Symptoms that predict chest X-ray results suspicious for lung cancer in UK primary care: results from a prospective study

Brindle, Lucy, Hamilton, W., Banerjee, Anindo and Dowswell, G. (2014) Symptoms that predict chest X-ray results suspicious for lung cancer in UK primary care: results from a prospective study. Winnipeg 2014 - 7th meeting of Ca-PRI, Winnipeg, Canada. 10 - 13 Jun 2014.

Abstract

Background/introduction:?Predictive values of lung cancer (LC) symptoms that inform UK LC referral guidelines have been calculated from GP records and databases, with the potential for recording bias by GPs. The Identifying Symptom Predictors of Chest and Respiratory Disease (IPCARD) self-completion questionnaire was designed, for use in prospective studies, to obtain accurate positive predictive values (PPVs) of patient-elicited potential LC symptoms. The IPCARD Feasibility Study provides the first prospectively collected, patient-elicited comprehensive symptom data to identify PPVs of potential LC symptoms in UK primary care.

Research question:?Which patient-elicited symptoms predict chest X-rays (CXR) suspicious for LC in a GP-referred CXR population (LC incidence 1%)?

Methods:?GP-referred CXR attendees (1414) at three UK sites completed IPCARD before CXR; LC diagnosis was obtained 6 months post-CXR. Multiple logistic regression was used to calculate PPVs of symptoms for abnormal CXR, adjusting for age and sex; and stratifying by smoking status and COPD.

Results:?Common chest symptoms – cough for longer than 3 weeks, generic chest aches/pains, and breathlessness – did not predict suspicious CXR. Weight loss and less common variants of chest pain (pain in side of chest/ribs, severe pain, and pain that ‘feels like indigestion – not associated with eating’ in patients with non-progressive/less severe pain) predicted CXR suspicious for LC in this high risk, referred population.

Discussion/conclusion:?Common chest systems, identified as referral criteria by NICE, and included in UK ‘Be Clear about Cancer’ public awareness campaigns, although potentially predicting LC in a lower risk pre-referral population, did not predict CXR suspicious for LC in this referred population with higher rates of non-malignant chest and respiratory disease. The possibility that weight loss, and variants of chest pain, might also predict LC in pre-referral primary care populations with chronic respiratory disease will be investigated in ongoing studies using IPCARD.

This record has no associated files available for download.

Contributors

Author: Lucy Brindle

Download statistics