Skip to Main content Skip to Navigation

Méthodologie d’analyse et de surveillance pour la prévention des arrêts maladie

Abstract : At a time when sick leave is a sign of growing ill-being for workers and a cost burden for the society, the systematic digitalization and distribution of data offers great opportunities for its prevention. We have therefore taken advantage of this opportunity to develop a range of prevention tools based on statistical analysis methods. In a first part, this work proposes an analysis of the mechanisms explaining sick leave among workers. The analysis of a national survey has first identified and prioritised their main determinants using random forest. Then, an analysis of administrative data had helped to identify absence trajectories that could lead to serious sick leaves thanks to sequential analyses and multi-state modelling. In a second step, tools were developed to identify abnormal situations of sick leave at company level. A company typology was first built to produce benchmark values for companies to accurately assess their situation. Finally, an algorithm for identifying absence peaks, adapted from epidemiological surveillance models, was finally developed to automatically identify companies in difficulty.
Document type :
Complete list of metadata
Contributor : ABES STAR :  Contact
Submitted on : Wednesday, February 24, 2021 - 4:24:20 PM
Last modification on : Friday, August 5, 2022 - 2:54:00 PM
Long-term archiving on: : Tuesday, May 25, 2021 - 6:51:01 PM


Version validated by the jury (STAR)


  • HAL Id : tel-03151304, version 1



Tom Duchemin. Méthodologie d’analyse et de surveillance pour la prévention des arrêts maladie. Statistiques [math.ST]. HESAM Université, 2020. Français. ⟨NNT : 2020HESAC027⟩. ⟨tel-03151304⟩



Record views


Files downloads