Diagnostic testing patterns for lung cancer-related symptoms in primary care: a retrospective cohort study.

OBJECTIVE

To examine diagnostic testing patterns for symptoms associated with lung cancer in Australian general practice.

DESIGN

Retrospective cohort study using a linked primary care database.

SETTING

Victorian general practices enrolled in the University of Melbourne's primary care data repository, Patron.

PARTICIPANTS

Deidentified patients aged 40 years and over who presented with symptoms associated with lung cancer between 2008 and 2022.

MAIN OUTCOME MEASURES: The proportion of patients who underwent pathology, imaging or referral or who were prescribed medication; differences across sociodemographic variables and the proportion of patients diagnosed with lung cancer.

RESULTS

Between 33% and 53% of patients underwent no investigations. Among those investigated, blood tests were most common for fatigue (83%) and anorexia (69%), while X-rays were more commonly performed for chest infections (40%) and haemoptysis (38%). Two-thirds of patients with haemoptysis were investigated, but only 22% were referred for a chest CT scan. When medications were prescribed, antimicrobials were often used, particularly for chest infections (79%) and haemoptysis (67%), while beta-2 agonists were primarily prescribed for shortness of breath (46%). Ever-smokers exhibited only marginally higher odds of being investigated (OR 1.05). Patients with higher socioeconomic advantage were 1.39 times more likely to undergo investigation than those with the least advantage (p<0.001). Presenting multiple times increased the odds of testing. Overall, lung cancer was diagnosed in 0.15% of patients, with haemoptysis preceding the highest proportion of cases (23/727; 3.2%).

CONCLUSION

There is significant variation in diagnostic testing for lung cancer-related symptoms in Australian primary care. Future efforts should focus on optimising testing for high-risk symptoms, rationalising investigations for non-specific symptoms, implementing safety-netting for low-risk but not no-risk symptoms and addressing socioeconomic disparities in testing rates.