to registries since 2016 is around 20% greater than the number included here.

Updated estimates from the Office for National Statistics (ONS) report 2,594 diagnoses of invasive cervical cancer during 2016 and 2,591 during 2017 (5), whereas the audit comprises 2,182 cases between April 2016 and March 2017 and 2,132 between April 2017 and March 2018. The most recent year of data included in the audit report is usually the least complete, with 2,055 cases reported here for April 2018 to March 2019 compared with 2,668 reported nationally for 2018. (3) SQAS staff work with hospital staff and National Cancer Registration and Analysis Service (NCRAS) colleagues to minimise these discrepancies and to make both data sources more directly comparable. 

Table B also presents a recent history of the number of cervical cancers included in each audit year. We have included this to illustrate the amount of new data received each year. Compared to national registrations, 80% of cancers diagnosed between April 2010 and March 2011 had been reported into the audit in time for inclusion in the 2011 report, 89% in the 2012 report and 77% in the 2016 report. By comparison, 81% of cancers diagnosed between April 2016 and March 2019 had been reported into the audit in time for inclusion into this report. Disruptions caused by the COVID-19 pandemic have impacted the exchange of data between SQAS and NCRAS affecting discrepancies between data sources.

Table B. Number of cases of cervical cancer included in each audit year compared with those reported nationally

Audit Year	Calendar year	Cases included in 2010 report	Cases included in 2011 report	Cases included in 2012 report	Cases included in 2013 report	Cases included in 2019 report	Cases included in current report	Cancer Registration
2007/08	2007	2,089	2,136	2,158	2,144	2,163	2,133	2,337
2008/09	2008	2,164	2,205	2,254	2,262	2,286	2,290	2,409
2009/10	2009	1,978	2,349	2,452	2,474	2,528	2,520	2,766
2010/11	2010	0	1,876	2,087	2,099	2,162	2,175	2,346
2011/12	2011	0	0	1,969	2,191	2,363	2,360	2,511
2012/13	2012	0	0	0	2,020	2,262	2,302	2,482
2013/14	2013	0	0	0	0	2,118	2,210	2,639
2014/15	2014	0	0	0	0	2,093	2,241	2,590
2015/16	2015	0	0	0	0	1,817	2,267	2,517
2016/17	2016	0	0	0	0	0	2,182	2,594
2017/18	2017	0	0	0	0	0	2,132	2,591
2018/19	2018	0	0	0	0	0	2,055	2,668
Total							26,867	30,450

Source: We have used the updated number of registrations from the Cancer Registration Statistics publication (2016,2017, 2018). As with the Audit, ONS receive notification of several additional cases after they have published their yearly statistics. Cancer registrations are published by calendar year; Audit data are by financial year (1 April to 31 March).

There is a trade-off between presenting data in a timely manner and the completeness of that same data.  For details on how timeliness affects our estimates of the International Federation of Gynaecology and Obstetrics (FIGO) stage and age at diagnosis please refer to the summary methods document (link). In this report we focus only on the most recently diagnosed cervical cancers by restricting the data in this report to 6,369 cases diagnosed between April 2016 and March 2019 and comparing them to 12,392 people without cancer.  However, where relevant we have used all cancers reported to the audit.

When a person with cervical cancer is identified for inclusion into the audit up to 4 people without cervical cancer get identified from the screening database. The people without cervical cancer are referred to as population ‘controls’. Controls are selected (or matched) if they are born within 2 years of the person with cancer and to either have the same GP or live in the same area. Most cases submitted (n=6,097) to the audit have at least 2 age-matched population controls (GP and district).  However, for a small number of cases (198), only one of these controls was identified, while 74 cases were submitted with no population control. For a defined subset of cases, up to 2 further controls were selected, resulting in 5,649 screened controls and 3,313 abnormal controls.

6. Age at which invasive cervical cancer is diagnosed

Figure 1 shows the number of cases of cervical cancer by age in the audit between April 2016 and March 2019, compared with the numbers reported nationally for January 2016 to December 2018. For the first time the peak number of cases in the audit is observed in the 30 to 34 year old age group (967 or 15.2%), followed closely by cases in people aged 25 to 29 (856 or 13.4%), and aged 35 to 39 (824 or 12.9%). National Cancer registration figures report the same age trend, Figure 1. The under-reporting of cases to the audit is better illustrated in Appendix A: data completeness, Table A-1 where audit cases reported by calendar year are compared to cases reported in the national statistics. It shows that under-reporting to the audit increases with age at diagnosis, so that we are missing more cases for people over the age of 65 than for those under 65.

In this audit, 85.2% of all cases of invasive cervical cancer fell within the age group eligible for cervical screening (25 to 64 years, see Appendix B: data tables, Table 3) and 14.4% were diagnosed at age 65 or older. In 2018, people eligible for screening made up 79.6% of all cervical cancer registrations in England. (3)

In December 2012, a change to the age at which people are first invited for screening came into effect. This saw people receive their first invitation to participate in screening at age 24.5 instead of 25. As a result, among people aged 20 to 25 inclusive, the proportion diagnosed with cervical cancer at age 24.5 to 25 increased from 35% (101 out of 291) prior to January 2013 to 74% (251 out of 340) of people diagnosed from January 2014 to December 2018. To reflect the change in policy throughout this report, from this point onwards, we include people diagnosed age 24.5 to 25 in the 25 to 29 age group. In this report there were 130 people diagnosed under the age of 25. Of the 130, only 25 (19.2%) were diagnosed before they would have been invited for screening for the first time.

Figure 1. Number of cases of invasive cervical cancer submitted to the audit compared with the number registered by ONS for 2016 to 2018, by age at diagnosis

7. FIGO stage of invasive cervical cancers

Table C shows the observed and estimated proportion of cervical cancer cases in each FIGO stage. FIGO stage information is missing for 4.3% of cases and clinical staging was not possible for 1.2%, therefore no staging data is available for 5.5% of cases (see Appendix A: data completeness, Table A-3). Figures in this report suggest that of the 273 (4.3%) cases with missing stage, staging data was not available for 38% (n=104) whilst for the majority the information may be available but not yet recorded within the audit.

The proportion of cases with missing stage has been reduced by 74% since the first audit report was published (from 21% to 5.4% in this report). We know that those cancers where the FIGO stage is unknown tend to be of a higher stage than those where stage is known (see summary methods document). Therefore, we estimate that had all the cancers been staged, 33% of cancers in the audit would be stage IA, 31% stage IB, 27% stage II or III and 9% stage IV (Table C).

The earlier the stage of cancer at diagnosis, the better the prognosis is for the individual and the less aggressive the treatment.

Table C. Cases of invasive cervical cancer in the audit between April 2016 and March 2019 audit, by FIGO stage

FIGO stage	Cases     (n)	Observed proportion (%)	Estimated percent distribution (%)
IA	1,993	31.3	32.9%
IB	1,873	29.4	31.0%
II & III	1630	25.6	27.1%
IV	526	8.3	8.9%
IB(NOS)*	74	1.2	-
None recorded	273	4.3	-
Total	6,369	100	100
* Cases reported as IB(NOS) are known to be stage IB or worse but detailed stage is not known

As a proportion of all cancers, invasive cervical cancer at FIGO stage II or worse was more likely to be diagnosed in people over the age of 50 than in those under 50, with stage IA disease becoming less frequent with increasing age. This is because people diagnosed with cervical cancer over 50 are much less likely to have been screened recently than those people of the same age without cancer. Their cancers are more likely to be diagnosed following investigation of symptoms and therefore present at a later stage. By contrast, between the ages of 25 and 49, the majority (44.6%) of cervical cancers were found to be stage IA (see Appendix B: data tables, Table 5a).

The difference in the stage at diagnosis by age is even more apparent when looking at rates of cervical cancer by FIGO stage (Figure 2). We used the observed rates of cervical cancer by age group reported in the Cancer Registration Statistics (2) in 2018 and applied the FIGO stage distribution by age observed in the audit dataset.

Detailed FIGO stage (for cases with known stage) by age group at diagnosis is presented in Figure 3. In 1,582 out of 1,993 cases of stage IA cancer, further details were provided, and these suggest that 93% are stage IA1 and only 7% stage IA2 (Appendix B: data tables, Table 4). 56% of stage IA1 cancers were diagnosed in people aged between 25 and 34, 34% were diagnosed in people aged 35 to 49 and 9% in those aged 50 to 64. Only 0.9% were diagnosed in people over the age of 65. Of the 1,355 people with stage 1B cervical cancer, further detail was known for 214. 69% of these people were diagnosed with stage IB1 cancer and 31% were stage IB2.

Figure 4 shows the estimated number of cervical cancer cases by FIGO stage, year of diagnosis, and age at diagnosis. Estimates for the last year of data included in the audit are dotted, as this more recent data is less complete, and we are less certain of the accuracy of the results. The ‘Jade Goody effect’ (a rise in the number of younger people attending cervical screening appointments following the diagnosis and untimely death of the reality TV star) can be seen in an increase in the number of stage IA and IB cancers diagnosed in people aged 25 to 34 (in particular) in between 2009 and 2010 (Figure 4b). No increase was observed in stage II or worse cancers.

The effect of raising the age at which people are first invited for screening from 20 to 25 can be seen from 2009 onwards (Figure 4b), with an increase in early stage IA cancers diagnosed in people aged 25 to 34 and a corresponding decrease in those under age 25 (Figure 4a).

Since people aged under 24.5 are not invited for screening, the likelihood of screen-detected cancer in this group is close to zero. Almost all individuals under age 25 have never been screened (98% of people without cancer and 100% of people with cancer, see Appendix B: data tables, Table 14). Therefore, it is mostly people who are investigated because of reported symptoms who are likely to be diagnosed at this age. Out of the 25 cases with recorded FIGO stage diagnosed in people aged 20 to 24.5, 33.9% were diagnosed as stage IB and a further 44.9% as stage II or worse.  To put this in context, there were 1,879 people with known stage diagnosed with cervical cancer aged 25 to 34 and only 12.8% were diagnosed with stage II or worse, Figure 3.

It is noteworthy that in the rate of cervical cancer in women aged 20 to 24 in 2015 (5) was 4.1 per 100,000 population, by 2018 (3) (Figure 2) the rate decreased to 1.9 per 100,000 population. Rates have also decreased (from 22.0 per 100,000 population) in women aged 25 to 29 (to 16.8 per 100,000 population). The decrease in rates is in line with expectations given that most women will have been offered vaccination against human papilloma virus (HPV) prior to age 20. We recently found that vaccination against HPV in England has resulted in an estimated 448 (95% CI: 339 to 556) fewer cervical cancers in women younger than 30 years by June 2019.(6) Among individuals who received the vaccine at ages 12 to 13 (aged younger than 25 in 2019) cervical cancer was almost eradicated. Therefore, the expectation is that rates in young women will continue to decrease whilst the highest rates of cervical cancer are expected in the cohort of women currently aged 30 to 34 since they have not benefited from vaccination and have historically lower screening coverage compared to older cohorts.

People over the age of 50 at diagnosis, and particularly those diagnosed after the age of 65, are more likely to be diagnosed with advanced stage cancer than younger people. Although the absolute number of cancers diagnosed in people aged 65 to 74 has fallen from 178 between 2007 and 2008 to 162 between 2018 and 2019 (Figure 4e, Table 2), there has been an increase in the proportion of cancers diagnosed as stage II or worse, from 64% between 2007 and 2008 to 78% between 2018 and 2019 (not shown). 30% of people diagnosed with cervical cancer aged 65 to 79 had not been screened between the ages of 50 and 64 compared with only 7% of people of the same age without cervical cancer (Appendix B: data tables, see Table 14). Of those screened between ages 50 and 64, only 25% of people with cancer had a negative test and were up-to-date with screening compared to 70% of people without cancer.

It is worth noting that people aged 75 at diagnosis (between 2007 and 2018) will have been aged between 45 and 56 when screening was first introduced in 1988. Therefore, most people aged over 75 are unlikely to have participated more than once or twice in the screening programme.

Figure 2. Observed incidence rates of cervical cancer per 100,000 people in England, 2018: by FIGO stage and age at diagnosis

Figure 3. FIGO stage of cervical cancer cases: percentage distribution, by age at diagnosis

Figure 4. Trends in the number of cervical cancer cases diagnosed from April 2007 to March 2019, by FIGO stage and age at diagnosis

8. Histology of invasive cervical cancers

Figure 5 shows the distribution of invasive cervical cancer cases by histological type. Most of the cases of cervical cancer show squamous histology (65%), while 18% are adenocarcinomas. Adenosquamous types are considerably less common (2%). Squamous carcinoma is more likely to be diagnosed as stage IA cancer than the other histological types. 37% of squamous carcinomas were stage IA compared to 23% of adenocarcinomas and 7% of adenosquamous cancers. 51% of the cases with undifferentiated and 28% of the other histological types were diagnosed as stage II cancer (with 34% and 22%, respectively, diagnosed as stage IV).

We have previously found (7), that while cytology screening is capable of detecting early-stage invasive adenocarcinoma and thereby prevents stage IB and worse adenocarcinomas, it is substantially less good at detecting glandular than squamous precancerous lesions. It therefore often fails to prevent stage IA adenocarcinoma. Hence squamous carcinoma is much more likely to be diagnosed at an early stage. Upon review of cytology test originally reported as negative, 6.9% were upgraded to ?glandular neoplasia (Table 19). The percentage upgraded to ?glandular neoplasia is double (12%) when the test originally reported as negative is taken within 2 years of diagnosis (Table 20).

Among cases with known histological type, there was no increase in the proportion of cancers diagnosed as adenocarcinoma between April 2007 and March 2019 (on average about 20%). Suggesting that the proportion of cancers in each histological type has remained constant over time (Figure 6).

Figure 5. Proportion of cervical cancer cases diagnosed between April 2016 and March 2019, by histology and FIGO stage

Figure 6. Trends in the proportion of cervical cancer cases, by histological type

9. Treatment of invasive cervical cancers

Figure 7 shows the distribution of treatment for cervical cancer, according to age (see also Appendix B, Table 10a). The most aggressive treatment employed in each case has been captured. Treatment was recorded in 4,926 cases (77.3%), and of these, the most common treatment was radiotherapy plus chemotherapy with or without a hysterectomy (29.7%), followed by cone biopsy or ‘loop’ excision (28.6%), and simple or radical hysterectomy (23.4%). Only 2.4% of cases were treated with a trachelectomy (which is a surgical treatment for cervical cancer intended to preserve fertility), this procedure is almost always performed in women under age 35 (Figure 7). 5.5% of known treatments were recorded as ‘none’. The proportion of cases with known treatment in the dataset has increased from 54.3% in the first report to 77.3% here (Table A-2b, Appendix A).

Age is an important consideration when deciding what treatment to offer people with early stage cancer (FIGO stage I). As people age, they become less likely to have fertility-sparing treatments for early stage cancer. However, for stage II or worse cancers, chemotherapy plus radiotherapy with or without a hysterectomy is the most common form of treatment regardless of age at diagnosis.

Filtering the results by age reveals that for people aged 50 to 64, the most common treatment was chemotherapy plus radiotherapy with or without a hysterectomy (34.7%), followed by hysterectomy alone (18.9%) (Table 11, Appendix B). By contrast, 33.8% of people under 50 had fertility-sparing treatment (cone biospy or loop excision, or trachelectomy) with only 20.3% undergoing a hysterectomy (simple or radical).

For people aged 65 to 79, chemotherapy plus radiotherapy with or without a hysterectomy (40.4%) was the most common treatment, followed by radiotherapy with or without a hysterectomy and no chemotherapy (14.2%) (Table 11, Appendix B). However, among people for whom treatment was recorded, 12.9% of people aged 75 or over reportedly received no treatment and 13.5% received palliative care (Figure 7). Given the substantially poorer relative survival of elderly cervical cancer patients nationally,(8) this appears to warrant further investigation. It should be borne in mind, however, that ‘no treatment’ may be recorded because a record of treatment was not found, rather than because the patient was not treated.

Figure 7. Proportion of cervical cancer cases by treatment received, by age at diagnosis

Figure 8 shows the distribution of treatment received for invasive cervical cancer by stage of disease and age at diagnosis. This graph shows that the majority of people diagnosed with FIGO stage IA1 cancer received cone biopsy/loop excision treatment, in particular those aged under 35 (89%). Loop excision is also the preferred treatment method for high grade abnormaliaties of the cervix (cervical intraepithelial neoplasia grade 2 or 3) detected through screening.

Fertility-sparing trachelectomy was mostly received by people with stage IA2 or IB1 cancers who were under age 50. The majority of people diagnosed with stage II or worse disease at any age received chemotherapy plus radiotherapy with or without a hysterectomy. It is worth noting that a very small proportion (1.5%) of cancers diagnosed at stage IA1 are recorded as having been treated with chemotherapy plus radiotherapy with or without a hysterectomy. Similarly, 1.7% of stage II or worse cancers are recorded as having been treated with a cone biopsy. These are clearly misclassifications of stage and/or treatment which we are endeavouring to correct for the future.

Figure 8. Proportion of cervical cancer cases by treatment received, by FIGO stage and age at diagnosis

10. Effect of HPV triage on diagnosis of cervical cancer

Roll out of HR-HPV testing to triage people who have a cytology sample showing borderline nuclear changes or low grade dyskaryosis, and HPV test of cure for all people treated for CIN with negative, borderline or low grade dyskaryosis, began in April 2012. People who are HR-HPV negative are returned to routine recall, but those who are HR-HPV positive are referred to colposcopy (people with high-grade cytology 6 months after treatment are referred immediately to colposcopy, without this additional HR-HPV test). (9,10)

By April 2013, most laboratories had rolled out HPV testing to triage people who have a cytology sample showing borderline nuclear changes or low-grade dyskaryosis. For the time period covered by this report, all eligible people will have had the opportunity to be routinely screened under the HPV triage and test of cure protocol. Following on from the initial findings reported in the 2013 to 2016 audit report, here we present the impact of HPV triage on

• the proportion of cytology tests that result in an early recall
• time between the borderline or low-grade dyskaryosis cytology and diagnosis
• FIGO stage by HPV triage status

The proportion of cytology tests that resulted in early recall decreased from 20% (836/4,105) to 5% (77/1,423) among people with cervical cancer and from 19% (1933/9,915) to 6% (87/1,414) among people without cancer (Figure 9). Among people without cancer the proportion of cytology tests with an early recall halved by 2014 and is now three-fold lower suggesting the HPV triage and test of cure protocol has significantly reduced the number of women needing early recall. This has the benefit of reducing the anxiety people experience with early repeat testing.

1,444 people with cervical cancer (5% of all cancers) had a borderline or low-grade cytology result within 2.5 years of diagnosis. Figure 10 presents the time from this test to diagnosis among these 1,444 people. As of April 2019, the proportion of people diagnosed within 3 months of a borderline or low-grade cytology increased from 20% of those with a test between April 2007 and March 2012 to 42% of those with a test between April 2016 and March 2019. Before HPV triage people would have returned 6 months after the first borderline or low-grade test for another cytology, at which time if the cytology was still abnormal, they would be referred. Hence HPV triage has ensured a timelier route to diagnosis.

Of the 1,444 people with a borderline or low-grade cytology, 570 (39%) had an HPV triage test within 2.5 years of diagnosis. A higher proportion of cancers were diagnosed as stage IA among people with an HPV triage test than among those with low-grade disease and no HPV triage, Figure 11. This suggests that quicker referral for further investigation has improved stage at diagnosis and therefore future prognosis.

Figure 9. Proportion of cytology tests that resulted in an early recall action code, by calendar year taken, for cases and controls

Figure 10. Time from first borderline or low-grade dyskaryosis test to diagnosis, by year the cytology test was taken: percentage of people with cervical cancer and a test within 2.5 years of diagnosis

Figure 11. Estimated proportion of cervical cancer cases by FIGO stage, by HPV triage status among those with a borderline or low-grade cytology within 2.5 years of diagnosis

11. Screening history of cases and controls

Screening histories in people diagnosed with cervical cancer and their controls (people without cervical cancer)

Figure 12 shows the distribution of screening histories in people diagnosed with cancer and their controls (who did not have cancer) by date of diagnosis and stage. Tests taken within 6 months of diagnosis are not considered when establishing peoples’ screening histories. An individual diagnosed within 6 months of diagnosis is ‘screen detected’ but if they had no other tests, they should not be considered to have a ‘screening history’. To compensate for this, people are considered to be up to date with screening if they have a test between 6 months and 3.5 (if age 25 to 49) or 5.5 years (if age 50 to 64) prior to diagnosis. We present the data in 2 time periods: April 2007 to March 2010 and April 2016 to March 2019.

Overall, between April 2016 to March 2019 the proportion of people diagnosed with cervical cancer who were up-to-date with screening was statistically significantly lower among cases than controls (18% to 24% compared with 62%).

The proportion of people without cancer who have no screening history has remained at around 16% since the start of the audit in 2007, while the proportion screened and up-to-date (according to current guidelines) has remained at around 60%. These statistics are based on data included in the audit and are not directly comparable to national coverage statistics (which are calculated differently).

The proportion of people diagnosed with cancer who are up-to-date with screening has decreased over time, and a larger proportion have no screening history (other than within 6 months of diagnosis). Among people diagnosed with stage IA cancers between April 2007 and March 2010, 22% had no screening history compared with 39% of those diagnosed between April 2016 and March 2019, Figure 12. In other words, the overall proportion of people diagnosed with stage 1A cancer who have no screening history has doubled over this 10 year time period. Similarly, among those diagnosed with stage IB cancers, the proportion with no screening history increased from 22% to 35% over the same time period. Among those with stage II or worse cervical cancer, the proportion with no screening history increased slightly (35% to 39%).

Figure 12. Screening history for cervical cancer cases (by FIGO stage) and controls, by date of diagnosis: proportion (excluding tests taken within 6 months of diagnosis)

There was evidence that the proportion of people aged 50 to 64 with no screening history (regardless of stage at diagnosis) has decreased over time. However, the proportion of people aged 25 to 34 and 35 to 49 with no screening history has increased over time (Figure 13). The increase is more evident for stage IA and IB cancers (around 25% and 12% respectively). Nevertheless, an 18% (10% at ages 35 to 49) increase is also seen for stage II+ at ages 24 to 34.

A closer look at individuals with no screening history (Figure 14), reveals a test within 6 months of diagnosis among the majority of FIGO stage 1A and 1B cancers among people aged 25 to 34. Among older individuals the proportion of cancers diagnosed following a test decreases. In particular, over 77% of individuals aged 35 to 64 with a FIGO stage II or worse cervical cancer have no tests prior to diagnosis.

Figure 13. Proportion of cervical cancer cases with no ‘screening history’, by FIGO stage, date of diagnosis and age at diagnosis

Between 2008 and 2009 the age of first cervical screening invitation in England was increased from 20 to 25. This reflected evidence that screening at ages 20 to 24 provides no population benefit in terms of cancer prevention. In 2012, the age of sending out the first screening invitation was changed again to 24.5 years to enable people to be screened by their 25th birthday.

This change in policy has resulted in an increase in the proportion of people aged 25 to 34 who attend their first screening test between the ages of 25 and 26, from 6.5% to 37.0%. For people aged 25 to 34 who are diagnosed with cancer, the proportion who are aged 25 to 26 at diagnosis increased from 9.2% to 33.1%.

Among people with no screening history who are aged 25 to 26 there has been no statistically significant change in the proportion who had a test within 6 months of diagnosis (88% (120 out of 137 people) between 2007 and 2010, 91% (378 out of 414) of those diagnosed between 2010 and 2013, 89% (345 out of 387) between 2013 and 2016 and 89% (203 of 226) of those diagnosed in the latest period (p=0.585)). In other words, although the numbers of people diagnosed aged 25 to 26 have doubled, it is still the case that most (88%) have their cancer detected because of their first screening test.

Figure 14. Cervical cancer cases diagnosed between March 2016 and April 2019 with no screening history, by age and FIGO stage: proportion with a test within 6 months of diagnosis

To explore this further, we looked at the proportion of people with cervical cancer with no screening history, by date of diagnosis and age (Figure 13). Further we assessed whether the odds ratio (OR) of being diagnosed with cervical cancer among those who have no screening history has changed over the 2 time periods, Table D.

Looking more closely at cancers in people aged 25 to 34 we found that the number of people with stage IA cancer in this age group was similar over the 2 time periods (1,035 and 1,056 people respectively). However, the proportion with no screening history other than within 6 months of diagnosis increased from 27% between 1 April 2007 and 31 March 2010 to 51% between 1 April 2016 and 31 March 2019 (Figure 13). Similar results were observed for stage IB cancers in people aged 25 to 34 over the same time periods (593 and 583 people respectively).

In people aged 25 to 34 the number of people with stage II cancer has increased by 42% over time (169 and 240 respectively). This is reflected in an increase in the proportion with no screening history other than within 6 months of diagnosis from 39% between 1 April 2007 and 31 March 2010 to 58% between 1 April 2016 and 31 March 2019. The majority (58%, Figure 14) have no tests prior to diagnosis suggesting a symptomatic route to diagnosis.

Among those diagnosed with cervical cancer aged 35 to 49, there has been a much more moderate increase in the proportion with no screening history, and no change in the total number of cancers over the 2 time periods (2,188 and 2,103 people respectively).

For people diagnosed at age 50 to 64, there has been a decrease in the proportion never screened with a concomitant increase in the proportion attending at longer intervals than recommended. A slight increase in the number of cancers diagnosed over the 2 time periods (943 and 1,185 people respectively) was also observed.

15.8% of individuals of screening age (858/5,427) diagnosed with cervical cancer have never had any screening tests at all. Of these, 50% (429 people) had a late stage diagnosis of stage 2 or worse. Of people of screening age with a stage 1B or worse cancer, almost two thirds (62.6%) had not been screened for at least 7 years prior to their diagnosis, compared with only 21.8% of people without cancer (Table 17).

Compared to people who attend screening every 3 years, the likelihood of being diagnosed with cervical cancer among those who have no screening history has increased over time at all ages and stages, Table D. Since screening has improved over time (through better quality assurance and more sensitive screening tests) the risk of cervical cancer among those previously screened as recommended has become lower increasing the odds among those who have no screening history. 

Overall, these findings mean that taking action to encourage individuals to attend for cervical screening should continue to be a high priority. Without cervical screening the number of cervical cancers diagnosed each year would be significantly greater. Currently 62% of cervical cancers are diagnosed at stage 1A or 1B. Diagnosing cervical cancer at an early (rather than late) stage should be considered a success as survival is over 98% and, in most cases, fertility can be preserved.

Table D. Odds ratio of cervical cancer diagnosis in people with no history of cervical screening compared to those attending screening as recommended (every 3 or 5 years) by diagnosis period

12. Deprivation levels among cases

For the first time, we explored deprivation level among cases using the English indices of deprivation in quintiles. A higher proportion of young individuals were in the most deprived quintile (39% of those under 25 years and 28% of those aged 25 to 49) compared with individuals aged 65 or older (21%), Table 22.

Deprivation by FIGO stage at diagnosis showed no clear trend when all individuals were considered (Table 23), however when restricted to people aged 25 to 64 a higher proportion of individuals with FIGO stage 2 or worse cervical cancer were in the most deprived quintile (p=0.04), Figure 15.

Individuals aged 25 to 64 in the lowest quintile of deprivation were more likely to never have been screened and those in the least deprived quintiles more likely to have been screened as recommended prior to diagnosis (p<0.001), Figure 16.

These audit results show the importance of initiatives to increase awareness of cervical screening and support participation in deprived populations.

Figure 15. Cervical cancer cases diagnosed between March 2016 and April 2019 aged 25 to 64 years, by deprivation quintile and FIGO stage

Figure 16. Cervical cancer cases diagnosed between March 2016 and April 2019 aged 25 to 64 years, by deprivation quintile and screening history

13. References

1.         NHS Cervical Screening Programme. Audit of Invasive Cervical Cancers. Sheffield 2006 [updated December 2006]. Available from: https://www.gov.uk/government/publications/cervical-screening-auditing-procedures.

2.         National invasive cervical cancer audit. September 2021.  Available from: https://www.gov.uk/government/publications/cervical-screening-auditing-procedures

1.         Public health England. Cancer registration statistics, England, 2018 (full release). Published 29 May 2020. Available from https://www.gov.uk/government/statistics/cancer-registration-statistics-england-2018-final-release

2.         NHS Cervical Screening Programme. The Aetiology of Cervical Cancer. Sheffield; 2005.

5.         Office for National Statistics. Cancer Statistics Registrations, England (Series MB1): 2015 to 2017 [Available from: https://www.ons.gov.uk/search?q=Cancer+Registration+Statistics%2C+England%2C+2014.

6.         Falcaro M, Castañon A, Ndlela B, Checchi M, Soldan K, Lopez-Bernal J, Elliss-Brookes L, Sasieni P. Lancet. 2021 Dec 4;398(10316):2084-2092.

1.         Castanon A, Landy R, Sasieni PD. Is cervical screening preventing adenocarcinoma and adenosquamous carcinoma of the cervix? Int J Cancer. 2016;139(5):1040-5.

8.         Cancer Research UK. CANCERSTATS: Cancer survival statistics by age, 2009. 2010 [Available from: http://info.cancerresearchuk.org/cancerstats/survival/age/.

1.         The British Society for Colposcopy and Cervical Pathology. HPV TESTING INTRODUCED TO THE NHSCSP 2013 [Available from: https://www.bsccp.org.uk/healthcare-professionals/professional-news/hpv-testing-introduced-to-the-nhscsp-1/.

10.       NHS Cervical Screening Programme. screening programme algorithm for HPV triage and test of cure algorithm 2013 (withdrawn April 2016) [available from: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/514429/hpv-triage-test-flowchart.pdf]

14. Glossary

Action Code	This field (downloaded as part of the screening history from NHAIS) denotes the action to be taken in response to the result of each cytology test. The codes are: A. Routine screening/call and recall. H. Result recorded, but no change in current action code. (This code is normally used when privately taken cytology tests are entered into the system; these tests are no longer entered into the screening system but are included historically). R. Early recall at an interval specified by the laboratory. S. Suspend recall pending referral.
Cases	People diagnosed with invasive cervical cancer in England.
Controls	People who have not been diagnosed with cervical cancer, who are registered with a GP in England.  They are matched by age and place of residence with a case.
Cervical Screening Audit and Disclosure Management Group	The Audit and Disclosure Management Group provides multi-disciplinary advice into the audit and its associated guidance
Confidence Interval	Confidence interval is a term used in inferential statistics that measures the probability that a population parameter will fall between 2 set values. The confidence interval can take any number of probabilities, with the most common being 95% or 99%.
Exeter call and recall system (NHAIS)	The system used to invite people for screening. Since 1988, it has stored screening records for all people registered with a GP
FIGO stage	The cancer staging classification developed by the International Federation of Gynaecological Oncologists (I, IA, IA1, IA2, IB, IB1, IB2, III, IIIA, IIIA, IV, IVA, IVB).
Cervical Screening Programme Lead (CSPL)	The named individual within each NHS trust who is responsible for collating cases of invasive cervical cancer and initiating the audit process.
Screening Quality Assurance Service (SQAS)	In April 2013, quality assurance of screening programmes in England became the responsibility of Public Health England (PHE). This resulted in the 9 Quality Assurance Reference Centres (QARCs) in England being reorganised into 4 regions rather than 9 and renamed the Screening QA Service (SQAS). On 1 October 2021, the SQAS became part of NHS England