Table 2 Key assumptions of the model

Changes in smoking behaviour are not modelled.

Programme uptake will be similar in real life to in the UKLS trial.

There is full concordance with screening programme (i.e. no missed appointments)

Health-related quality of life similar for preclinical and diagnosed lung cancer (stratified by stage).

Health-related quality of life similar for clinically presenting and screen-detected lung cancer of the same stage.

Health-related quality of life for diagnosed lung cancer is constant until death.

Natural history of lung cancers is similar across all included individuals.

Lung cancers progress through stages in numerical order without skipping any stages.

Sensitivity of LDCT is independent of patient and tumour characteristics.

Lung cancer mortality: screening cannot be less effective than no screening.

Mortality from preclinical lung cancer assumed to be negligible.

Lung cancer incidence in participating population similar to incidence in general smoking (current and former) population.

Survival in participating population similar to survival in general population (stratified by stage).

Incidental findings not modelled.

True-positive results lead to immediate diagnosis and treatment.

False-positive and indeterminate results are treated equivalently.

Non-attendance of screening was not explicitly modelled

Additional cancers caused by radiation exposure not modelled.

Risk prediction is dependent only on prevalence of occult lung cancer or short-term incidence (within 3 years).