Early Diagnosis

Early Diagnosis of Cancer

While it is known that the earlier that cancer is diagnosed, the better the outcome for the patient, the process of symptoms, consultation with GP, referral to hospital and diagnosis is a complex one, and it is not clear how to achieve earlier symptomatic diagnosis of cancer. Research within the Policy Research Unit on this subject includes:

  • Studies of the predictive power of symptoms or patient histories reported at primary care for various cancers.

    These involve retrieving large amounts of data from general practice databases, to find symptoms and histories which arise at GP consultations in the months before a diagnosis of cancer. The results are used to build up a series of warning signals for GP's to refer patients whose symptoms are associated with certain cancers. Results have already strongly influence NICE guidelines for referral for further investigation.

  • Estimation of the effect of factors such as age, ethnic group and socioeconomic status on GP's referral habits.

    This was done via virtual consultations on the internet. GP's viewed actors playing patients with various symptoms and demographic features. We found that older patients and black patients were less likely to be referred for further investigation. An online education tool has been developed as a result, and this is being trialled in three medical schools.

  • Identification of factors associated with late or emergency presentation with cancer.

    This involves a major data linkage exercise between general practice, cancer registries and other sources. Finding those factors which are strongly associated with late or emergency presentation, will inform policies to avoid these in the future by promoting earlier diagnosis.

  • Possible new markers of cancer in primary care 

    We found that thrombocytosis was associated with subsequent diagnosis of several types of cancer. Work is ongoing to ascertain how the use of platelet counts could promote diagnosis of cancer at an earlier stage.

  • Abdominal pain presenting in primary care

    We plan a large study based on primary care records of abdominal pain in association with other features, and likelihood of cancer.

  • Clinical decision support tools

    We are working on a qualitative study of use and acceptance of a clinical decision support computer aid for upper GI cancer including action taken as a result of the aid, and follow-up interviews 6 months after use of the tool.



Diagram: Quantification of the risk of uterine cancer in symptomatic primary care