My primary research interests are in the development and application of methods for population-based cancer survival analysis, particularly the estimation and modeling of relative survival. I am also interested in cancer survivorship, especially the physical and psychosocial impact of cancer and its treatment. I have a general interest in statistical aspects of the design, analysis, and reporting of epidemiological studies as well as a general interest in studies of disease aetiology, with particular focus on cancer epidemiology along with perinatal and reproductive epidemiology.
Methods for population-based cancer survival analysis
My primary research interests are in the development and application of methods for population-based cancer survival analysis, particularly the estimation and modeling of relative survival. The general aim of my research program is to develop and apply statistical methods for estimating and modelling cancer patient survival using registry data, with particular emphasis on methods for presenting survival statistics in a manner relevant for clinicians and patients.
Patient survival is the most important single measure of cancer patient care (the diagnosis and treatment of cancer) and together with incidence and mortality is one of the key measures of cancer control. The optimal method for monitoring and evaluating the effectiveness of cancer patient care is through the population-based study of cancer patient survival, which is only possible using data collected by population-based cancer registries. It is standard in population-based studies to use relative survival as the measure of cancer patient survival (Dickman and Adami, 2006). Relative survival is the ratio of observed (all-cause) to expected survival proportion and provides a measure of excess mortality associated with diagnosis of cancer. Excess mortality is the difference between the observed (all-cause) mortality and the mortality that would have been expected if the patients were not diagnosed with cancer. It has the advantage that cause of death information is not required and that it captures mortality both directly due to the cancer as well as indirectly due to the cancer (e.g., increased risk of non-cancer mortality caused by the treatment). In cancer clinical trials it is standard to estimate cause-specific survival but this is less frequently used in population-based studies since information on cause of death is not as accurate as it is in clinical trials.
In recent years I have collaborated closely with Prof. Paul Lambert at the University of Leicester. We have developed and applied cure models for relative survival and methods for estimating the probability of death due to cancer in the presence of competing risks. We have investigated the assumptions underlying relative survival, and evaluated approaches to predicting cancer patient survival. We have also developed methods for partitioning excess mortality, that we have applied to studying treatment-related mortality among patients with Hodgkin lymphoma. We have developed freely-available user-friendly software (primarily Stata but also SAS) to implement our methods and hold courses each year (http://cansurv.net/) to train cancer researchers working in the area. We have applied our methods to a wide range of cancer sites; in recent years have had a particularly successful collaboration with a group of haematologists at Karolinska Hospital Solna lead by Prof. Magnus Björkholm.