Causal inference in cancer epidemiology
The Norwegian health surveys and health registries represent a unique treasure. By using techniques from modern causal inference to study the complex associations between exposure over time and cancer risk, we will learn more about the exposure-cancer associations and help optimize public health strategies towards cancer prevention.
About the project
The Norwegian health surveys and health registries represent a unique treasure, which needs to be more extensively exploited with efficient modern analytical tools, moving us closer to more targeted prevention for each individual patient. Recently developed methodology in causal inference provides such tools.
Cutaneous malignant melanoma is the cancer with the greatest increase (19%) from the previous five-year period (2003-2007) to the last period (2008-2012) in Norwegian women. The Cancer Registry of Norway recently expressed concern about the high incidence of melanoma in Norway: After a period of decreasing or non-increasing melanoma incidence rates in the 1990s, a steady increase is observed over the last 10 years. Primary prevention is the most effective way to fight melanoma, since this cancer is mainly caused by ultraviolet (UV) exposure in sun-sensitive subjects. Age at UV exposure is important.
The Norwegian Women and Cancer cohort study, has since 1991 included about 170 000 women. UV exposure (sunburns, bathing vacations and solarium use) in all previous age decades, host factors (hair and eye color, the propensity to sunburn, the ability to tan, freckles and nevi) and several other variables were reported at inclusion. Because a long follow-up is needed, repeated assessments of UV exposure are carried out.
To understand the causal pathway from UV exposure to melanoma diagnosis, we will apply recent methodological developments for mediation analysis based on counterfactuals in a survival setting.
- Professor Adele Green, PhD, QIMR Berghofer Medical Research Institute, Queensland, Australia and CRUK Manchester Cancer Institute, UK.
- Associate Professor Theis Lange, PhD, University of Copenhagen, Denmark
- Professor Eiliv Lund, MD, PhD, University of Tromsø, Norway.
- Professor Elisabete Weiderpass, MD, PhD, Karolinska Institutet, Sweden.