Estimating time trends of quartz exposure in Swedish iron foundries –predicting today’s and yesterday’s exposure

Introduction
In Swedish foundries there is long tradition of enforced workplace surveys with measurements of quartz from the late 1970s onwards. The availability of exposure information over almost 40 years presents a unique opportunity to study trends over time and to evaluate the validity of exposure models based on shorter periods over time.

Methods
The database of 2333 quartz exposure measurements from the11 Swedish iron foundries divided into 11 specific job titles were separated in 3 models regarding time periods reflecting work conditions and production processes; historical model (1968-1989), development model (1990-2004) and validation model (2005–2006). A linear mixed-effect model for repeated measurements was used to describe trends over time. A linear regression analysis was performed to investigate agreement between the models. For each combination of job title and company size the average exposure was estimated.

Results
A strong reduction in exposure was seen in the periods 1968-1974 (51 % reduction) and 1975-1979 (28 %). In later periods quartz exposure was reduced by at best 8 % per 5 years. Without adjusting for this trend over time, predicting early historical exposures from our development model resulted in a significant regression coefficient of 2.42 and an R-square of 0.81, indicating underestimation of historical exposure levels. Comparing our development model with our validation model resulted in significant regression coefficient of 0.31, indicating overestimation of current exposure levels.

Discussion
To investigate long-term trends in quartz exposure over time, overall linear trends could be described by using mixed model analysis. For use as a basis for creating individual dose measures in purpose to predict historical exposures, a resolution of time periods, jobs, company and size is necessary to provide. Our comprehensive measurement database was also able to detect systematic changes in concentration levels, well reflected in our mixed models for different time periods.