OBJECTIVES: Undiagnosed obstructive sleep apnea (“OSA”) is associated with decreased workplace performance and increased mortality.  While the diagnosis and treatment of OSA in symptomatic individuals is highly cost-effective, it is unknown whether screening for OSA in the workplace can be cost-effective. METHODS: We modeled three strategies in a hypothetical cohort of 50 year old men:  (1) No screening or intervention for OSA. (2) Refer all individuals for lab-based diagnostic polysomnography, followed by continuous positive airway pressure (“CPAP”) therapy in those diagnosed with OSA.  (3) Screen individuals with a validated instrument (Berlin Questionnaire) delivered via email, followed by referral for polysomonography for those who screen positive and CPAP therapy in those diagnosed with OSA.  Costs of managing the screening program, as well as various incentives to improve survey response, were also included in the model.  Estimation of treatment benefits were taken from a previously published model (Pietzsch, et al).  We took a societal perspective and a lifetime horizon.  RESULTS: The incremental cost-effectiveness ratio (“ICER”) for the Berlin Questionnaire strategy compared to no screening, was $41,749/QALY.  By comparison, the ICER for the all-polysomnography strategy compared to no screening was $66,711/QALY.  We then considered eight possible strategies to improve Berlin Questionnaire response rates and plotted them on a cost effectiveness frontier, and found that with maximum enhancement of survey response, the ICER decreased to $32,484/QALY for the Berlin Questionnaire strategy.  The cost of screening (not including diagnosis) without survey enhancement was $11/person; with maximal survey enhancement, it was $40/person.  One-way sensitivity analysis found that the ICER was most sensitive to the size of the screening population, the prevalence of OSA, and the clinical benefit of OSA treatment. CONCLUSIONS: Screening for OSA in the workplace using the Berlin Questionnaire can be cost-effective, particularly with use of survey response enhancement techniques.