A principal factor analysis to characterize agricultural exposures among Nebraska veterans

Our plain-English summary of A principal factor analysis to characterize agricultural exposures among Nebraska veterans, by L. Weissenburger-Moser, J. Meza, F. Yu, et al (Journal of Exposure Science & Environmental Epidemiology, 2017).

The study

• Authors: L. Weissenburger-Moser, J. Meza, F. Yu, O. Shiyanbola, D.J. Romberger, T.D. LeVan
• Journal: Journal of Exposure Science & Environmental Epidemiology
• Published: 2017;27(2):214-220
• Cited by: 10 articles, as counted by NIH iCite
• Read it: PubMed · Publisher · Free full text

Abstract

Agricultural workers are at an increased risk of developing chronic respiratory disorders. Accurate estimation of long-term agricultural exposures based on questionnaires has been used to improve the validity of epidemiologic investigations and subsequent evaluation of the association between agricultural exposures and chronic diseases. Our aim was to use principal factor analysis (PFA) to distill exposure data into essential variables characterizing long-term agricultural exposures. This is a cross-sectional study of veterans between the ages of 40 and 80 years and who worked on a farm for ≥2 years. Participant characteristics were: 98.1% were white males with a mean age 65±8 (SD) years and 39.8% had chronic obstructive pulmonary disease. The final model included four factors and explained 16.6% of the variance in the exposure data. Factor 1 was a heterogeneous factor; however, Factor 2 was exclusively composed of exposure to livestock such as hogs, dairy and poultry. Factor 3 included exposures from jobs on or off the farm such as wood dust, mineral dust, asbestos and spray paint. Crop exposure loaded exclusively in Factor 4 and included lifetime hours of exposure and maximum number of acres farmed in the participants' lifetime. The factors in the final model were interpretable and consistent with farming practices.

In plain English

This cross-sectional study used questionnaire answers from 681 veterans aged 40 to 80 who had worked on a farm as adults for at least 2 years. They were recruited from March 2008 to December 2013 at the VA Nebraska Western Iowa Health Care System. The authors used a statistical method, principal factor analysis, to sort self-reported exposures into a few patterns. These covered farm work, farm tasks, livestock, crops and "other exposures" including wood dust, mineral dust, asbestos, solvents, spray paint and welding fumes. The study did not test links to disease, and the asbestos question did not name a fiber type. In the first three of four models, asbestos grouped with other dust, solvent and paint exposures. In the final model, which explained 16.6% of the variation, the text describes that group as rock dust and spray paint, although the abstract lists asbestos in it. The authors concluded the method was effective. They stated that recall bias was probable and that interviewer bias was possible because of telephone versus in-person interviews. They also noted a mostly white male population and that exposures were not directly measured.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Cross-sectional study; exploratory statistical analysis (principal factor analysis) of exposure questionnaire data
• Who was studied: 681 veterans aged 40 to 80 who had worked on a farm as adults for at least 2 years, recruited in primary care clinics of the VA Nebraska Western Iowa Health Care System from March 2008 to December 2013. 263 answered Questionnaire 1 (by telephone) and 418 answered Questionnaire 2 (in person). People diagnosed with asthma, lung cancer or interstitial lung disease were excluded. The prevalence of COPD was 39.8%.
• Exposure: Self-reported agricultural exposures: farm work duration and intensity, farm tasks, livestock, crops, diesel and gas equipment, and "other exposures" on and off the farm. The other exposures were wood dust, grain dust, silica/mineral dust, asbestos (no fiber type given), smoke other than cigarette, chemical solvents, spray paint and welding fumes. They were recorded as yes/no, total years or total lifetime hours.
• Main results:
  • Model 1 (Questionnaire 1, 263 participants): four factors explained 24.4% of the variance. Factor 1 was made up only of the "other exposures" from a job on or off the farm, including asbestos.
  • Model 2 (Questionnaire 2, 418 participants, yes/no exposures in the 20s and 40s): four factors explained 14.5% of the total variance. Factor 4, explaining 2.7%, included asbestos with wood dust, rock dust, chemical solvents and spray paint.
  • Model 3 (Questionnaire 2, lifetime exposures in years): three factors explained 10.5% of the total variance. Factor 3, explaining 2.3%, included lifetime years of exposure to asbestos, wood dust, rock dust, chemical solvents and spray paint.
  • Model 4 (Questionnaire 2, lifetime hours and maximum livestock or acres): four factors explained 16.6% of the variance. In the results text, Factor 3 (2.7% of variance) included lifetime hours of rock dust and spray paint.
  • In a sensitivity analysis of Model 4 split by COPD status, Factors 1 and 2 showed similar clustering patterns in the total population, in those with COPD and in those without COPD.
• The authors' conclusion: The authors concluded that principal factor analysis was an effective method for characterizing exposure patterns in this population, and that the resulting factor patterns were interpretable and logical in terms of farming practice. They stated that Model 4, with the most detailed exposure information, captured the highest percentage of variance. They identified years worked on a farm, farm tasks, livestock, crops and "other exposures" as the key questionnaire items. They described examining links with respiratory diseases such as COPD as the next step.
• Limitations the authors state:
  • Recall bias is probable because participants recalled lifetime exposures retrospectively.
  • Interviewer bias is possible because Questionnaire 1 was given by telephone and Questionnaire 2 in person.
  • Generalizability is limited: participants were primarily older white male veterans, whose exposures may differ from those of younger workers.
  • Agricultural exposures were not directly measured.

Related studies

• Association of Agricultural, Occupational, and Military Inhalants With Autoantibodies and Disease Features in US Veterans With Rheumatoid Arthritis — A.V. Ebel, G. Lutt, et al. (2021)
• Asbestos plaques in a typical Veteran's hospital population — J.A. Miller, J.V. Zurlo (1996)
• Living with mesothelioma: A qualitative study of the experiences of male military veterans in the UK — S. Ejegi-Memeh, L. Darlison, et al. (2021)
• Chronic bronchitis among nonsmoking farm women in the agricultural health study — M. Valcin, P.K. Henneberger, et al. (2007)
• Asbestos exposure and mesothelioma mortality among atomic veterans — J.E. Till, H.L. Beck, et al. (2022)
• Increased risk of lung cancer in pesticide-exposed male agricultural workers — E. Barthel (1981)

Browse more research in Journal of Exposure Science & Environmental Epidemiology, or return to medical research.