Dp. Loomis et al., ORGANIZATION AND CLASSIFICATION OF WORK HISTORY DATA IN INDUSTRY-WIDESTUDIES - AN APPLICATION TO THE ELECTRIC-POWER INDUSTRY, American journal of industrial medicine, 26(3), 1994, pp. 413-425
Industry-based cohort studies require systems for organizing work hist
ory data. Although the ultimate goal may be to assess the hazards of s
pecific exposures, classification of the job titles that comprise work
histories serves an important descriptive purpose in itself and is of
ten necessary before exposure data can be obtained. A system we have c
reated for organizing jobs in a study of 135,000 workers at five elect
ric power companies highlights conceptual and practical issues in mana
ging work history data for epidemiological studies. Job characteristic
s including function, location, and authority were used to develop a s
ystem of 28 occupational categories. Comprehensibility, flexibility, a
nd efficiency were important criteria in designing the system. Assessm
ent of exposures was an implicit goal; the same categories will define
job-exposure matrices for numerous agents. A combination of computer
algorithms and expert judgment was used to assign individual job title
s to the categories. This system facilitates examining the effects of
various agents and controlling for confounding. The 28 categories can
be collapsed and regrouped to analyze disease risks in relation to exp
osures to magnetic fields and other agents; even exposures not previou
sly considered could be brought into the study with this generic syste
m for organizing the electric power industry. (C) 1994 Wiley-Liss, Inc
.