Systematic critical appraisal and clinical epidemiology as represented by the Cochrane Collaboration have provided the methodology for the formulation of clinical ‘practice guidelines’ in many specialties and in primary care. Practice guidelines are valuable in establishing the standard of care, controlling costs, and improving consistency of outcomes. Critics charge that practice guidelines suppress innovation, constrain the practice of medicine, ignores differences in patient population, and are overly rigid given individual patient characteristics. The application of clinical guidelines to occupational medicine in the United States was pioneered by the American College of Occupational and Environmental Medicine (ACOEM), which developed and maintains the ACOEM Practice Guidelines (APG) using a Cochrane-like system of expert panels reviewing peer-reviewed evidence. The APG has become a widely accepted authoritative reference used especially workers’ compensation systems to establish the standard of care. Problems include: Uneven evidentiary base (abundant and uninformative in the case of back pain, scanty and largely irrelevant for elbow and other issues); absence of clinical trials for occupationally-relevant outcomes; obstacles to performing outcomes studies and workplace intervention trials for injuries; characteristics of occupational disease such as latency, infrequency, and certainty of diagnosis; opposition by practitioners, mostly outside of occupational medicine; rigidity in application, not taking into account co-morbidity, patient characteristics, and complications; over-reliance on meta-analysis (which could be corrected by Bayesian analysis). Effectiveness of practice guidelines in occupational medicine has not been systematically studied except for cost, however. Performance assessment has been anecdotal. After having overseen the early APG program, the author’s view was that practice guidelines for were a beneficial but partial solution, addressing different issues than in general medicine. The opinions expressed here are those of the authors and do not reflect a position of ACOEM.