Modelling the Relative and Absolute Risks of Radiation-Induced Cancers (Q3759778)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 4009617
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | Modelling the Relative and Absolute Risks of Radiation-Induced Cancers |
scientific article; zbMATH DE number 4009617 |
Statements
Modelling the Relative and Absolute Risks of Radiation-Induced Cancers (English)
0 references
1987
0 references
radiation-induced cancers
0 references
relative risk
0 references
absolute risk
0 references
parametric family of models
0 references
cohort studies of patients
0 references
ankylosing spondylitis
0 references
atomic bomb survivors
0 references
diagnostic procedures
0 references
additive hazards model
0 references
cancer incidence
0 references
cancer mortality
0 references
model discrimination
0 references
proportional hazards model
0 references
chi-square goodness of fit
0 references
long term carcinogenic risk
0 references
epithelial tumours
0 references
breast cancer
0 references
leukaemia
0 references
0.755937397480011
0 references
0.7554136514663696
0 references
0.7528696060180664
0 references
0.7466109395027161
0 references
Data from studies of the long term carcinogenic risk following exposure to radiation are usually available in the form of numbers of deaths and person-years at risk. The individuals are divided into groups defined by the dose received, and within each dose group stratified by other variables which affect the hazard rate. Some of these variables are intrinsic, such as sex and age at exposure; others are time-varying, such as time since exposure. Let \(Y_ i(j)\) denote the numbers of deaths from the cause of interest in the i th stratum for the j th dose group. Similarly, let \(P_ i(j)\) and \(\lambda_ i(j)\) denote the corresponding number of person-years at risk and hazard rate, respectively. The distribution of \(Y_ i(j)\) conditional on \(P_ i(j)\) can then be approximated as Poisson with mean: \(\mu_ i(j)=P_ i(j)\lambda_ i(j)\) independently for different i and j. The authors consider first the general risk model: the case of two dose groups, one high and one low or unexposed, is considered here, with the model: \[ \lambda_ i(1) = [\lambda_ i^{\gamma}(0)+\{1+\exp (x^ T_ i\beta)\}^{\gamma}-1]^{1/\gamma}. \] The parameters are estimated and the model is checked; \(\chi^ 2\) goodness of fit is used. Examples are given, such as non-sex-specific epithelial tumours, breast cancer, leukaemia, and the results of the paper are discussed.
0 references