Cancer mortality trends in the Umbria region of Italy 1978-2004: a joinpoint regression analysis

Abstract

Background: The aim of the present paper was to analyse cancer mortality in the Umbria region, from 1978 to 2004. Mortality trends depend on a number of factors including exposures, health system interventions, and possibly artefact (e.g. classification change, variations of data completeness). Descriptive data on mortality only allow for generation of hypotheses to explain observed trends. Some clues on the respective role of possible mortality determinants may be found comparing mortality with incidence and survival data.

Methods: Mortality data for the periods 1978-1993 and 1994-2004 were supplied by the National Institute of Statistics (ISTAT) and the Regional Causes of Death Registry (ReNCaM) respectively. Sex and site-specific mortality time trends were analysed by the "joinpoint regression" method.

Results: For all sites combined, in both sexes, the standardised rate was first increasing before the end of the eighties and decreasing thereafter. Gastric cancer mortality showed a different trend by gender; that is the rate constantly decreased over the period among females while, for males, it was first increasing up to 1985 and decreasing thereafter. Liver cancer trend showed a pattern similar to gastric cancer. Large bowel cancer showed a gender specific trend, that is it was increasing among males and stable among females. Also lung cancer mortality varied by gender: it started to decline after 1989 among males but was steadily increasing over the study period among women. A decreasing trend for female breast cancer mortality began in 1994. Prostate cancer mortality trend is the only one showing two significant joinpoints: mortality decreased up to 1990, then it increased up to 1998 and, finally, was decreasing.

Conclusion: Overall cancer mortality was decreasing in both sexes in Umbria and this favourable trend will probably continue and further improve since population screening against breast, cervix, and large bowel cancers were recently introduced. Besides gastric cancer, tobacco-related cancers and prostate cancer mainly contributed to mortality reduction in males, whereas breast cancer mainly contributed to declining mortality in females.

Figure 1

Observed standardized rates per 100.000 inhabitants (▲ males; ○ females) and 'best' joinpoint model estimates (solid line males; dashed line females) for selected cancer sites by gender.

Figure 2

Observed standardized mortality rates per 100.000 inhabitants and 'best' joinpoint model estimates for female breast, ovary, prostate, male head and neck (▲ observed; solid line joinpoint), oesophagus (△ observed; dotted line joinpoint) and larynx cancers (◆ observed; mixed line joinpoint).