Abstract
BACKGROUND: Awareness of detail in ambiguous complaints may help GPs suspect cancer when a malignancy is present. AIM: To study the contribution of symptoms and patient characteristics to GPs' suspicions of cancer being present, and to what degree these suspicions were confirmed. DESIGN AND SETTING: Prospective cohort study of patients in 283 rural and urban general practices throughout Norway. METHOD: During patient consultations (over a period of 10 days) GPs registered whether there was a suspicion of cancer when a patient presented with at least one of seven focal symptoms and three general symptoms commonly considered to be warning signs of cancer. Follow-up questionnaires were sent to GPs 6-7 months later, requesting information on any subsequent diagnosis of cancer in these patients. RESULTS: Out of 51 073 patients, 6321 presented with warning signs of cancer; of these, 106 had a subsequent cancer diagnosis. Of the patients presenting with warning signs, 1515 (24%) patients were suspected of having cancer; this was correct for 3.8% of suspected cases. Of the 106 patients diagnosed with cancer who presented with warning signs, cancer was suspected in 58 (54.7%). GPs' correct cancer suspicions were six times more frequent than their erroneous lack of suspicion. Multiple symptoms, previous cancer, comorbidity, and multiple consultations increased the probability of cancer, but only multiple symptoms and previous cancer increased suspicion. Suspicion led to an increase in the number of diagnostic procedures undertaken. The proportion of cancer cases where GPs recorded a lack of suspicion was relatively small, but important. CONCLUSION: Selected symptoms appropriately resulted in GPs suspecting cancer. Comorbidity and multiple consultations were underestimated by GPs as factors associated with cancer. Cancer suspicion should rely on symptoms in combination with other relevant information.