NOTE: This page is archived for historical purposes and is no longer being maintained or updated.
D3.1 Dealing With Variability of Hazards in Occupational Injury Epidemiology-Kromhout H, Loomis D
Injury epidemiology has been very traditional with respect to exposure assessment. Evaluation of risks has been based on job titles or personal attributes, like sex and age. Nevertheless, we know that hazardous circumstances are not permanently present or present with the same intensity. In order to make better inferences on causative factors leading to injuries, a better understanding of the variability in these factors will become a necessity. Much can be learned from recent advances in related fields of occupational cancer and respiratory disease epidemiology. Understanding of variability patterns in chemical and physical exposures has improved strategies for assessing and assigning exposure. For example, recent research on exposure to magnetic fields showed that health effects would have gone undetected had more traditional approaches to exposure assessment been applied. Variation in exposure has two fundamental dimensions: person and time. Variability between groups of people is a fundamental requirement for most epidemiological research. Exposure may also vary within groups and within individuals, however. The dimensions of exposure variability can be described quantitatively by the expression Xij(t) = f(m + a i + bj + dt),where X(t) is instantaneous individual exposure at time t, : is the overall mean exposure level, and a , b, and d respectively represent deviations from : associated with being a member of group i, being the j-th person in that group, and temporal fluctuation of exposure at time t. Opportunities for improved hazard assessment in the field of occupational injury epidemiology will be sketched based on these concepts and experience with other workplace agents.
D3.2 Variability, Measurement, and Analysis of Hours of Exposure in a Cohort of Fishers-Marshall SW
Occupational epidemiologists and industrial hygienists have developed methods for modeling exposures to environmental agents, such as dusts and chemicals, but these methods have not been widely applied to injury hazards. We illustrate the general approach using preliminary data from a cohort of fishers in Eastern North Carolina. Hours spent working on the water was assessed on a weekly basis throughout the fishing season. We partition the variation in weekly hours worked into two components, between-worker variance (FB2) and within-worker week-to-week (Fw2) variance, using a multilevel random effects model. Risk ratios (RR) and confidence intervals (CI) for weekly hours worked were estimated from the model.
The ratio of between-worker variation (sB2=1.40) to within-worker variation (sw2=0.40) was 3.5, indicating that the majority of the variation in weekly hours worked was due to differences between fishers. The range of variation was quantified by calculating, from the model, the ratio of the 97.5th percentile to 2.5th percentile of the distribution of hours worked. This ratio was 102.8 hours for the between-fishers component and 11.9 hours for the week-to-week within-fishers component.
Crabbing was associated with the least time spent on the water (RR=1, referent category), while clamming (RR= 2.91; 95%CI: 1.46-5.83) and shrimping (RR=2.05; 95%CI: 0.64-6.54) required the greatest time. Finfishing (RR=1.46; 95%CI: 0.97-2.19) and oystering (RR=1.46; 95%CI: 0.70-3.04) were intermediate. Mixed operations required more time on the water than operations involving only one type of fishing (RR=1.81; 95%CI: 1.24-2.63). Future analyses will examine the role of seasonal and meteorological factors.
Whenever they are on the water, fishers are exposed to many important but non-modifiable hazards, e.g. unpredictable changes in local weather conditions. Increasing our understanding of the reasons for variation in work patterns presents an important (and possibly the only) opportunity for reducing exposure to these environmental hazards.
D3.3 Measuring Transient Occupational Injury Exposures-Sorock GS, Lombardi DA, Hauser R, Eisen EA, Herrick R, Mittleman MA
Case-only designs, such as case-crossover studies, use cases as their own controls and compare exposures immediately before the injury (hazard period) to earlier control periods in the same individual. This study design eliminates confounding by differences between individuals, allowing investigators to focus on transient exposures occurring in the workplace. Some examples of transient risk factors that may be studied include personal protective equipment, staffing level, and antihistamine use. These transient risk factors are the focus of the investigation rather than fixed risk factors like age, gender, industry, and occupation. Such transient factors may be more modifiable than fixed ones that can be evaluated as potential effect modifiers.
We used the case-crossover study design and interviewed 1,128 subjects (874 men and 254 women) a median of 1.3 days after an acute traumatic hand injury. The majority of subjects (69%) reported at least one of eight transient exposures during the 90 minutes before the injury. Some problematic areas were noted: 1) knowing the average duration of each transient exposure under investigation before selecting the control periods to avoid correlation of transient exposures over both the hazard and control intervals; 2) selecting control periods when the injury occurred at the beginning of the work shift or in the first month of employment; and 3) relying on self-reported exposures and their timing and the lack of any gold standard for comparison. These problem areas will be illustrated and plans for questionnaire reliability and validity assessments will be discussed.
|NOIRS 2000 Menu Page|
|NIOSH home page||CDC home page||NORA Traumatic Injury home page|