Analyses of database contents

The seed of our trials register was the systematic search strategy we built.

To first create a gold standard of articles that such a search strategy ought to find, we went through two whole volumes (years 2000 and 2001) of 19 biomedical journals of both occupational health specialty journals and non-occupational health journals that deal with occupational health topics.

Out of the 11 022 articles published in these 19 journals during these two years we found 149 occupational health intervention studies, yielding a prevalence rate of 1.35 occupational health articles per 100 articles published.

When we classified these 149 studies by study type we found 32 randomised controlled trials, 42 controlled studies, 10 interrupted time-series studies and 65 uncontrolled before-after studies.

Using our search strategy and these examples we began to scour PubMed for more studies. Since then we have developed the search strategy and our classification system further but the process of finding intervention studies, coding them and storing them in our database still continues. We have complemented our systematic searching with handsearches and other methods. That is why we now have 1896 references there.

Below is a picture of the relative proportions of different study designs according to the outcomes studied in the intervention studies included in our trials register. You can see that occupational exposures and diseases have been studied the most.