The GMS contract’s cancer indicators underline the importance of the GP’s role in providing support and coordinating care for these patients, says Dr John Donald

For many years cancer has been a major clinical priority in the UK in both primary and secondary care. This is borne out by the huge resources that have been put into diagnosis, treatment and follow up and in reducing waiting times for referral for cancer patients in all four countries.

The GMS contract acknowledges that most of the active management of cancer patients occurs in the secondary sector and that the indicator set is not evidence based (Table 1, below).1

Table 1: Clinical indicators for cancer
indicator no
Clinical indicator Points Qualifier Preferred
Read code
Exception reporting and suggested Read code Payment
Cancer 1
The practice can produce a register of all cancer patients diagnosed after 1 April 2003 6 This must be kept updated.
All cancers should be included except non-melanotic skin lesions

Exception reporting, cancer quality indicators 9h8..
Excepted from cancer quality indicators:

  • patient unsuitable 9h81
  • informed dissent 9h81

Non-melanotic skin cancers: 9h81.

Ongoing management
Cancer 2

The % of patients with cancer diagnosed from 1 April 2003 with a review by the practice recorded within 6 months of confirmed diagnosis. This should include an assessment of support needs, if any, and a review of co-ordination arrangements with secondary care 6 Should be recorded within 6 months of confirmed diagnosis 8BAV (was 8CLO)   25-90%

However, primary care provides a major supportive and co-ordinating role once diagnosis has been established, and to do this it is essential that each practice knows which of its patients have cancer. Therefore, the main emphasis of the indicators is on setting up a disease register and regularly reviewing previously diagnosed patients.

Although a disease register is common to all indicator sets, the methodology for finding the patients to populate a register differs for each one.

As far as review is concerned, it is expected that each patient on the register is reviewed within 6 months of diagnosis with the emphasis on support needs and coordination of care with the secondary sector.

Primary care obviously has a key role in the early recognition of danger symptoms and in diagnosis and early referral but it would not be easy to design indicators to measure this. In addition, the indicators do not reflect the necessity of accurately coding patients who are at risk as a result of their family history, for example, those with first-degree relatives with breast cancer or relatives with familial polyposis. From this, the indicators only demand what most would accept as minimum care for this patient group.

The cancer register

Practices wishing to tackle the clinical indicators for cancer may find themselves in one of the following situations:

  • The practice already has complete or near-complete clinical Read coding
  • The practice has little or no Read coding
  • The practice wishes to do more than just meet the basic contract requirements.

Complete or near-complete Read coding

It is expected that each GP computer system supplier will provide search templates, including exception codes, for each quality indicator. However, the real work will be in identifying and coding patients within the practice who have cancer.

Useful starting searches will depend on how the practice has recorded cancer diagnoses in the first place, but searching for codes including .B...,, BB.., B0-B6, Byu.. will fill in some gaps.

Proxy drug searches for cancer patients will also not be very easy as more than 70 different drugs for cancer treatment are listed in the BNF. In addition, many of these will be prescribed and given in secondary care.However, it may be worth doing some searches based on the more common drugs such as tamoxifen and the anti-androgens.

In addition, the practice should ask all staff members, including extended staff such as community nurses, to contribute to a list of cancer patients. Fortunately, for an average-sized practice the numbers will not be large and it will be the more recently diagnosed patients who are the most relevant for the contract and the most easily remembered.

If your computer supplier has not provided search templates, it is worth coding all found cases with a single ‘generic’ cancer code such as .B... to help identify these patients more easily in the future.

It is important to include the exception codes at the time of case finding via computer searches and to record patients who are unsuitable for including in future searches. An important example of this is applying an exception code to all patients with non-melanotic skin cancers. A guide to general exceptions can be found in paragraph 3.30 of the supporting documentation.

Little or no Read coding

In Scotland, practices can obtain Scottish cancer registry records for their registered patients who have cancer.2 Practices in England and Wales could try approaching their local regional cancer registry to see if they can obtain this information.The problem with this approach is that the central information may not be up to date, and to be useful it must be correct as at 1 April, 2003. Otherwise, the methods of case finding are the same as those described above.

Practices that wish to do more than just meet the contract

This situation is beyond the scope of this article. However, once patients are identified it is worth coding and recording as much detail as possible about the exact site and nature of the cancer – see the ‘morphology’ codes under ‘.B...’ of the Read code dictionary.

At the same time, it is worth adding codes for chemotherapy (8BAD0), radiotherapy (5151), metastatic cancer (BB13.) and palliative care (ZV57C). Recording these codes at the time of the original case finding will prove valuable when applying the second (review) cancer indicator.

Updating the register

If the register is to continue to be a useful tool and to earn quality points under the new contract, it must be kept up to date. Practices should ensure that they have a system for correctly coding any new diagnosis of cancer or significant treatment for patients with cancer.

This is generally done via letters received from the clinics; the diagnoses are highlighted by the GP and then passed to administrative staff for coding. A list of the relevant cancer Read codes should be produced for staff to refer to, to ensure that codes are accurately recorded.

The register can be verified by comparing national prevalence rates for all cancers for all ages against national prevalence statistics. For example, in Scotland, the national prevalence rate for all cancers for all ages and both sexes is 2.41%, which for an average practice of 5000 patients would give around 100 cases.2 Practices with a higher than average age band would expect to find more cases.

Reviewing identified patients

Cancer indicator 2 relates to the percentage of patients with cancer, diagnosed from 1 April 2003, who have had a review by the practice recorded within 6 months of confirmed diagnosis. The review should include an assessment of any support needs and a review of co-ordination arrangements with secondary care.

The contract guidance admits "The IT solution for this indicator is not straightforward because it is difficult to tie the date of review to the date of diagnosis in computer searches.”1 It suggests that practices could search for patients added to the register between 6 and 12 months previously and undertake a manual search of notes to see if a cancer review has been held.

The documentation also suggests an alternative: to search for patients added to the cancer register between 6 and 12 months previously who also have an 8BAV coding (not 8CLO as originally stated in the contract guidance) in the past 12 months.

While this search will not produce data that directly meet the indicator, it will identify a cohort of cancer patients and whether a cancer review has taken place up to 12 months after diagnosis (depending on when the diagnosis was made). A further difficulty with this indicator is that this code should have been recorded already from October 2003 (6 months after April 2003) for patients diagnosed after 1 April 2003.

Essentially, practices should apply the code for ‘Cancer diagnosis discussed’ (8BAV) to every situation where a cancer patient is encountered by any staff member, so all staff need to be alerted to this requirement. Alternatively, the practice could organise regular multidisciplinary cancer meetings to keep all staff updated about their cancer patients. At review, practices could take the opportunity to elicit any family history of cancer and codes inserted into the records of any family members who are also looked after by the practice.3

This indicator may be verified by randomly selecting a number of patients’ case records in which a review has been recorded, to confirm that an assessment of support needs and review of co-ordination arrangements with secondary care have been undertaken and recorded.

The South East of Scotland Cancer Managed Clinical Network (SCAN) has produced two excellent guides, one of which is a summary the other very detailed, to setting up cancer registers and patient review.4


  1. Investing in General Practice:The New General Medical Services Contract. Supporting documentation.
  2. Information and Services Division, Common Services Agency, Scotland.
  3. Rose P,Murphy M,Munafo M et al.Improving the ascertainment of families at high risk of colorectal cancer: a prospective GP register study. Br J Gen Pract 2004; 54: 267-71.

Guidelines in Practice, June 2004, Volume 7(6)
© 2004 MGP Ltd
further information | subscribe