Intended for healthcare professionals


Is the two week rule for cancer referrals working?

BMJ 2001; 322 doi: (Published 30 June 2001) Cite this as: BMJ 2001;322:1555

Not too well

  1. Roger Jones, Wolfson professor of general practice (roger.jones{at},
  2. Greg Rubin, professor of primary care,
  3. Pali Hungin, professor of general practice
  1. Guy's, King's, and St Thomas's School of Medicine, 5 Lambeth Walk, London SE11 6SP
  2. University of Sunderland, Sunderland SR2 7BW
  3. University of Durham, Durham DH1 3HN

    Mortality rates in the United Kingdom for several cancers compare unfavourably with those in other countries,1 and this may be explained at least partly by British patients having more advanced disease at the time of treatment than their European counterparts. Morbidity and mortality can be reduced through primary and secondary prevention, including screening; by early detection; and by prompt and effective treatment. Last year, in the context of the NHS's disgracefully long waiting lists, the government pledged that patients with suspected cancer would be seen by a specialist within two weeks of referral by their general practitioner. The aim was to reduce delays between presentation, diagnosis, and treatment. The two week rule has now been implemented across a range of specialties, supported by widely distributed information about risk factors and criteria for rapid referral for suspected cancer. How is it working?

    Local rapid access referral mechanisms have been established, including specially designed forms and direct electronic access to outpatient appointments for patients meeting prespecified criteria. In addition, the government established the cancer services collaborative programme, creating nine cancer networks covering about 15 million people, aimed at optimising systems of care and improving the experience and outcomes of care for patients with suspected or diagnosed cancer.

    The two week rule should reassure patients that suspicious symptoms will be investigated promptly and also help general practitioners with their referral decisions. There have been suggestions, however, that the proposals go beyond the evidence base. For example, there is no good evidence that delays measured in weeks in diagnosing colorectal cancer have an impact on staging or health outcomes,2 though delays of 3–6 months in breast cancer patients are associated with reduced survival.3 The presence of macroscopic haematuria, an urgent referral criterion, is a poor predictor of renal tract malignancy.4 In general practice the predictive values of symptoms suggestive of cancer are poorly defined. Some, such as the presence of an abdominal mass or weight loss in patients with gastrointestinal symptoms, are likely to be markers for inoperability, while modestly raised levels of prostate specific antigen, another referral criterion, have poor predictive value for prostate cancer.5 Although factors including age, family history, and change in bowel habit have been the subject of research, 6 7 their diagnostic utility remains unclear. Indeed, symptoms predictive of cancer are now the subject of a major call for research from the NHS research and development programme.

    In April this year the health ministers announced substantial reductions in delays before referral, achieved by redesigning referral mechanisms, developing protocols for “one stop” investigation and treatment, and supporting more effective multidisciplinary team care. However, several abstracts presented at the recent scientific meeting of the British Society of Gastroenterology suggested that, at least for urgent gastroenterological referrals, the two week waiting standard is being met at the expense of a substantial increase in the waiting time for routine referrals, while not necessarily identifying treatable causes of cancer. Two separate audits undertaken in district general hospitals found that only one third to a half of patients with colorectal cancer were referred through the two week arrangements, 8 9 one of which also identified a doubling in routine waiting time to 64 days.8 A retrospective study of patients with colorectal cancer admitted with no referral guidelines to a single surgeon found that about 30% of cases would not have met the referral criteria.10 Fast track systems for suspected bowel cancer can work, but their introduction needs to be supported by increased resources, particularly for investigation, staging, and treatment.

    Hospitals are now required to provide feedback to general practitioners about the working of the two week rule, and a major emerging issue relates to the appropriateness of the guidelines themselves. It is becoming clear that patients can meet the national guidelines and still be an “inappropriate” referral in the opinion of the consultant, while patients who do not meet the guidelines can present with a high suspicion of cancer, appropriate for urgent evaluation. In one trust in the south of England a recent audit has indicated that about 66% of referrals for breast, skin, and lower bowel problems were appropriate, while 80-100% of referrals for suspected gynaecological, upper gastrointestinal, lung, and urological malignancies were judged appropriate (J Price, personal communication). Information about the opportunity costs, in terms of delayed investigation and treatment for less urgent cases, will emerge more slowly.

    In the absence of highly sensitive and predictive symptom clusters for diagnosing cancer, the existing referral criteria establish de facto risk thresholds for rapid referral and assessment. The positive predictive value for colorectal cancer of persistent rectal bleeding and change in bowel habit, for example, is about 30%. As a result, existing guidelines do not benefit the significant numbers of patients with lower risk symptoms who nevertheless will prove to have cancer and may now experience longer waiting times than before. Those not immediately referred might have a less well defined clinical picture but, paradoxically, might harbour malignancy at an earlier stage, with better prospects of cure. This situation is unlikely to match the public's expectation of the NHS as investigating any symptoms that could indicate cancer and is inconsistent with government policy elsewhere which advocates healthcare interventions at much lower levels of risk—for example, in the primary prevention of coronary heart disease.

    We must hope that this initiative will result in genuine clinical gain and improved survival. It will require further refinement through high quality research on the predictive values of symptoms, symptom clusters, and other risk factors in patients presenting in primary care. Careful audit of the criteria and methods for urgent referral will be needed, together with more imaginative use of information technology and the continuing development of multidisciplinary care of patients with cancer across the primary-secondary care interface.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    View Abstract