作者: R. C. Tiwari , K. Ghosh , A. Jemal , M. Hachey , E. Ward
关键词:
摘要: Every January for more than 40 years, the American Cancer Society (ACS) has estimated total number of cancer deaths that are expected to occur in United States and individual states upcoming year. In a collaborative effort improve accuracy predictions, investigators from National Institute ACS have developed tested new prediction method. The method was used create mortality predictions first time Statistics, 2004 Facts & Figures 2004. authors present conceptual overview previous state-space (SSM), they review results rigorous testing determine which provides accurate observed years 1997 1999. methods compared using squared deviations (the square predicted minus values) each sites published as well all combined. At national level, were not consistently lower every site either method, but average (averaged across sites, sex) substantially SSM During period 1999, estimates usually greater numbers combined several major probably because less sensitive recent changes rates (and associated counts) occurred early mid 1990s. improved particularly evident prostate cancer, changed dramatically late 1980s state two comparable. Based on these results, elected use annual at levels.