NIOSHTIC-2 Publications Search
Statistical model for prediction of retrospective exposure to ethylene oxide in an occupational mortality study.
Hornung-RW; Greife-AL; Stayner-LT; Steenland-NK; Herrick-RF; Elliott-LJ; Ringenburg-VL; Morawetz-J
Am J Ind Med 1994 Jun; 25(6):825-836
Three steps were presented to be used in developing a model to predict exposure levels of ethylene-oxide (75218) (EtO) in the sterilization industry. The estimates were used to calculate the cumulative exposure to EtO for workers in an effort to then conduct an exposure response assessment in an epidemiologic mortality study. The first step in building the model was to determine the amount of industrial hygiene data that was available and suitable for use in the development of the exposure prediction model. The second step concerning building of the model itself took into consideration several variables including calendar year of operation, job, location or department, type of product produced, annual pounds of EtO gas mixture used, vessel volume, engineering controls in place, administrative controls, and process controls. Once developed, the model was subjected to rigid evaluations in an effort to verify that the model can be used reliably in prediction of historical exposures. The model as developed predicted EtO exposures within 1.1 parts per million (ppm) of the validation data set with a standard deviation of 3.7ppm. The arithmetic and geometric means of the 46 measurements in the validation data set were 4.6 and 2.2ppm, respectively. The authors conclude that the model as developed outperformed the panel of industrial hygienists relative to the validation data in terms of both bias and precision.
NIOSH-Author; Risk-analysis; Epidemiology; Laboratory-workers; Health-care-personnel; Mortality-data; Statistical-analysis; Mathematical-models; Industrial-hygiene
Issue of Publication
American Journal of Industrial Medicine
Page last reviewed: April 12, 2019
Content source: National Institute for Occupational Safety and Health Education and Information Division