The magnitude of the cancer problem, as conveyed by the incidence, mortality and prevalence figures, was studied by analysis of the material of the Finnish Cancer Registry from 1953-1970. The prevalence was defined as the number of persons who had cancer, or had ever had cancer, and were living on D
Summarizing indices for comparison of cancer incidence data
β Scribed by Richard Doll; Paula Cook
- Publisher
- John Wiley and Sons
- Year
- 1967
- Tongue
- French
- Weight
- 610 KB
- Volume
- 2
- Category
- Article
- ISSN
- 0020-7136
No coin nor oath required. For personal study only.
β¦ Synopsis
The desirability of achieving agreement on a method of summarizing cancer incidence is emphasized, and the advantages and disadvantages of the use of the traditional standardized incidence rates are discussed.
It is concluded that: I ) No single index is capable of replacing the individual sex-and age-specific incidence rates, and these should always be presented when basic cancer incidence data are published.
2 ) For aetiological studies age-specific incidence rates can usefully be summarized in two indices, one ( a truncated standardized incidence) showing a standardized incidence over a restricted age range, and the other indicating the rate at which cancer incidence increases with age.
3 ) An appropriate truncated rate for most epithelial cancers is one for the age range of 35 to 64 years, using weights of 6 , 6 , 6 , 5 , 4 and 4 for the individual 5-year age groups, derived from Segi's (1960) world standard population. For other cancers, age ranges of 0 to 14 years and of 0 to 44 years would be more suitable, with weights of 12, I0 and 9,
π SIMILAR VOLUMES
This paper extends four goodness-of-ΓΏt measures of a generalized linear model (GLM) to random e ects and marginal models for longitudinal data. The four measures are the proportional reduction in entropy measure, the proportional reduction in deviance measure, the concordance correlation coe cient a