ISO TC 260, Human resource management, has developed a technical specification in which metrics for the incidence of occupational accidents are standardized internationally. At its meeting on 12 November 2019, the executive board of EUROGIP, the French occupational safety and health organisation, expressed strong reservations with respect to this project, and drew up a position paper (pdf), excerpts of which are reproduced here.
The specification ISO/TS 24179, Human resource management – Occupational health and safety metrics, is intended to support international standard ISO 30414, Human resource management – Guidelines for internal and external human capital reporting, which was published in 2018. The aim is to define metrics with which human resources managers in companies are able to demonstrate their added value. Examples of these metrics are the number of occupational accidents (whether they lead to incapacity for work or not) and fatalities at the workplace.
Rigorous recording and monitoring of occupational risks and accidents in the interests of better prevention is in principle desirable. Global metrics of accidents would however in no way enable companies around the world to be compared and the quality of their prevention measures thereby to be assessed. On the contrary, such metrics could even be counter-productive, for the reasons set out below.
The concept of an occupational accident
In most countries, the concept of an occupational accident is closely linked to the country’s accident insurance system. Multiple systems may exist side-by-side within a single country (e.g. statutory, private or sectoral accident insurance). Numerous examples show how widely the concept of an occupational accident varies. Some systems exclude from the definition all accidents that do not lead to incapacity for work or that result in absence from work only for a specified number of days. Others include all accidents. Furthermore, some systems do not recognize accidents as occupational accidents if they are caused by wilful action or misconduct on the part of the accident victim; in other systems, the company’s liability does extend to such cases. The precise criteria for reporting of an accident to be mandatory thus differ widely from one country to another.
Where a sudden death occurs at the workplace (e.g. due to a heart attack), many systems rule out its recognition as an occupational accident unless a link to the work can be demonstrated. In other systems, a death at the workplace is classified a priori as an occupational accident unless a non-occupational cause (such as a prior medical condition) is demonstrated.
These differences in the burden of proof already influence the results significantly, even in comparisons between only a small number of countries. In comparisons on a global scale, they would lead to even greater bias. Moreover, companies in systems with more comprehensive recognition of and compensation for accidents would be penalized when their results are compared to those of companies in countries whose policies for such compensation are much more restrictive or virtually non-existent.
In addition to the structural biases described above, comparisons of the number of occupational accidents would be statistically meaningful only for companies with a large workforce. An occupational accident is statistically a very rare event in a small or medium-sized enterprise (SME). Comparisons of metrics for the accident frequency in SMEs worldwide are therefore meaningless, since this would require reliable statistics for each enterprise over a number of decades. Depending on how contractors, subcontractors and temporary staff are included in the reporting, companies may be tempted to outsource particularly hazardous tasks.
ISO metrics such as these, which purchasers would use when evaluating suppliers, do not permit fair comparisons. Contrary to the actual objective, they would inevitably disadvantage companies or locations in countries with the most highly developed systems for occupational safety and health and accident insurance. In a worst-case scenario, these metrics, which are not necessarily consistent with the requirements of social insurance systems, could create an incentive, backed by an international standard, not to report accidents.
It is therefore imperative that the proponents of this initiative to develop a technical specification give consideration to the numerous biases and risks.