Summary Section 1 of Quebec’s Regulation respecting Occupational Health and Safety (ROHS) defines an “enclosed area” (confined space) as any area that is completely or partially enclosed and is not designed for human occupation, nor intended to be, but may occasionally be occupied for the performance of work. A worker may therefore enter such a space that (i) is not a workstation, (ii) has restricted access and (iii) represents a risk to health and safety. The issue of work in confined spaces cuts across a wide range of sectors: municipal, manufacturing, chemical, military, agricultural, construction and transportation. The most common confined spaces in industry are tanks, reservoirs, silos, vats, manholes, pits, sewers, piping and the tank cars or trucks that have certain characteristics defined in the regulations. Workers enter confined spaces to perform maintenance, manufacturing or other tasks (e.g., construction industry). The occupational health and safety hazards are primarily atmospheric, biological, physical and ergonomic. The risks run by workers who enter these confined spaces are potentially high because of the confinement, inadequate natural ventilation, need to work in isolation, and access, rescue and communication problems. Accidents are common. In Quebec, for instance, between 1998 and 2011, 40 fatalities occurred in 32 confined space accidents, accounting for 4% of the investigation reports of the Commission des normes, de l’équité, de la santé et de la sécurité du travail (CNESST). Confined space work is regulated in Quebec with respect to worker qualifications, hazard identification, atmospheric control, entry monitoring and rescue procedures. Canada has a standard for confined space work: CSA Z1006 – Management of Work in Confined Spaces. In practice, before starting work in a confined space, a qualified person must conduct a risk assessment in order to take appropriate risk elimination and reduction measures. The purpose of this research project is to contribute to the prevention of accidents in confined spaces by helping companies apply existing regulations. Two specific objectives are to (i) gain a better understanding of confined space risk management and identify issues based on the literature and field observations, and (ii) develop a confined space risk analysis and work categorization tool that meets the needs defined in the first stage of the project. The research method includes (i) a critical review of the literature on confined space risk management, (ii) an analysis of fatal confined space accident investigations in Quebec in order to identify failings that led to the accidents and (iii) visits to 15 companies and organizations that manage confined space entries for their workers and subcontractors. The findings showed that, first, the number of fatal accidents caused by an equipment energy control problem highlights the importance of mechanical hazards in confined spaces. A more multidisciplinary approach would therefore seem desirable. Second, the risk estimation and assessment stages are seldom dealt with formally in the literature, with the exception of atmospheric risks. The literature focuses primarily on identifying hazards related to different kinds of work in confined spaces. The main risk analysis tools suggested in the literature (e.g., checklists, risk matrices) are often incomplete and do not take into account certain specific factors such as the physical characteristics of the confined space, rescue conditions, the variety of hazards, or the physical and psychological condition of the person entering the space. Furthermore, none of the organizations visited estimated risks, relying instead solely on the experience of the permit issuer. In some cases, this way of working can lead to inaccurate assessments of risk (e.g., omission or underestimation) and possibly to inadequate risk reduction measures. Third, in the literature, the concept of similar confined space, which is intended to lighten the burden of risk analysis, is not supported with practical assessment criteria. The idea of categorizing confined spaces to facilitate risk management and communication is described in the literature, but is not used much in the field. Fourth, field visits revealed that most rescue procedures had neither been tested nor made available to the local fire department. Last, it should be noted that the literature stresses the importance of rescue procedures, worker training and the conditions to be met before entering confined spaces, but pays little attention to the safe design of such spaces, although that is the risk control measure that deserves the greatest emphasis. On the basis of these findings and standard ISO 12100 – Safety of Machinery – General Principles for Design – Risk Assessment and Risk Reduction, a five-step risk assessment tool was developed for confined spaces, in order to meet the project’s second objective. Step 1 of the tool consists of a list of 26 closed-ended questions intended to characterize the confined space, its environment and working conditions. The purpose of step 2 is to describe the accident process related to the risks identified by the tool user. Step 3 facilitates risk estimation using a risk matrix and criteria tailored to the context of confined spaces. For this purpose, risk estimation tool design criteria recently proposed with respect to machine safety were applied. Step 4 provides a graphic categorization by risk families and levels. Last, step 5 consists of a feedback loop that estimates residual risks once the risk reduction measures have been selected. The tool can be used to determine, on the basis of explicit criteria, whether two confined space assignments are actually identical, with a view to simplifying risk reduction work if possible. The tool can also help decide, a priori, on the basis of predetermined criteria, whether rescue without entry is possible and whether the residual risks are acceptable. Twenty-two confined space experts were asked to test the tool’s usefulness and relevance. It was also compared with other types of tools recommended in the literature or by businesses for analysing risks associated with work in confined spaces. The distinguishing characteristics of the tool are (i) the exhaustiveness and multidisciplinarity of its risk identification, (ii) the detailed selection criteria used to estimate risks, (iii) the use of the risk analysis results and (iv) the impact of the confined space risk reduction measures on risk parameters. This study will provide support for designers, prevention specialists and rescuers in their respective efforts to improve the health and safety conditions of workers who must enter confined spaces. The tool can be used to design a confined space or to assess an existing one.