Grain Dust and Health: A Competing Risk Analysis for the Grain Workers in Saskatchewan

Abstract

Grain dust industry workers are exposed to a number of work-related hazards, including high levels of endotoxin, microorganisms and dust. Multiple studies have reported immunological, toxicological and clinical effects of occupational exposure to grain dust contaminants. The study aims to determine the effects of various prognostic and demographic factors on health-related outcomes among the grain industry workers in Saskatchewan. Statistical analyses of the grain dust data can be carried out in a competing risk framework. In this context, competing risk is defined when an individual has a chance of getting one or more events to emulate with event of interest (e.g. death, time to relapse, time o disease type etc.). The competing risk analysis involves fitting the Cox PH model separately for each event type, treating the other (competing) event types as censored in addition to those who are censored from loss to follow-up or withdrawal. One of the assumptions of competing risk analysis is that censoring is independent of events regardless of the different ways that censoring can occur, including failure from competing risks other than the event-type of interest.

We define three competing events for the grain dust industry workers in Saskatchewan: chronic cough or phlegm, shortness of breath and allergy. Each worker can experience any of these events over the follow-up period from 1978 to 2005. We then consider seven covariates to assess their effects on the hazards of each of these three events: age, history of health problem (yes/no), history of asthma (yes/no), body mass index (BMI), forced expiratory volume in one second (FEV1), FEV1/FVC ratio which is the proportion of the amount of air exhaled in the first second (FEV1) to all of the air exhaled during a maximal exhalation (FVC) and smoking.

Our competing risk analyses reveal that FEV1/FVC ratio and smoking are highly significant to the risk of developing chronic cough or phlegm (p-value = 0.0238 and 0.0009 respectively). For shortness of breath, history of asthma and smoking are found significant, with p-values 0.0481 and 0.024, respectively. Results also indicate a high impact of age and FEV1 on allergy.

Our analyses are based on a relatively small sample (n = 226), and therefore caution should be applied to generalize our findings. Nevertheless, our findings could be useful for policy makers to make the environment of grain industries safe and secure for the workers with respect to standards and guidelines. Results could also be useful for human awareness.