A taxonomy of health networks and systems: Bringing order out of chaos

Objective. To use existing theory and data for empirical development of a t axonomy that identifies clusters of organizations sharing common strategic/ structural features. Data Sources. Data from the 1994 and 1995 American Hospital Association Ann ual Surveys, which provide extensive data on hospital involvement in hospit al-led health networks and systems. Study Design. Theories of organization behavior and industrial organization economics were used to identify three strategic/structural dimensions: dif ferentiation, which refers to the number of different products/services alo ng a healthcare continuum; integration, which refers to mechanisms used to achieve unity of effort across organizational components; and centralizatio n, which relates to the extent to which activities take place at centralize d versus dispersed locations. These dimensions were applied to three compon ents of the health service/product continuum: hospital services, physician arrangements, and provider-based insurance activities. Data Extraction Methods. We identified 295 health systems and 274 health ne tworks across the United States in 1994, and 297 health systems and 306 hea lth networks in 1995 using AKA. data. Empirical measures aggregated individ ual hospital data to the health network and system level. Principal Findings. We identified a reliable, internally valid, and stable four-cluster solution for health networks and a five-cluster solution for h ealth systems. We found that differentiation and centralization were partic ularly important in distinguishing unique clusters of organizations. High d ifferentiation typically occurred with low centralization, which suggests t hat a broader scope of activity is more difficult to centrally coordinate. Integration was also important, but we found that health networks and syste ms typically engaged in both ownership-based and contractual-based integrat ion or they were not integrated at all. Conclusions. Overall, we were able to classify approximately 70 percent of hospitalled health networks and 90 percent of hospital-led health systems i nto well-defined organizational clusters. Given the widespread perception t hat organizational change in healthcare has been chaotic, our research sugg ests that important and meaningful similarities exist across many evolving organizations. The resulting taxonomy provides a new lexicon for researcher s, policymakers, and healthcare executives for characterizing key strategic and structural features of evolving organizations. The taxonomy also provi des a framework for future inquiry about the relationships between organiza tional strategy, structure, and performance, and for assessing policy issue s, such as Medicare Provider Sponsored Organizations, antitrust, and insura nce regulation.