BREAST-CANCER 5-YEAR SURVIVAL IN NEW-SOUTH-WALES WOMEN, 1972 TO 1991

This study of breast cancer survival is based on analysis of five-year relative survival of 38 362 cases of invasive breast cancer in New So uth Wales (NSW) women, incident between 1972 and 1991, with follow-up to 1992, using data from the population-based NSW Central Cancer Regis try. Survival was ascertained by matching the registry file of breast cancers against NSW death certificates from 1972 to 1992, mainly by au tomated probabilistic linkage. Absolute survival of cases was compared with expected survival of age- and period-matched NSW women. Proporti onal hazard regression analysis was used for examination of the effect s on excess mortality of age, period of diagnosis and degree of spread at diagnosis. Relative survival at five years increased from 70 per c ent in 1972-1976 to 77 per cent in 1987-1991. Survival improved during the 1970s and in the late 1980s. Regression analysis suggested that p art of the improved survival in the late 1980s was due to lesser degre e of spread at diagnosis, whereas the improved survival during the 197 0s may have been due to treatment. Survival was better for those aged 40-49 years (RR = 0.86) and worse for those aged greater than or equal to 70 years (RR = 1.22) compared with the referent group (60-69 years ). Excess mortality was much less for those with invasive localised di sease than those with regional spread (RR = 3.1) or metastatic cancer (RR = 15.5) at diagnosis. For the most recent period (1987-1991), rela tive five-year survival was 90, 70 and 18 per cent, respectively, for the three degree-of-spread categories.