A comprehensive location-allocation method for specialized healthcare services

This paper focuses on the development, solution, and application of a location-allocation model for specialized health care services such as the treatment and rehabilitation necessary for strokes or traumatic brain injuries. The model is based on our experience with the Department of Veterans Affairs’ integrated service networks. The model minimizes the total cost borne by the health system and its patients and incorporates admission acuity levels, service proportion requirements, and admission retention rates. A common resource constraint is introduced at the facility level since treatment of multiple acuity levels involves the pooling of common resources. Realistic instances of the model with 20 potential service locations, 50 admission districts and up to five open treatment units for three levels of severity are solved in about 300 seconds. The applicability of the model is tested by an extensive managerial experiment using data derived from one of the Department of Veterans Affairs specialized healthcare services. We investigate the effects of five critical factors: (1) the degree of service centralization, (2) service level mandates by acuity, (3) lost admission cost by acuity, (4) facility overload penalty cost by acuity and (5) target utilization level by acuity and treatment unit. We examine the countervailing forces present in making healthcare service location decisions and the resulting tradeoffs from the implicitly multiobjective nature of the system. The experiment and analysis demonstrate that the major factors of the experiment have a significant bearing on the optimal assignment of admission districts to treatment units.