In most cases of shoulder injury at work, causation analysis is not clear-cut and requires detailed, thoughtful, and time-consuming causation analysis; traditionally, physicians have approached this in a cursory manner, often presenting their findings as an opinion. An established method of causation analysis using six steps is outlined in the American College of Occupational and Environmental Medicine Guidelines and in the AMA Guides to the Evaluation of Disease and Injury Causation, Second Edition, as follows: 1) collect evidence of disease; 2) collect epidemiological data; 3) collect evidence of exposure; 4) collect other relevant factors; 5) evaluate the validity of the evidence; and 6) write a report with evaluation and conclusions. Evaluators also should recognize that thresholds for causation vary by state and are based on specific statutes or case law. Three cases illustrate evidence-based causation analysis using the six steps and illustrate how examiners can form well-founded opinions about whether a given condition is work related, nonoccupational, or some combination of these. An evaluator's causal conclusions should be rational, should be consistent with the facts of the individual case and medical literature, and should cite pertinent references. The opinion should be stated “to a reasonable degree of medical probability,” on a “more-probable-than-not” basis, or using a suitable phrase that meets the legal threshold in the applicable jurisdiction.