| dc.description.abstract | Occupational Health and Safety Management Systems (OHSMS) play a fundamental
role within organizations. According to the International Labor Organization (ILO,
2020), despite greater attention to the issue by governments and companies, the
quantity of work-related accidents and illnesses, as well as their respective economic
impact, remain significant. Technical, behavioral, and economic aspects are affected
by the performance of these systems and, consequently, the general result of
organizations. In this context, the performance evaluation of the OHSMS represents
an important mechanism for reviewing the implemented measures concerning the
results obtained. However, this field of study has been neglected as conceptual
divergences are frequently observed in the proposed assessment instruments,
compromising the quality of the analysis of the results. Furthermore, organizations
evaluate the performance of the OHSMS predominantly through its efficacy (i.e.
achievement of objectives previously defined), without taking into account the
resources used, and without evaluating the impact of each initiative concerning the
results obtained. Therefore, we defend the thesis that the performance of OHSMS
should be complementarily evaluated through the analysis of efficiency, and that the
Data Envelopment Analysis (DEA) can contribute to this objective. This research
follows the paper-based thesis (PBT) model and has as its primary objective the
evaluation of the efficiency of Occupational Health and Safety Management Systems
(OHSMS) using DEA. The study contributes to filling theoretical gaps about the
conceptual accuracy of performance measures used in the OHSMS literature,
proposing a conceptual basis for this field of study. In addition, the proposed model
directly contributes to greater assertiveness in decision-making by organizations, by
aggregating the technical efficiency and the efficacy on the evaluation of the OHSMS.
The main limitations of the study are the lack of application of the model in external
benchmarking and the fact that the data used come from a single case. | en |
