Evolution of the occurrence of accidents at work in France from 1974 to 1994 : a number of observations and questions, what answers?

Reviewing the evolution of the main national safety results (frequency and gravity rates) from 1974 to 1994 highlights an improvement in accident levels at work.
Beyond this observation, several points in question can, however, be distinguished : differentiated evolution of the results of groups of activities, dispersion of results, increase in seriousness, regression during the period 1987 - 1990.
Such uncertainties serve as a reminder of the necessity both of giving the statistical data a meaning beyond a mere observation, and of imposing a change to the paradigm ; this change is illustrated by the explanations put forward for the aforementioned regression. The return to an improvement is also a reminder of the existence of institutional regulation mechanisms within the firm, bearing in mind that neither their reactivity nor extent is known.