Long-term trends in occupational exposure: Are they real? What causes them? What shall we do with them?

The Beatles already knew it in 1967 when they sang‘‘It’s getting better all the time’’. Scientific evidencethat this holds true for occupational exposures hasonly been obtained very recently (Symanski et al.,1998a,b). Circumstantial evidence has been aroundfor much longer. Just a chat with an experiencedworker or a glance at old workplace photographsconveys the same notion.Symanski et al. reviewed the literature for longi-tudinally collected sets of data that allowed them toquantify long-term trends in exposure to chemicalagents. On the basis of 696 data sets published overa 30-year period (1967–1996), they observed a me-dian annual trend of y6% with an inter-quartilerange of y1% to y11%. Such a trend results ap-proximately in a halving of airborne concentrationsevery decade. Obtaining a clear insight in the under-lying dynamics was hampered by lack of infor-mation on measurement strategies, monitoring andanalytical methods. So, were the observed trendsreal or an artefact of changes in study methods?In the current volume of this journal three papershave been published describing long-term trends inoccupational exposures across three industriesbased on large sets of exposure data from the pav-ing, the carbon-black and the rubber manufacturingindustries (Burstyn et al., 2000; van Tongeren et al.,2000, Vermeulen et al., 2000). Two of these studieswere based on data collected in a similar fashionover a period of almost a decade (van Tongeren etal., 2000, Vermeulen et al., 2000). The third studywas based on a database that comprised individualmeasurement data from surveys in severalEuropean countries collected over more than twodecades (Burstyn et al., 2000). Remarkably, allthree studies present downward long-term trendsthat fall almost entirely within the inter-quartilerange described by Symanski and co-workers(1998b) (paving: y6% for bitumen fume, y14% forbitumen vapour and y11% for PAH; carbon-black:y7% to y13% for inhalable dust and y6% toy10% for respirable dust; rubber manufacturing:y6% for inhalable particulate and y7% for dermalexposure). Another recent study of trends in ex-posure to wood dust in the USA describes a similarannual reduction of 7% over the 1979–1997 period(Teschke et al., 1999).The paper by Vermeulen et al. in this issue isvery unusual. Not only the companies, measure-ment strategy and sampling and analytical methodsremained unchanged over time, but also detailed in-formation was collected on the evolution of theproduction process and control measures during thenine year period. Using this information, theauthors were able to explain more than 80% of thehistoric decline in mean inhalable particulate con-centration. For dermal exposure, only 30% of thedecline was explained by these factors. Eliminationof a source was proven to be the most e•ective con-trol measure for both inhalation and dermal ex-posures, reducing them 67% and 66%, respectively.Engineering measures to control exposure weresomewhat less e•ective, but were still associatedwith statistically significant exposure reductions of34% and 49%, respectively. Surprisingly, measuresaimed at reduction of emission did not appear tocontribute in general to a decrease in exposurelevels. Measures classified under this headingresulted both in decreased and increased exposure.Thus, it seems that the postulated long-term