Abstract In 2001, the IRSST asked Dr. Rivard's research team to evaluate the PRÉVICAP program (PRÉVention des situations de handICAP au travail) in place in four Québec rehabilitation centres and involving 11 regional offices of the CSST. This pilot project, which was launched in 2000, was the result of an agreement between the CSST, the IRSST, and the Réseau en réadaptation au travail du Québec (RRTQ). It allowed for the management, under the PRÉVICAP program, of 571 workers who had sustained an employment injury between 2001 and 2004. The PRÉVICAP program is based on a series of innovative discussions and projects conducted in the field of occupational rehabilitation by the Université de Sherbrooke in the early 1990s, and is designed to foster workers' return to their pre-injury jobs. It consists among other things, of the early and interdisciplinary management of workers with musculoskeletal injuries (MSIs) who are at risk of long-term disability, in partnership with all the stakeholders impacted by the injury (the worker who sustained the employment injury, employer, CSST case managers and health professionals). The purpose of the evaluation was to provide the CSST with the elements needed to make an informed decision as to whether it would be in the CSST's interest to adopt the PRÉVICAP model to deal with the problem of MSI-related occupational disability, in light of the conditions required for its implementation (implementation analysis), its effectiveness (impact analysis) and its costs and performance (economic analysis). Several research methods were used, including a multiple-case study to document the level of and variations in program implementation in the four pilot regions, and a quasi-experimental study to assess the program's effectiveness and performance by comparing the situation for workers enrolled in the PRÉVICAP program (experimental group) to the situation for those who were not (control group) over a period of three years following the employment injury. The results of the implementation analysis showed that the program was successfully implemented in the CSST's rehabilitation centres and regional offices involved in the pilot project. The interviews of the OHS professionals at the CSST and the PRÉVICAP centres revealed that the implementation process was similar in all four regions. Case management usually began late relative to the injury event and was lengthy, starting an average of six months after the event and lasting an average of six months. The same problems were encountered in each region. Both the decision to implement the pilot project and the implementation process itself were perceived as being too centralized; there was no consensus as to the value of the program and the criteria defining the target population; the stakeholders did not clearly understand their role and communication among them was sometimes arduous, making it difficult to establish a partnership; the program was cumbersome from an administrative standpoint; and the active participation of all the stakeholders, particularly of the workers and their employers, was sometimes difficult, if not impossible, to obtain. The impact and economic analyses involved comparing the workers who had benefited from at least one PRÉVICAP intervention (10 hours of services) to workers who had received the usual services. Our evaluation suggests that the PRÉVICAP program produces better results than conventional management. The PRÉVICAP workers returned sustainably to their pre-injury job nearly three times faster and in greater numbers (55% versus 29% at two years post-event) than those in the control group. They also stopped receiving disability indemnities 1.7 times faster, which on average translated into a savings of five and a half months of income replacement indemnities (IRIs) over three years. Again on average, the cost of the PRÉVICAP program was high, i.e. $19,000 per worker, and the total cost of management over a three-year post-event period was 13% higher for a worker enrolled in the program than for one who received only the usual services ($60,873 versus $53,990). The difference in costs drops to 4% ($53,242 versus $51,003) if we exclude from the comparison the 22 workers whose cases entailed very high costs (over $119,000). Taking into account the efficiency gains in terms of indemnification days (income replacement indemnities) saved, the performance of the case management process including the PRÉVICAP program was statistically equal to that of management without the program if we consider that each day saved was worth $10; this performance is statistically superior, with an average estimated savings of $10,000 per worker if we are prepared to consider that each day saved is worth $60. The results suggest particularly high program effectiveness and performance in cases where the worker had not been indemnified by the CSST in the five years prior to the current indemnification period. Furthermore, the PRÉVICAP workers were very satisfied at having completed the program and more satisfied with the services received from the CSST than were the workers in the control group. Three years post-event, pain levels and functional disability levels were still high in both the PRÉVICAP and control group workers. The PRÉVICAP workers with back injuries were more impaired than the control group workers, whereas the reverse was true for the workers with neck and/or upper limb injuries. Again, three years post-event, the PRÉVICAP and control group workers made similar and even greater use of the medication, home support services and/or equipment required due to their injury. Workers with MSIs and who have been receiving indemnities for several months represent a vulnerable population in terms of long-term disability, a costly situation for the indemnifying agency. In fact, among the workers indemnified for an MSI, 20% are indemnified for more than three months, yet they account for 75% of the IRI costs. Our evaluation provides the first scientific evidence regarding the value of a PRÉVICAP-type program for such a population, and more specifically, for workers with no history of indemnification, who represent approximately three-quarters of this population. The PRÉVICAP program involves several stakeholders. In Québec, the large-scale implementation of a program of this nature poses a number of challenges. The results of the evaluation suggest that, as an innovation or new practice, the program would have greater chances of being accepted at the CSST by the heads of the regional offices and the rehabilitation counsellors if they were to participate in the decisions and processes related to the implementation right from the outset, if they subscribed to the philosophy underlying the intervention, and if they understood how the program works, as well as the respective roles of the various stakeholders and the program's target population. Since the program's success is largely contingent upon the beliefs and attitudes of the worker, employer and attending physician regarding the program, it may be advisable to develop a clear strategy for promoting the program in the eyes of these stakeholders. One component of this strategy would be passing on the scientific evidence available about the program's impact and performance. Lastly, it may be worthwhile to think of ways to improve the partnership and communication among the stakeholders and to lighten the administrative procedures associated with the program. The results of this evaluation may help in decision making about the relevance of implementing the PRÉVICAP program and about the implementation process itself, ultimately with the aim of promoting the return to work of individuals on long-term disability.