Email updates

Keep up to date with the latest news and content from BMC Public Health and BioMed Central.

Open Access Research article

Methods and feasibility of collecting occupational data for a large population-based cohort study in the United States: the reasons for geographic and racial differences in stroke study

Leslie A MacDonald1*, LeaVonne Pulley2, Misty J Hein1 and Virginia J Howard3

Author Affiliations

1 National Institute for Occupational Safety and Health, 4676 Columbia Parkway, MS R15, Cincinnati, OH 45226, USA

2 College of Public Health, University of Arkansas for Medical Sciences, 4301 West Markham Street, Little Rock, AR 72205, USA

3 School of Public Health, University of Alabama at Birmingham, 1665 University Boulevard, Birmingham, AL 35294, USA

For all author emails, please log on.

BMC Public Health 2014, 14:142  doi:10.1186/1471-2458-14-142

Published: 10 February 2014

Abstract

Background

Coronary heart disease and stroke are major contributors to preventable mortality. Evidence links work conditions to these diseases; however, occupational data are perceived to be difficult to collect for large population-based cohorts. We report methodological details and the feasibility of conducting an occupational ancillary study for a large U.S. prospective cohort being followed longitudinally for cardiovascular disease and stroke.

Methods

Current and historical occupational information were collected from active participants of the REasons for Geographic And Racial Differences in Stroke (REGARDS) Study. A survey was designed to gather quality occupational data among this national cohort of black and white men and women aged 45 years and older (enrolled 2003–2007). Trained staff conducted Computer-Assisted Telephone Interviews (CATI). After a brief pilot period, interviewers received additional training in the collection of narrative industry and occupation data before administering the survey to remaining cohort members. Trained coders used a computer-assisted coding system to assign U.S. Census codes for industry and occupation. All data were double coded; discrepant codes were independently resolved.

Results

Over a 2-year period, 17,648 participants provided consent and completed the occupational survey (87% response rate). A total of 20,427 jobs were assigned Census codes. Inter-rater reliability was 80% for industry and 74% for occupation. Less than 0.5% of the industry and occupation data were uncodable, compared with 12% during the pilot period. Concordance between the current and longest-held jobs was moderately high. The median time to collect employment status plus narrative and descriptive job information by CATI was 1.6 to 2.3 minutes per job. Median time to assign Census codes was 1.3 minutes per rater.

Conclusions

The feasibility of conducting high-quality occupational data collection and coding for a large heterogeneous population-based sample was demonstrated. We found that training for interview staff was important in ensuring that narrative responses for industry and occupation were adequately specified for coding. Estimates of survey administration time and coding from digital records provide an objective basis for planning future studies. The social and environmental conditions of work are important understudied risk factors that can be feasibly integrated into large population-based health studies.

Keywords:
Occupations; Occupational exposure; Stressful events; Social class; Cohort studies; Epidemiologic methods; Data collection; Stroke; Cardiovascular disease