Summary This study explored the potential for using the toxic agent exposure data collected in Québec by the Réseau public québécois en santé au travail (Québec public occupational health network) for more than 30 years. At the start of this project, it was estimated that more than 800,000 measurements had been carried out by the network's teams during this period, constituting an exceptionally rich pool of data that could be developed as a computerized database. In fact, this type of occupational exposure database (OEDB), several of which have been established in various countries, can be operated at the local and national levels for purposes of exposure monitoring, risk analysis, establishing standards, compensation, and scientific research. In Québec, the IRSST maintains an electronic database (LMS system) containing all of the analytical results related to the measurements made by CSSS (health and social services centre) occupational health teams since 1985. Intended for managing laboratories, its use as an OEDB is limited, due to the absence of some information on the exposure circumstances that led to each analysis, for example the worker's job. All of the contextual information is contained in the paper files of the occupational health teams, who in 1994 began using a computerized system, SMEST (Surveillance Médico-Environnementale de la Santé des Travailleurs, worker medico-environmental health surveillance) to record the exposure data collected in the companies in their territory. This database, containing the measurements entered inconsistently by the practitioners, was replaced between 2006 and 2007 with the current SISAT system (Système d'Information en Santé Au Travail, occupational health information system). Finally, it is important to mention the creation in 1992, in the framework of a research project, of the HYGIENE database containing all of the exposure data from the health files of establishments in the territory of the former community health department of Sacré-Coeur Hospital in Montréal, and covering the period from 1978 to 1992. The network's data were therefore the subject of several computerization projects or activities without a complete OEDB being created. This was the general objective of the present project of studying the feasibility and resources necessary for the creation of a Québec OEDB covering all of the data from the network's health files for the period preceding SISAT's introduction. More specifically, the aim of the project was: 1) to establish a "map" of the Network's historical measurements by compiling the collected samples with a stratification by measured substance, region involved, industrial activity sector, and time period covered, 2) to estimate the value of the SMEST database as an OEDB, by creating a descriptive profile of the data and by determining the proportion of the practitioners' measurements that were in fact recorded, 3) to illustrate the use of advanced statistical modeling methods to establish exposure profiles from the HYGIENE database data, and 4) to carry out a pilot computerization project on the exposure data in order to characterize the resources necessary for the creation of a potential exhaustive Québec OEDB. In total, The LIMS made it possible to document some 557,000 analyses associated with air samples over the 1985-2008 period, corresponding to 13,370 establishments covering 608 activity sectors. CSST Groups I and II contained 68% of the data. The Montréal, Montérégie and Mauricie-Bois-Francs regions represented 54% of the records, with the development or application of the health programs specific to the establishments accounting for 60.5%. Sixteen substances accounted for more than 10,000 data, representing 56% of the records. A tool (ÉDALI) freely accessible on the IRSST's Web site was created in the framework of this project to allow practitioners to explore these data. The SMEST data mainly covered the 1992-2000 period and consisted of some 16,000 quantitative evaluations on substances, and involved 3679 establishments and 2007 jobs. Five agents (welding fumes, carbon monoxide, styrene, crystalline silica and acetone) represented 44% of the data. Only 16% of the LIMS records could be associated with an evaluation in SMEST. Conversely, 51% of the SMEST data could be associated with a LIMS record. In total, the coverage rate was mediocre and varied greatly, depending on the agent, the period, and the region. HYGIENE statistical modeling involved the 2997 personal exposure measurements for the 46 solvents documented in the database. It revealed a reduction in levels over time, and the influence of various factors such as the duration of measurement, the season, the strategy used, and ventilation. The modeling also established a profile of the exposure levels for several job-economic activity combinations. Finally, a pilot database was created over a 4-month period from the computerization of the health files held by the CSSS de la Montagne (de la Montagne health and social services centre) in Montréal. This OEDB contains 4970 analytical results, on 118 agents in 192 files and 50 establishments. This exercise identified the contextual information most frequently documented in the files and highlighted the need for structuring profession and economic activity coding. In total, the computerization rate was 80 to 100 measurements per day, or a rate three times faster than during the creation of the HYGIENE database 20 years ago. By applying this rate to our estimate, based on the LIMS and the HYGIENE experience of 730,000 data in the network's files over the 1981-2008 period, including direct-reading instrument measurements, it is estimated that the creation of an exhaustive database would mobilize between 6 and 8 coding hygienist person-years in the Montréal region and between 29 and 39 for all of Québec. These estimates probably represent an overestimation of the real resources due to the economies of scale that would be achieved in a large project and mainly the improvement in the data input assistance tools available during the pilot study. In conclusion, this project led to the enhancement of IRSST laboratory data by producing a measuring activity mapping tool. It showed a very insufficient SMEST coverage rate for it to be used as an OEDB. The work on the HYGIENE database and on the pilot database revealed the feasibility of constructing a Québec OEDB covering all of the data from the network's health files in order to develop an information system on occupational exposures in Québec in the last 30 years.