Research Article: Performance and Cost-Effectiveness of Computed Tomography Lung Cancer Screening Scenarios in a Population-Based Setting: A Microsimulation Modeling Analysis in Ontario, Canada

Date Published: February 7, 2017

Publisher: Public Library of Science

Author(s): Kevin ten Haaf, Martin C. Tammemägi, Susan J. Bondy, Carlijn M. van der Aalst, Sumei Gu, S. Elizabeth McGregor, Garth Nicholas, Harry J. de Koning, Lawrence F. Paszat, Steven D. Shapiro

Abstract: BackgroundThe National Lung Screening Trial (NLST) results indicate that computed tomography (CT) lung cancer screening for current and former smokers with three annual screens can be cost-effective in a trial setting. However, the cost-effectiveness in a population-based setting with >3 screening rounds is uncertain. Therefore, the objective of this study was to estimate the cost-effectiveness of lung cancer screening in a population-based setting in Ontario, Canada, and evaluate the effects of screening eligibility criteria.Methods and FindingsThis study used microsimulation modeling informed by various data sources, including the Ontario Health Insurance Plan (OHIP), Ontario Cancer Registry, smoking behavior surveys, and the NLST. Persons, born between 1940 and 1969, were examined from a third-party health care payer perspective across a lifetime horizon. Starting in 2015, 576 CT screening scenarios were examined, varying by age to start and end screening, smoking eligibility criteria, and screening interval. Among the examined outcome measures were lung cancer deaths averted, life-years gained, percentage ever screened, costs (in 2015 Canadian dollars), and overdiagnosis. The results of the base-case analysis indicated that annual screening was more cost-effective than biennial screening. Scenarios with eligibility criteria that required as few as 20 pack-years were dominated by scenarios that required higher numbers of accumulated pack-years. In general, scenarios that applied stringent smoking eligibility criteria (i.e., requiring higher levels of accumulated smoking exposure) were more cost-effective than scenarios with less stringent smoking eligibility criteria, with modest differences in life-years gained. Annual screening between ages 55–75 for persons who smoked ≥40 pack-years and who currently smoke or quit ≤10 y ago yielded an incremental cost-effectiveness ratio of $41,136 Canadian dollars ($33,825 in May 1, 2015, United States dollars) per life-year gained (compared to annual screening between ages 60–75 for persons who smoked ≥40 pack-years and who currently smoke or quit ≤10 y ago), which was considered optimal at a cost-effectiveness threshold of $50,000 Canadian dollars ($41,114 May 1, 2015, US dollars). If 50% lower or higher attributable costs were assumed, the incremental cost-effectiveness ratio of this scenario was estimated to be $38,240 ($31,444 May 1, 2015, US dollars) or $48,525 ($39,901 May 1, 2015, US dollars), respectively. If 50% lower or higher costs for CT examinations were assumed, the incremental cost-effectiveness ratio of this scenario was estimated to be $28,630 ($23,542 May 1, 2015, US dollars) or $73,507 ($60,443 May 1, 2015, US dollars), respectively.This scenario would screen 9.56% (499,261 individuals) of the total population (ever- and never-smokers) at least once, which would require 4,788,523 CT examinations, and reduce lung cancer mortality in the total population by 9.05% (preventing 13,108 lung cancer deaths), while 12.53% of screen-detected cancers would be overdiagnosed (4,282 overdiagnosed cases). Sensitivity analyses indicated that the overall results were most sensitive to variations in CT examination costs. Quality of life was not incorporated in the analyses, and assumptions for follow-up procedures were based on data from the NLST, which may not be generalizable to a population-based setting.ConclusionsLung cancer screening with stringent smoking eligibility criteria can be cost-effective in a population-based setting.

Partial Text: The National Lung Screening Trial (NLST) showed that screening with low-dose computed tomography (CT) can reduce lung cancer mortality [1]. Although the sensitivity of CT screening in the NLST was reported to be over 90% across the three screening rounds, the reported specificity ranged from 73.4% in the first round to 83.9% in the third round [1]. Overall, 23.3% of the CT screens in the NLST were false positive, which often required additional follow-up CT examinations and, infrequently, invasive procedures (such as a biopsy, bronchoscopy, or thoracotomy) to determine the malignancy of one or more suspicious pulmonary nodules detected by CT screening [1]. Lung cancer screening with three annual screens, as performed in the NLST, was reported to be cost-effective by US standards, yielding estimated cost-effectiveness ratios of US$52,000 per life-year gained and US$81,000 per quality-adjusted life-year gained [2,3]. However, although the cost-effectiveness of lung cancer screening in a population-based setting has been examined previously, it has not been examined extensively [4–9].

This simulation study indicates that lung cancer screening can be cost-effective in a population-based setting when eligibility is restricted to high-risk groups. In contrast, utilizing loose eligibility criteria yields nonoptimal and potentially cost-ineffective scenarios, as the cost-effectiveness of lung cancer screening is highly dependent on scenario characteristics, primarily the smoking eligibility criteria. Scenarios that utilize stringent smoking eligibility criteria are more cost-effective than scenarios that utilize less restrictive smoking eligibility criteria due to a focus on individuals at higher risk of developing lung cancer. This greatly reduces the number of screening examinations while still screening those at highest risk. Thus, the level of lung cancer risk at which an individual is eligible for lung cancer screening should be considered before implementing lung cancer screening policies. Future research should investigate the cost-effectiveness of lung cancer screening selection based on accurate lung cancer risk prediction models using suitable risk thresholds [41–43].



Leave a Reply

Your email address will not be published.