Abstract
INTRODUCTION: Lung cancer (LC) accounts for 26.4% of all cancer deaths in Hong Kong (HK). Lung cancer screening (LCS) with low-dose computed tomography (LDCT) can reduce LC mortality. The cost-effectiveness of LDCT screening in high-risk individuals on the basis of smoking history has previously been investigated. However, nearly half of patients with LC in HK never smoke, indicating a different LC epidemiology compared with Western countries, where most LC cases are associated with smoking. We conducted a cost-effectiveness analysis for LCS, utilizing local data and expanding the target population to include we not only high-risk individuals identified on the basis of smoking history but also those identified through other risk factors. METHODS: A decision tree combined with a state-transition Markov model was developed to simulate identification, diagnosis, and treatments for high-risk individuals, from a health care provider perspective. The selection criteria and screening effectiveness for high-risk individuals on the basis of smoking history were obtained from the Dutch-Belgian Lung Cancer Screening Study, targeting heavy smokers aged 50 to 74 years; whereas the Taiwan Lung Cancer Screening in Never-Smoker Trial was used to model high-risk individuals on the basis of factors other than smoking history. Local LC survival and cost data were used to populate the model. The willingness-to-pay threshold used in the study was US$24,302 to US$40,202 per quality-adjusted life-year (QALY). RESULTS: Screening led to additional early LC detected, and LC mortality reduction, compared with no screening. Over a lifetime horizon, the incremental cost-effectiveness ratio for high-risk individuals on the basis of smoking history was US$14,122 per QALY. The incremental cost-effectiveness ratio for high-risk individuals on the basis of factors other than smoking history was lower at US$9610 per QALY. CONCLUSION: LCS with LDCT can be considered cost-effective in HK for high-risk individuals on the basis of smoking history and factors other than smoking history, contributing to the health benefits of the population. Our findings support a population-based LCS for all high-risk individuals identified through criteria beyond smoking history.