URANIUM MINERS DATA
(used in examples 1.6, 4.2, 4.5, and 4.16)
Background:
The Colorado Plateau uranium miners cohort consists of 3347 miners who worked underground at least one month in the uranium mines of the four-state Colorado Plateau area and were examined at least once by Public Health Service physicians between 1950 and 1960. The miners were traced for mortality outcomes through December 31, 1982, by which time 258 lung cancer deaths were observed. Job histories were obtained from each of the mining companies and combined with available data on annual mine radon levels to construct each miner's radon exposure history. A smoking history was taken at the first examination and updated with each subsequent exam.
For computational convenience, the cohort results reported in examples 4.2 and 4.4 are based on a nested case-control sample with 50 controls per case (cf. the comment in example 4.16). In example 4.16 we report estimates using 50 controls per case together with estimates based on simple and counter-matched nested case-control samples with 1 and 3 controls per case. (Note that the data sets given below do not cover example 4.14.)
Data sets
3 controls (to be given)
1 control (to be given)
3 controls counter-matched (to be given)
1 control counter-matched (to be given)
Coding for the data sets with simple random sampling of the controls
setno: risk set number
ccind: case-control indicator (0 = control, 1 = case)
time: age of the case
samp: number in sampled risk set (i.e. number of controls + 1)
atrisk: number at risk
ttlsmk: cumulative smoking (in 1000 packs)
ttlrdn: cumulative radon (in 100 WLM)
Coding for the data sets with counter-matched sampling of the controls (to be given)
Acknowledgement:
We thank Bryan Langholz for making the data available.