Anesthesia care in the United States is provided by a number of different practice models, that could involve nurses, physician’s assistants, and physicians working in teams or alone. Much of the variation in care models evolved over time based on tradition, experience and political influences, rather than by design using evidence and cost-effectiveness analysis. In recent years, there has been increasing adoption of a team care model that requires less direct supervision from an anesthesiologist, as opposed to a physician led, medically-directed model. This model has lower labor costs, since it requires fewer, high-cost, anesthesiologists. In addition, this model allows for more flexible assignment of practitioners, based on the patient’s risk. This dissertation analyzed the patterns of distribution across different hospital sizes and types, and 4 different care models: (i) all physician providers; (ii) a high physician supervision model; (iii) a low physician supervision model; and (iv) all nurse anesthetist. Nurse anesthetist only practices are heavily centered around small, rural hospitals, whereas most other models are found in large, urban areas. Both surgical complexity and comorbidity scores were higher in physician or physician-led groups, but by very small margins. Analysis of 48 hour mortality showed higher death counts by hospital for large and teaching hospitals, and either supervised model. Overall differences between surgical complexity, comorbidity, and perioperative mortality sometimes attributed solely to anesthesia models, are significantly related to hospital level characteristics, regardless of the provider.