Towards a better understanding of occupational accidents in Belgium

project carried out with the support of the Fund Abefradoc - Benevermedex, managed by the King Baudouin Foundation

Authors
Affiliation

Tom Geens, PhD, MsSc

Liantis

Marco Vandamme, MsSc

Heidi Janssens, PhD, MD

Dieter D’haeyere, MsSc

Maarten Van Mol, MD, MsSc

Published

October 29, 2025

Doi

10.5281/zenodo.17256999

Keywords

Occupational Accidents, Workplace Accidents, Accidents at work, Workplace injuries, Determinants, Factors, Cost, Occupational Safety, Occupational Risk, Commuting Accidents, Accident Frequency, Accident Severity

Preface

Contact

Liantis external service for prevention and protection at work
Sint-Clarastraat 48
B-8000 Bruges
Belgium

press & communication
+32 50 474197
communicatie@liantis.be

research & analytics
+32 50 474805
research@liantis.be

privacy & data protection
+32 50 474690
dpo@liantis.be

Funding

This project was carried out with the support of the Fund Abefradoc - Benevermedex,
managed by the King Baudouin Foundation (2022-J1162090-229330).

Availability

This report is available in two formats: a .pdf version and an .html version.

Both output formats have been produced using Quarto, an open-source scientific and technical publishing system for creating dynamic documents, reports, presentations, and websites using plain text formats such as Markdown. Quarto supports multiple programming languages and integrates with computational notebooks, enabling reproducible research workflows. It facilitates the embedding of code, outputs, and visualizations directly within the narrative, promoting transparency and reproducibility in scientific communication.

Intellectual property disclaimer

All content in this report -including text, figures, tables, and data- is the property of Liantis, unless otherwise stated. This work is protected under applicable copyright and intellectual property laws.

Third-party content is credited accordingly. If you believe any material has been used improperly, please contact Liantis for resolution.

Reproduction, distribution, or transmission of any part of this publication -by electronic, mechanical, photocopying, recording, or any other means- is strictly prohibited without prior written permission from Liantis.

The views and interpretations expressed in this study are those of the authors and do not necessarily reflect those of the funding bodies.

You may reproduce material from this study in non-commercial publications and presentations, provided both of the following conditions are met:

For more information, please contact research@liantis.be.

Privacy disclaimer

For this report, personal data as defined by the European General Data Protection Regulation (GDPR) has been processed. An analysis has been made on compliance with GDPR for all processing activities, such as e.g. collecting, structuring, storing, adapting and analyzing the data and for all categories of personal data, including special categories as defined by article 9 of the GDPR. Since the goal of the processing was scientific research, all processing activities were executed in accordance with the relevant GDPR articles and recitals. Article 89 (1), but also recitals 50, 52, 53, 62, 156 and 159 are especially relevant.

All GDPR principles were checked for compliance, hereafter a few examples are shown.

  • Lawfulness, fairness and transparency: a lawful ground for processing in art. 6 -and where applicable art. 9- was defined, and transparency requirements were checked, where art. 14, 5, (b) confirms that the requirements for the right to information were not applicable.
  • Purpose limitation: data from multiple sources was combined, which is justifiable since it was for scientific research. Health data was already processed for “preventative or occupational medicine”, as mentioned in art. 9, 2, (h) and payroll data was processed as well. Both are not incompatible with the initial purposes based on art. 89 (1) of the GDPR.
  • Data minimization: Only the data necessary for the research was processed.
  • Storage limitation: the necessary personal data was used to make a meaningful statistical analysis, pseudonymized where possible. An aggregation of the data was made as quick as possible to minimize the personal data processing and risk that comes with it and no personal data is used in the results of the report. This aligns perfectly with recital 162.

For more information, please contact dpo@liantis.be.

Why read this report?

According to Bart Steyaert (FEDRIS)

Bart Steyaert, Database Service - Studies and Development Department, FEDRIS

From the viewpoint of FEDRIS, this report, Towards a Better Understanding of Occupational Accidents in Belgium, offers a new and crucial view into a socially significant issue: the safety and well-being of workers. Occupational accidents are not only a source of human suffering but also place a substantial burden on our social security system and the economy. By deepening our understanding of the circumstances and consequences of workplace accidents, the report contributes to a more informed dialogue about prevention, re-integration, and policy development. It reinforces the importance of evidence-based measures to reduce workplace risks and improve outcomes for both workers and employers.

This collaboration with Liantis has been particularly valuable in achieving this goal. Their scientific cell brought technical expertise and advanced data processing capabilities that helped refine and enrich the raw data available to us. FEDRIS’s role in validating and clarifying the classification of work absences and occupational accident outcomes ensured a robust foundation for analysis. The result is a high-quality dataset and analysis that not only highlight existing risk patterns but also point to concrete levers for action. For FEDRIS, this opens doors to further collaboration with stakeholders, and offers new avenues for research and policy, especially around themes like partial return to work, sector-specific risks, and long-term injury impact. The integration of data science methods in this project illustrates the path forward toward a more proactive, evidence-based approach in managing and preventing occupational accidents in Belgium. In addition, this project provided useful feedback for FEDRIS, for instance with respect to improving the documentation of the used variables in the analysis that is presented in this report.

According to Jelle Neyens (Liantis)

Jelle Neyens, Chief Client Officer – Large Enterprises, Head of Occupational Health, Liantis

Occupational accidents remain a persistent and complex challenge across sectors and demographic groups. This study aims to advance our understanding of such incidents in Belgium by moving beyond traditional descriptive statistics and applying a rigorous multilevel analytical framework.

I am proud to present Towards a Better Understanding of Occupational Accidents in Belgium, which addresses a tangible source of human suffering and significant business impact. Drawing on a decade of data and -depending on the analysis- covering up to 200,000 occupational accidents across 20 sectors and nearly 70,000 companies, the research applies mixed-effects models to identify patterns and determinants of accident occurrence, reporting behaviour, and other relevant outcomes. By accounting for repeated observations within individuals and clustering within organizations, the models correct for biases and reveal the nuanced interplay between personal and structural determinants.

The strength of this study lies not only in the breadth of the dataset but also in the depth of collaboration among the involved parties. A thorough understanding of the data processing pipeline -spanning collection, validation, and cleaning- has enabled the team to resolve inconsistencies and ensure analytical robustness. This shared expertise has been instrumental in transforming raw administrative records into meaningful insights and is actively supporting the optimization of existing processes within and across the respective fields of work.

This report is not merely a statistical exercise; it is intended as a foundation for evidence-based decision-making. We hope it will inform policy development, enhance workplace safety strategies, and stimulate future research collaborations. The insights presented herein are designed to empower stakeholders -including insurers, employers, policy-makers, and researchers- with actionable knowledge that can drive meaningful change.

The range of conclusions emerging from this study confirms certain long-held assumptions while also revealing new insights. As with any robust scientific inquiry, it raises further questions that merit exploration. For instance, while it is commonly assumed that workers in smaller companies are less likely to take time off to recover from injuries, the data show that when they do, their recovery periods tend to be significantly longer. This observation prompts further investigation: Are their injuries more severe? Do larger firms offer more reintegration options? Or do they benefit from stronger HR policies that ensure better follow-up for absent workers?

This study not only validates and challenges existing beliefs; it also encourages us to ask new questions. These questions are essential as we continue striving toward a future in which occupational accidents and their consequences are systematically reduced.

Enjoy your reading.

How to read this report?

A good place to start is short summary (an abstract available in English, Dutch, or French, located immediately after the preface) or the extended summary, a more extended overview spanning several pages, found at the very end of the report just before the references section.

Both summaries follow the classical structure of the report, which is organized into five main sections, each serving a distinct purpose:

  1. Literature Overview. Provides background on occupational accidents, including definitions, a review of relevant literature, and identification of knowledge gaps. Recommended for readers seeking context and theoretical framing.

  2. Data Quality Report. Describes the origin, structure, and preparation of the data used in the analysis. It explains how different data sources interact, outlines encountered limitations, and details the steps taken to ensure data readiness for modelling. This section also assesses the quality of the source datasets and the additional outcomes and determinants gathered. For occupational accidents, this includes the content quality of XML notification messages sent via FEDRIS and KSZ automated flows, as well as the processed information stored in Liantis databases. Essential for readers interested in understanding the analytical foundation.

  3. Data Analytics Report. Presents the statistical results that underpin the study’s conclusions. It includes descriptive statistics, visualizations, and both univariable and multivariable modelling approaches to explore the occurrence and severity of occupational accidents. The analysis focuses on how various determinants contribute to different outcome measures, with results interpreted in light of existing literature. Recommended for readers seeking detailed insights into the modelling and analytical rationale.

  4. Practical Compendium. Summarizes key findings and offers practical advice for stakeholders such as employers, employees, and policymakers. Particularly relevant for readers interested in actionable insights and main takeaways.

  5. Recommendations for Knowledge Dissemination and Training. Provides initial suggestions for translating findings into training and communication strategies. Limited in scope and intended as a starting point for further development.

Since the report is extensive, readers are encouraged to focus on the sections most relevant to their interests and needs. Navigating such a large document can be challenging, so it’s helpful to keep the keyboard shortcuts Alt + Left Arrow and Alt + Right Arrow in mind to jump backward and forward between previously visited locations.

To further support smooth navigation, the authors have included a series of suggestions throughout the report, emphasizing different types of important messages. These selections are highlighted using four distinct types of Callout Blocks. Each type is visually unique, serves a specific purpose, and includes its own Previous and Next buttons to enable faster and easier navigation.

The four types used in this report are:

Insights and idea’s

Green boxes with a bulb icon highlight an insight or an idea for further research or analysis.

Information summaries

Blue boxes with an i-icon highlight a summary with important numbers to note or remember from the preceding paragraph(s).

Warnings

Orange boxes with an exclamation mark icon in a triangle highlight a warning e.g. concerning data quality issues or limitations of the analysis.

Dangers

Red boxes with an exclamation mark icon highlight dangers like strategic and ethic issues and require carefull attention on how to act upon the presented information.

What this study does not address

This study does not examine the specific characteristics of occupational accidents, such as their causes, types, or circumstances. Common accident mechanisms such as falls or slips are not addressed in this report. Instead, the focus lies on the outcomes of occupational accidents and the determinants associated with their occurrence and severity. The aim is to provide a broader understanding of the factors influencing these events and their consequences.

Given the wide range of possible outcomes, the analysis prioritizes generic indicators (e.g., the likelihood of reporting an occupational accident) over specific ones (e.g., probability of fatality or occurrence of a particular type of injury such as a fracture). This approach allows for a more generalizable and scalable interpretation of the data.

For detailed insights into the causes and types of occupational accidents in Belgium, readers are referred to the FEDRIS annual reports and other relevant literature in the field.

Acknowledgements

We extend our sincere gratitude to the Statistics Team and the Database Service of the Studies and Development Department at FEDRIS for their invaluable support throughout this project. Their expertise in clarifying the structure and interpretation of whole and partial (temporary) work absences, as well as their provision of validated data on accepted and refused occupational accident cases, was essential to our analysis. This study would not have been possible without their dedicated assistance and collaboration.

Our special thank herefore goes out to:

  • Bart Steyaert
  • Jocelyne Landries

We gratefully acknowledge the valuable contributions of our colleagues at Liantis, whose in-depth knowledge of processes and data flows related to occupational accidents significantly enriched this project. Their insights -ranging from operational and technical to managerial and communicative aspects- were instrumental in bringing this work to a successful conclusion.

We would like to specifically thank:

  • Bart Muylle – ESPP operational flow checks, proofreading and reality checks
  • Karlo Van der Plaetse – Service development potential and reality checks
  • Marc Durivet – PS operational data flows, internal documentation and payroll data
  • Koenraad Keirsschieter – IT PS-specific payroll calendar queries
  • Matthias Debeuf – IT ESPP operational OA flows and internal documentation
  • Britt Mallentjer – RS specific insurance queries, proofreading and reality checks
  • Isabelle Mattelaer – RS specific insurance queries, proofreading and reality checks
  • Christel Belmans – RS operational flows and insurance data
  • Ludo Houf – ESPP operational reporting and analysis
  • Christa Inghelbrecht – Internal prevention advisor, reality checks and client case study
  • Inge Taelman – HR expert advice, payroll flow and reality checks
  • Jan Vrielynck – ESPP permanent incapacity degree, medical context and client case study
  • Roger Collier – ESPP study implications for safety practice, original draft proofreading
  • Bart Geerts – ESPP literature study, original draft proofreading
  • Jos Schockaert – ESPP motivated SME policy advice
  • Thomas Hanson – Legal advice concerning the project
  • Wouter Bruggeman – IT PIM, storage and compute tips related to the project VM
  • Thomas Geens – IT backup and restore of the project VM
  • Alies De Kock – Marketing, PDF report front and back cover design
  • Arne Sevenhant – Marketing, HTML report look and feel and roll-out
  • Martijn Vanhinsberg – External communication preparation and roll-out
  • Jelle Neyens – Direction sponsorship

We also appreciate the valuable feedback and suggestions from professionals employed with Liantis customers that were willing to participate in case studies:

  • Jean-Marie Maes – internal prevention advisor, AZ Sint-Lucas
  • Koen Stoffelen – former head of HR administration, AZ Sint-Lucas
  • Wouter Spenninck – internal prevention advisor, Jan Yperman Ziekenhuis

Index