Investigating dose response relations in occupational mortality studies: something to keep in mind.

Investing dose response relations in occupational mortality studies: something to keep in mind SIR,-In their recent article (1987;44:642-4) Swaen and Volovics discuss the correct method for calculat- ing person-years when a dose response relation is examined.They explain that when exposure is measured by an amount of time in a particular environment, care is needed in attributing the person- years to correct levels.Two points may be worth making.Firstly, it should be noted that this issue was raised with some cogency in correspondence in the Lancet over ten years ago in relation to a study of vinyl chloride monomer and mortality.The original authors'2 did not use a correct method and showed a spurious inverse trend in mortality with increasing exposure.Their error was pointed out by others.?5 The debate forms a useful demonstration of the implications of the right and wrong approaches.Secondly, the need for appropriate allocation of the person-years may be seen clearly by analogy with age.If a worker is followed up for 40 years, from age 30 to age 69, it must be obvious that in calculating a standardised mortality ratio one should not allocate 40 years to the 65-69 years category.To do so, however, would involve the same error as allocating the entire person-years of experience of a worker with 10 or more years of exposure to the 10 + category.