ABSTRACT A microsimulation model is an ideal tool to evaluate the impact of the new health screening programme. Health policymakers in New Zealand are reviewing international evidence on lung cancer screening with the aim of reducing lung cancer mortality through early detection. In this study, we developed and calibrated a New Zealand specific microsimulation model for the natural history of lung cancer. This is the first time a lung cancer microsimulation model has been developed for New Zealand. We adapted simulation steps devised for the MIcrosimulation Lung Cancer (MILC) model (Chrysanthopoulou, 2017). The simulation steps include a Gompertz tumour growth component, disease progression based on log-normal distributions and a survival component. Simulated lung cancer rates at different disease stages (i.e. local, regional, and distant) and at diagnosis were used to calibrate and validate model parameters to data from the New Zealand (NZ) lung cancer registry. We used the Bayesian approach to calibrate the model parameters using the Hamiltonian Monte Carlo (HMC) approach, which is not commonly employed in the field of cancer modelling. The calibrated model was able to predict general trends in lung cancer incidence by age group when compared to lung cancer incidence in NZ.
Read full abstract