CANDID

CANcer DIagnosis Decision rules

Key areas of concern for both doctors and patients in primary care when suspecting a patient has cancer is preventing delays in diagnosis, getting high risk patients referred first and keeping unnecessary investigations to a minimum. There have been few valid studies to assist decision-making in primary care, either to get a patient referred quickly or to assist in making sure an anxious patient is effectively reassured.

This study seeks to work out which of the symptoms and examination findings are the most effective in predicting lung or colon cancer. To decide the best clinical information to collect in the study we will interview patients and also get consensus from a group of experts. Then we will recruit 20,000 patients who consult their GP - half with respiratory symptoms and the other half with low bowel symptoms.

Clinical information will be collected using standardised internet based forms. Willing patients will complete lifestyle questionnaires and provide blood samples (including for genetic analysis). The National Cancer Registry will then be monitored to see which patients develop cancer, and statistical analysis will determine the most important clinical variables that predict cancer. The clinical prediction 'rules' or decision aids developed from these studies will then be tested with a further 2000 patients for each condition for validity.

CANDID is led by Professors Paul Little (University of Southampton) and Tom Fahey (Dublin) at the with additional recruitment Centres led by Dr Alastair Hay at the University of Bristol and Professor Richard Hobbs (University of Oxford), Professor Irwin Nazareth (University College London), Dr Evangelos Kontopantelis at the University of Manchester and Professor Danielle van der Windt at the University of Keele.