The data were collected in intervals termed cycles. The periods of cycles and the number of grain workers that participated in each cycle were as follows: October 1978 to September 1981, cycle 1 (n = 5,702); October 1981 to September 1984, cycle 2 (n = 5,491); October 1984 to September 1987, cycle 3 (n = 3,713); October 1987 to September 1990, cycle 4 (n = 2,847); and October 1990 to September 1993, cycle 5 (n = 3,079). The Environmental Surveillance Program ran parallel to the Grain Dust Medical Surveillance Program. Dust concentrations in workplaces were monitored. The specifications of this program are given elsewhere.
The number of observations contributed by Canadian grain workers stratified by region and cycle are given in Table 1. Data were collected on the following: company; province; region; type of grain elevator; age; height; weight; smoking information; lung function measurements (FEV1, FVC, mid-maximum flow rate, and FEV1/FVC ratio); respiratory symptoms (chronic cough, chronic sputum, chronic wheeze, and chronic dyspnea); and physician. Chronic symptoms were defined as follows: chronic cough, with cough in the morning or during the day or night for > 3 months a year for 2 years; chronic sputum, with phlegm in the morning or during the days for > 3 months a year for 2 years; chronic wheeze, with chest wheezing or whistling most days or nights; and chronic dyspnea, with shortness of breath on effort on the level or walking up a slight hill. read more
Mean and SDs were used to describe the distribution of continuous variables. The x2 test was used for comparisons of the categorical variables.
Marginal logistic regression models based on the generalized estimating equations approach were fitted using a procedure (PROC GENMOD; SAS Institute; Cary, NC) to determine significant predictors and then predict the probability of respiratory symptoms developing over time.
The generalized estimating equations approach, which incorporates within-subject correlation, was used to fit the multivariable model in order to determine the significant predictors of the respiratory symptoms. Marginal models for each of the respiratory symptoms (chronic wheeze, chronic dyspnea, chronic cough, and chronic sputum) were used to assess the longitudinal changes in the prevalence of respiratory symptoms among the Canadian cohort of male grain workers. In these models, the covariates used were: baseline height, current weight, current exposure (years in industry), smoking status, region, cycle, and an interaction term between years in industry and smoking status. These covariates for the multivariable model were selected based on standard model building strategies.
Table 1—Number of Observations Contributed by Canadian Grain Workers Stratified by Region and Cycle
|Regions||Cycle 1||Cycle 2||Cycle 3||Cycle 4||Cycle 5||Total|