Reducing the recovery time for workers who are injured or disabled by a workplace accident is a key policy goal. This has motivated the promotion of employer-based return to work programs, despite a lack of systematic evidence on the effectiveness of such programs. The authors combine data on duration of time out of work for workers' compensation claimants with information on employer return to work programs to estimate the impact of the programs on time out of work. Discrete-time hazard estimates suggest that the workers in a program return approximately 1.4 times sooner compared to workers injured at a firm without a program. This corresponds to a reduction of between 3-4 weeks in median duration of time to return to work for all workers in our sample. The effect is stronger for men than for women, likely due to occupational differences between the two groups, and are robust across different specifications. Back-of-the-envelope calculations suggest that these programs are cost effective for large employers. More work is needed to determine whether these programs could be adopted successfully by smaller firms.