Wolfson Institute of Preventive Medicine

Search - ZOO
Search - K2
Search - Categories
Search - Contacts
Search - Content
Search - Newsfeeds
Search - Weblinks

National Cervical Screening Audit

Questions and Answers


Section 1 . Updates to the National Audit

Because of difficulties in drafting the new guidance prior to the end of March 2012 we propose a staged introduction of the changes outlined in the new protocol. This staged introduction will allow us to carryout three training events where the changes to the protocol will be explained. The training events are scheduled as follows: 18th of June in London, 25th of June in Leeds and 2nd of July in Birmingham.

Cases diagnosed before the 1st of April 2012

1. Use the old forms for cancers diagnosed up to 31/03/2012
2. If you have not yet initiated a slide review for a case diagnosed prior to 01/04/2012, there is no need to complete this review the back log
Cases diagnosed from 1st of April 2012 to the end of June 2012

1. Use the new forms (and database) to collect the following details: case details, Exeter download (cytology history), colposcopy and histology. These forms are available on the NHSSCP website.
2. No slide review for cancers diagnosed during this period is expected
Cases diagnosed from July 2012 onwards
1. Follow the guidelines as set out in the Protocol Changes to the Audit of Invasive Cervical Cancers 2012-13

Cases diagnosed from April 2013 onwards
1. Follow the guidelines as set out in the soon to be published: Protocol Changes to the Audit of Invasive Cervical Cancers 2013, mainly it involves the addition of colposcopy review.

2013 additions:
Definition of the role of the HBPC: it has taken us a year to put together this short outline of the role of the HBPC. We hope you find it useful.
Colposcopy review: will finally be incorporated into the audit and full guidance on how to carry out the review will accompany the updated protocol
Guidance on histology review: the histology review forms have been modify so that they make more sense and guidance on how to complete them will be included in the new protocol.
2012 additions:
AJ-CRUK Exeter download – several new fields have been added to the download to facilitate data collection. They include: full postal address including postcode (to facilitate creation of IMD), date of invitation and HPV result. These changes will be incorporated into the access database so that the file can be uploaded automatically as usual
Cytology review – this exercise will have a much stronger educational focus than in the past. There will be clear separation between audit as part of clinical disclosure to patients and national case-control audit. The changes should ensure the review process is more streamlined and less time consuming
Histology review – diagnostic samples and all those taken after such a sample need not be reviewed. Fields to be completed upon review will now follow the RCPath minimum dataset.
Essential Fields list – IMD and treatment have been added to the essential fields list. Cancer histology review has been removed.

All updates to the audit forms will be published on the NHS Screening website.

For QARC staff only:
Any new version of the Audit database will be uploaded onto the secure audit website,

The national colposcopy group have agreed on the guidance for colposcopy review and data collection form. This will be incorporated into the audit from April 2013.

Important changes to the cytology and histology review have been approved by the Audit Management Group. Details of the changes were published in the 2012 audit protocol.
A copy can be found at

Does the punch biopsy in the example below need to go to external histology review?
Women refered as a result of an abnormal test, has a punch biopsy (high grade CGIN) a month before diagnosis (adenocarcinoma on a LLETZ). Upon local review the punch biopsy was upgraded to cancer.
Current guidelines say "An external review should only be performed where an abnormality is detected that was not formerly reported and/or where earlier detection would have led to further clinical review or treatment in that clinical unit, rather than discharge of the patient back to the GP"
The audit management group agree that although an abnormality not formerly detected was reported the managment of this patient would not have changed substantially. Therefore no external review is needed.

Section 2. Who and what information should be included in the audit

Any woman who has entered the Audit should remain in it. It should be clearly documented on the histology review form that she was subsequently found NOT to be a case. In these women we would be especially interested in knowing what treatment she received (if any) and whether the "cancer" was screen detected. If the review was done so close to diagnosis that the woman was never entered into the audit (i.e. no screening histories have been obtained), then there is no point in entering her. If the woman has been registered with the Cancer registry, it would be interesting to know how the registries deal with these situations. It was felt that in general the woman should be informed that she did not in fact have cancer (and in most cases that her treatment would have been the same even if the correct diagnosis had been made originally).

We do not consider women not registered with a GP as being part of the National Screening Programme (they will not have been invited at regular intervals). There is no system in place for obtaining controls in these circumstances (and therefore no need to obtain them). Keeping a record of these women might be important in the context of what the minister said (all women have a right to have their case reviewed). In the context of the National Audit, it is important to know how many women diagnosed with cancer are in this situation so we can monitor what percentage of women diagnosed with cancer they represent. A simple list so that we can keep track of them is sufficient, including (where possible) their date of diagnosis, stage and histology.

No, cervical cancer reoccurrences should not be included in the audit

Yes, if the woman was registered with an NHS GP she should be included regardless of where the diagnosis was made.

If it is possible (and the private laboratory is willing) private histology samples should be reviewed according to guidelines

Yes, where possible private colposcopy appointments should be recorded.

Women in HM forces or in prison are at higher risk of developing cervical cancer than the general population, however they are not registered with an NHS GP and they do not receive standard NHS screening. Due to the effort involved in finding information on these women and the fact that they are not part of the NHS screening programme the audit management group recommended that they should not be included in the audit.
At the moment we recommend that QARCs keep a note of the year and age of the women at diagnosis for future reference, but they will not be included in the audit.

When the histology report is unclear and other supporting tests do not provide much clarification we suggest that the HBPC discusses the case with the clinical team and/or the MDT. If it still remains unclear, as a rough guide we say that if the case will be reported as a cervical cancer to the Cancer Registry we want it registered in the Audit.
When the diagnosis was not made on histology, please follow the Cancer Registry guidelines when determining the appropriate date of diagnosis.

Section 3. Running AJ-CRUK

It is possible to run AJ-CRUK with a dummy NHS number, but this will only be applicable to women who are registered with a GP. Otherwise there will be no screening record on Exeter.

Including women in prison into the screening programme has always been a problem. Although these women are offered screening and colposcopy the results are not recorded on Exeter. The National Screening Office is working on this problem but no solution has been found. In the context of the National Audit, it would be useful to know how many women diagnosed with cancer are in this situation so we can monitor what percentage of women diagnosed with cancer they represent. A simple list so that we can keep track of them is sufficient, including (where possible) their date of diagnosis, stage and histology.

Most QARCs come across this problem because the database will not let you enter duplications of NHS numbers. If you are not entering NHS numbers for controls it is very likely that you will never realise that a control has now become a case. If you become aware of this problem please make a note so that we can keep a record of how often it happens.

To re-run the AJ-CRUK job within the same PCT you need to contact Exeter directly and request that the flag be removed from the case and the controls selected the first time the job was run. Only then will you be able to run the job again.

If a case is run on AJ-CRUK in two different PCTs, the consequence will be two Study IDs and two set of controls for the same case. If the QARC does not identify this before entering the case into the database, we are usually able to identify duplications nationally whilst cleaning the data. We inform the QARCs immediately if this happens. It is not possible to create duplication within the same PCT unless specially requested from Exeter.

Section 4. Database queries and updates

The latest version of the database is V13, last updated April 2012. It can be downloaded from the secure website at any time
This version superseded version 12 - which had a problem dealing with postcodes that were not in the right format.

If these tumours are to be registered separately by the Cancer registry, then please enter the histological sample twice in the biopsy form and enter the review twice as if they were separate samples even if they were diagnosed on the same histological sample.

Please enter two colposcopic appointments with the same date and other details and just vary the procedure type.

As cities and towns grow and re-develop new postcodes are issued. This happens about once a year. We are not able to update the database on such a regular basis and therefore some postcodes will not be recognised and it will not be possible to obtain an Index of Multiple Deprivation.
Postcodes are updated everytime a new version of the database is released.

Section 5. Background to the audit

The essential fields reflect the minimum data which needs to be available in order for an Audit to take place. Failure to obtain these fields makes the data unsuitable for audit purposes. Over and above these fields, it is the QARC's decision how broad an area the audit will cover.
We expect changes to the essential fields list as the audit evolves.

The AJ-CRUK programme will automatically download the GP number (locally defined identifier for a particular GP) for each woman. If it is necessary to request additional screening details from a woman's GP, having the GP number will facilitate identification of the right GP. This field in not essential and should not be submitted to the National Audit.

If you would rather not provide the full date of birth we would ask you to provide the month and year of birth rather than age. It is difficult to ascertain at which time the age was calculated and this will lead to problems at the time of analysis.

It is a measure of mobility – a risk factor. It could provide an indication as to whether screening should be offered at the time a woman registers with a GP. During analysis it can be used to limit screening histories between a case and her control to the period they were both registered with the same GP. This field in not essential.

The index of multiple deprivation (IMD) can be obtained by entering a woman's postcode into the Access Database. It has ten levels and is based on the English Indices of Deprivation 2004 produced by the Office of the Deputy Prime Minister (ODPM), details of which can be found on their website (

These indices are based on Super Output Areas (new geography for the collection and publication of small area statistics) developed from the 2001 census. The first SOA data to be published are the ODPM Indices of Deprivation 2004. The plan is to create three SOA levels. At the beginning of 2005 only one level existed, lower layer SOA. Each lower layer SOA includes at least 1,000 people and 400 households.

ODPM Indices of Depravation 2004 rank super output areas from most to least deprived. In order to match the SOA to the postcode it was necessary to obtain data from the Economic and Social Data Service (ESDS). ESDS have a directory called All fields Postcode Directory, which links the postcodes to a Output Area (165,665 in England; 9,769 in Wales) and OAs to the lower layer SOAs (6,780 in England;413 in Wales).

Providing the SOA rank for each woman could potentially pinpoint her area of residence. We therefore grouped the ranks into deciles, which are displayed by the Access database. The SOA ranks included in each decile are listed in Table 1.

The end user licence for ESDS requires us to share the data with other ESDS registered users only. We must also let them know off all publications that include their data.

The postcodes in the database get updated everytime a new version of the database is released. Unfortunately postcode updates often happen more than once yearly and the database cannot be updated that frequently. Postcode updates also mean that the ranks included in each decile change slighly over time.


The Scottish Index of Multiple Deprivation 2004 are based on “data zones” and was created by the Scottish Executive. In the same way as with the English ranks I have used the “data zones” provided by ESDS to create a Scottish IMD for the database. A list of ranks included in each decile can be found in Table 1.


The Welsh Index of Multiple Deprivation is based on Electoral Division levels in 2000. This Index was put together by the Index Team at Oxford University. The Electoral Divisions are matched to the postcode in the same way as the English ranks. A list of ranks included in each decile can be found in Table 1.

Table 1. Deprivation Ranks included in each Decile

Ranks based on ODPM's Indices of Deprivation 2004 Access Database Deciles Ranks based Welsh Index of Multiple Deprivation 2000 Access Database Deciles Ranks based on Scottish Index of Multiple Depravation 2004 Access Database Deciles

min 1/max 32,482


min 1/ max 865


min 1/max 6,505






























































NOTE: indices are not comparable with each other.

Please use the UK Association of Cancer Registries Algorithm. (
Justification: This is used as a proxy for the date of diagnosis for analytical purposes, and to derive age at diagnosis (with birth date), again for use in analyses of trends and of survival. (This date was formerly known as "anniversary date").
Definition: The date of the first event (of the six listed below) to occur chronologically should be chosen as the incidence date.
Order of declining priority:
1) Date of first histological or cytological confirmation of this malignancy (with the exception of histology or cytology at autopsy). This date should be, in the following order:
date when the specimen was taken
or date of receipt by the pathologist
or date of the pathology report
Date of admission to hospital because of this malignancy.
When evaluated at an outpatient clinic only: date of first consultation at the outpatient clinic because of this malignancy.
Date of diagnosis, other than 1, 2 or 3.
Date of death, if no information is available other than the fact that the patient has died because of malignancy
Date of death, if the malignancy is discovered at autopsy.
- Whichever date is selected, the incidence date should not be later than the date of the start of the treatment, or decision not to treat, or date of death
- The choice of the incidence date does not determine the coding of "basis of registration"
- Registries must give precedence to an event of higher priority if it occurs within three months of an earlier or initially chosen date of lower priority
- For patients with CIN III or GIN III use the date of the biopsy confirming the diagnosis
- Cases with only a cytology date need careful review before registration and allocation of incidence date.

Please use the pre-treatment FIGO stage (as recorded at the MDT meeting). It will be possible to enter “IN” (greater than micro-invasive or 1B+) and X (unknown), however as a region you may insist that a more complete FIGO stage is obtained. Both the national forms and the database now allow for an updated FIGO stage to be entered at any time.

We are not collecting data from the Screening Classification requested in Appendix 2 because it is subjective and not validated. The national screening programme has been working on an update to this classification. The new classification will be objective and it will also be possible to use the audit database to classify cases automatically.

This section aims to collect treatment data to see if those women with FIGO stage 1A are more likely to receive fertility preserving treatment (a goal of the screening programme). As a broad national survey it will tell us of the variation in treatment given for each FIGO stage. For this section please use the most radical treatment a woman received after initial diagnosis and FIGO staging. However if initial surgery reveals a more extensive disease requiring further surgery, use the latter. We do not want to wait for a long follow-up for this section and we are only interested in the treatment received at first diagnosis and not in reoccurrence or secondary tumours.
If the diagnostic LLETZ/LOOP/CONE removed the cancer, please still report this as the treatment.
We have added the options "palliative care", "none" and "not known" which should eliminate confusion.

We are looking for up to 5 controls for each case. Two controls are selected independently of the screening histories. Additional controls are selected if the case had a history of abnormal smears or if she appears to be screen-detected. The idea of the “abnormal control” is to compare screening histories following the first abnormal smear. The “screen-detected” controls make allowance for the fact that a screen-detected case must have been for screening at least once.
GP Control (Study ID ends with 2): a woman with the same age (within one year of birth) and the same GP practice as the case.
District control (Study ID ends with 3): a woman with the same age and from the same screening programme, but a different GP practice as the case. This control is, where possible, matched on the first half of the post-code. The word “district” is a historical legacy – as with all controls, this woman comes from the same call-recall system as the case.
Adjusted screened control (Study ID ends with 4): when a case appears to be screen detected the computer selects a screened woman. It considers all the smears she had in the previous 5.5 years and randomly chooses one. If that smear is within six months of the cases’ index smear, the woman is included. It is thought this will reduce frequency bias of unadjusted screened controls (women screened more often are more likely to be selected as controls).
Abnormal control (Study ID ends with 5): when a case has a previous abnormal smear (<6m before diagnosis) the computer selects a control with an abnormal smear. The earliest abnormal smear in the case is identified and matching is based on the date and result of that smear.
Unadjusted screened control (Study ID ends with 6): similar to the adjusted screening controls, except that the control simply has to have a smear within six months of the case’s index smear.

Return to top