Control of smoking in occupational epidemiologic studies: Methods and needs

✍ Scribed by Dr. Aaron Blair; Kyle Steenland; Carl Shy; Maureen O'Berg; William Halperin; Terry Thomas

• Publisher: John Wiley and Sons
• Year: 1988
• Tongue: English
• Weight: 152 KB
• Volume: 13
• Category: Article
• ISSN: 0271-3586
• DOI: 10.1002/ajim.4700130102

✦ Synopsis

Many occupational epidemiologic studies lack information on smoking habits of study subjects [Blair and Spirtas, 1981; Steenland et al, 19841. In fact, it is unusual for cohort mortality studies to include information on tobacco use. The absence of information on smoking complicates interpretations since tobacco plays a major role in the etiology of many chronic diseases including cancers of the lung, esophagus, larynx, mouth, bladder, and other sites, as well as nonmalignant respiratory disease and cardiovascular disease as reported by the Surgeon General [ 19791. Differences in smoking prevalence between particular occupations and the comparison population may result in inaccurate assessment of the relationship between disease and the occupational exposure of interest.

The historical nature of most occupational cohort studies makes it difficult to obtain information on smoking. Occupational cohorts are usually assembled from company or union records which consist primarily of work histories and rarely contain information on tobacco use. Today this is changing somewhat and it is not unusual to find some information on smoking available in medical records of employers. Recording of smoking habits is a fairly recent practice, however, and information is seldom available for before the mid-1970s. Procedures to obtain information on smoking habits of cohort members can be expensive and complicated since many subjects will have terminated employment, moved to new geographic locations, or died. Location of these subjects is a difficult process.

In the absence of information on smoking, any observed excesses of smokingrelated diseases noted among particular occupational groups could be interpreted as due to heavy use of tobacco. Likewise, deficits of these diseases could be due to a