Objective: This paper presents the stages of development of an occupational biomechanical exposure matrix and preliminary reliability and validity indicators.
Methods: The expertise-based job exposure matrix, called "MADE" (for "difficult physical conditions and job matrix"), was developed from the French and international classification of jobs for 17 biomechanical exposures. Three pairs of investigators independently rate the frequency and intensity of exposure of each job from 0 to 5; discordant scores within pairs (mean difference >20% for each biomechanical exposure considered) were discussed collectively. Reliability was assessed by the difference between the initial ratings, the number of revised jobs, and the correlation between the two classifications (with transcoding). Validity was studied by correlating variables similar to those from existing United States and Danish matrices.
Results: The difference between the mean scores of the pairs was considered to be "fair" (less than one point). Of the 51 paired results studied (17 exposures - 3 pairs of investigators), four coefficients were found to be less than 0.6, and 14 were between 0.6 and 0.7. Forty-nine of the 1,169 job categories were reviewed. Correlation coefficients between the initial classification and transcoding were greater than 0.7. Correlation coefficients between the French, United States and Danish MEE variables were fair to good.
Conclusion: An expertise-based job exposure matrix with fairly reliable indicators has been developed, opening up the prospects to improve certain some fields of public health, at both national and international levels.