Increasingly many firms have started to implement programs intended to improve the workers' health and the psychosocial work-environment, as well as other attributes of labor quality. Motivated by the need for evaluating to what extent the programs affect a firm's productivity performance, this study discusses a model for analyzing the contribution of labor quality attributes toward firm productivity. To assess the contribution from the labor quality attributes, we model firm productivity as the outcome of two separate processes within a firm: the physical production process and the labor quality process. Firm productivity is measured by a Malmquist-like productivity index and is computed by Data Envelopment Analysis. Based on bootstrap methods we analyze potential statistical bias and provide bias-corrected productivity estimates. The labor quality attributes are first modeled at an individual worker level as latent variables using Item Response Theory, and then aggregated to a firm-level. The model is empirically validated using data from three manufacturing plants that participated in a coordinated worksite health promotion program. Over a 4-year period (2000–2003), we observed a general improvement in efficiency of 2–5%, half of which could be attributed to an improvement in workers' health and psychosocial work-environment. A key benefit with the model is that it is practical, easy to implement, and very fast to compute. The model also constructively contributes to the discourse on sustainability by providing a framework for deriving meaningful metrics and providing tangible measurements on the effect of sustainability-related issues.