Cancer incidence in the United Kingdom: projections to the year 2030
Quantifying the future burden of cancer incidence and mortality, in terms of expected numbers of cases and deaths, or rates of disease, is important in optimising the allocation of resources for screening, diagnostic, therapeutic and palliative services and to provide a baseline from which the success of future interventions can be judged (Bray and Moller, 2006). The future number of cases (or deaths) is a consequence of two components of change; the evolution of the size and age structure of the population (demographic component) and changes in the rates of disease. The latter component is generally evaluated by some form of projection – likely future trends are estimated on the basis of those observed in the past. A distinction should be made between what we refer to as ‘projections' and ‘predictions'. We use the term ‘projections' for the extrapolation of observed trends in incidence rates that do not explicitly take into account expected changes in factors influencing risk (in a positive or negative way). By contrast, we use ‘predictions' to be statements about the future that take into account all known trends and likely future changes in risk factors. Predictions necessarily involve some sort of modelling: incorporating data on postulated future change in factors that affect risk, the components of the population likely to be involved, the size of the effect (relative risk) and latency between exposure and outcome. Several statistical methods have been used to make projections of future cancer burden. While short-term projections may assume continuation of past rates of change – generally within specific age groups – projections over the longer term need to model more complex interactions, incorporating the past components of change due to the effects of age, calendar period and birth cohort, and making explicit assumptions about the extent to which past changes are likely to operate into the future (Moller et al, 2003; Sedjo et al, 2007; Olsen et al, 2008). In the United Kingdom, previous studies have used such models to generate cancer mortality projections up to the year 2025 (Olsen et al, 2008), and cancer incidence projections up to the year 2020 for England (Moller et al, 2007). In this paper, we use data on the incidence of cancer in Great Britain (GB: England, Scotland and Wales) between 1975 and 2007 to make projections of the future rates of 28 types of cancer in the United Kingdom up to 2030, and estimate the corresponding future burden in terms of numbers of cases, using national population projections. Compared with the paper by Moller et al (2007), we use data for the United Kingdom (England, Scotland, Wales and Northern Ireland) rather than England, include data for 2004–2007 and make projections for prostate and bladder cancers (which they avoided) and use methodology that takes into account changes in breast screening when modelling breast cancer rates.