Modelling and forecasting lung cancer incidence and mortality in Saudi Arabia

• Main Author: Alzahrani, S. M.
• Published: University of Salford 2016
• Subjects: 616.99
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.707014

The aim of this research is to forecast the rates of lung cancer incidence and mortality in the Kingdom of Saudi Arabia using data on lung cancer diagnosis between 1994 and 2009. Lung cancer data, including incidence and mortality, were obtained from Saudi Cancer Registry at the Ministry of Health. The Central Department of Statistics & Information at the Ministry of Planning also provided data on person characteristics, such as age, gender and ethnicity. These data serve as a basis for modelling the effect of gender, ethnicity, and age at diagnosis, and region on incidence and mortality. For comparison of incidence and mortality rates between region and over time, standardised rates are used in this thesis, based on a hypothetical standard population, in our case the world standard population. We use several modelling approaches. The first part of the analysis uses two approaches. The first approach concentrates on Box–Jenkins methodology, and the second approach uses dynamic regression modelling including both finite and infinite lag models to forecast lung cancer incident cases. The second part focuses on age-period-cohort modelling including both incidence and mortality rates of lung cancer, and using two methodological approaches, namely spline functions and Bayesian dynamic models, for the incidence and mortality respectively. Lung cancer is rarely diagnosed in people under 30 years of age in Saudi Arabia, but incidence rises sharply thereafter peaking in the 65-69 years age group. Males have a 79% greater incidence rate of lung cancer than females across the entire dataset when adjusting for the other effects. The average age standardised incidence rate in 2009 was 3.8 per 100,000 population whereas the average age standardised mortality rate was 1.9 per 100,000 population in the same year. The highest number of cases of lung cancer were reported in the Western region at 187 and in Riyadh at 144 cases and the majority of cases were diagnosed in winter (December - March). The forecast incidence rate of lung cancer is expected to decrease in men but to increase in women over the next ten years. This is perhaps due to the increase in the proportion of female smokers. The male age standardised rate of lung cancer incidence is forecast to fall from 4.6 in 2010 to 2.4 per 100,000 by 2020, whereas the female age standardised rate is forecast to increase from 2.0 in 2010 to 2.2 per 100,000 by 2020. On the other hand, the overall mortality rate of lung cancer (with 95% credible interval shown) is forecast to increase to 2020 from 1.8 (1.61, 1.94) in 2010 to 3.04 (0.13, 5.94) per 100,000 population. Age has a strong association with lung cancer mortality, suggesting age-related causes such as accumulative exposures to smoking over time may be the main reason for increasing lung cancer mortality in Saudi Arabia. This is the first study to forecast lung cancer incidence and mortality in Saudi Arabia. It will help the Saudi Arabian Ministry of Health to understand the rate of future lung cancer incidence and mortality and the overall effects of the population classes, and to plan healthcare provision accordingly. The data are limited because the Saudi Cancer Registry has only been in existence since 1992. Therefore, we can expect the precision of forecasts to improve as further data are collected.