|Home | About | Journals | Submit | Contact Us | Français|
We thank Cherrie and colleagues for their comments on our paper. They have presented a simulation of a hypothetical group of women and applied a mesothelioma risk, based on time since exposure with clearance of fibres over time, in addition to the death rate from all causes. This simulation gave about seven expected mesotheliomas in the period 1996–2003. As they acknowledge, we did note that the absence of mesotheliomas in this time period indicated a larger reduction in mesothelioma incidence than would be predicted from an elimination model. In their model they took the clearance half‐life as 20 years; we consider a faster rate of clearance would be equally justified, and using a half‐life of 10 years there would have been 4.0 mesotheliomas expected in 1996–2003 in the cohort with known duration of employment. The probability of observing zero would then have been 0.018. As 45 of the 65 mesothelioma occurred in the cohort with known duration of employment, this observation and the simulation by Cherrie and colleagues are in close agreement.
The second point raised by Cherrie and colleagues was the importance of taking into account competing causes of death. As they state these are implicitly adjusted for because the occurrence of death from another cause removes that person from contributing further years at risk after their death. This is not a perfect adjustment because our information on exposure was limited to duration—we had no information on intensity. As they suggest the more intensely exposed members of the cohort might have died of lung cancer and thus have been differentially removed from the group at risk of developing mesothelioma. This is an interesting suggestion but, in the absence of data on intensity of exposure, it is impossible to address the size of the effect that this might have had. However, as there were only about 20 attributable deaths from lung cancer among over 500 from all other causes, any effect would surely have been trivial.