A Taxonomy of Hospitals Based on Partnerships for Population Health Improvement

Abstract

The U.S. health care delivery systems and hospitals face growing pressures and opportunities to engage in hospital-community partnerships that are a key component in addressing population health issues and a broad range of social determinants of health. However, the extent to which hospitals develop partnerships with community organizations varies widely, and little is known about the varying levels of hospitals’ partnerships or their integration of non-medical and medical services. This study aims to develop a taxonomy of hospitals based on their partnerships with external organizations and to assess the relationship between hospitals’ structural characteristics and their partnership development. Institutional theory suggests that hospitals generally explore community partnerships in response to widespread social and market forces. Contingency theory predicts that some hospitals are more likely to collaborate with non-health sector partners, whereas others may continue to work with familiar health care sector partners to focus on medical care, and still others may establish a balance between these two strategies. The theoretical framework of exploration vs. exploitation offers a persuasive lens through which to analyze hospitals’ approaches to developing hospital–community partnerships and addressing non-medical social determinants of health. Strategic decisions on partnership development can be interpreted as exploratory, exploitative, or ambidextrous. Developing a taxonomy is a useful approach for identifying variations in hospital–community partnerships by clustering hospitals into internally homogeneous types that are differentiated across the clusters. Based on 1,238 valid cases from the 2015 American Hospital Association Population Health Survey, this study uses survey items asking about the strength of partnerships for 36 partner organizations. Excluding three variables with low reliability, 33 variables are categorized into six partner groups by factor analysis. Cluster analysis is performed to develop a taxonomy of hospitals based on their partnerships with the six partner groups. The validity of the resulting taxonomy is confirmed through discriminant analysis. Cross-tabulation/chi-square independence tests are performed to compare hospitals’ structural characteristics - number of beds, governing authority, and teaching status - across the four cluster types. Finally, multinomial logistics regression analysis is used to investigate the relationships between hospitals’ structural characteristics and their partnership development. Of 1,238 hospitals, 26.1% are classified as an exploratory group that develops more collaborative partnerships with non-health sector organizations. Exploitative hospitals (18.3%) focus on partnerships with health care sector organizations. Ambidextrous hospitals (27.0%) develop collaborative relationships both inside and outside the health care sector. Finally, independent hospitals (28.6%) do not establish strong partnerships. Larger hospitals, not-for-profit hospitals, and teaching hospitals are more likely to be classified as exploratory or ambidextrous. The four-cluster taxonomy can provide policymakers and hospital managers with a better understanding of the considerable variation in hospital-community partnerships and enables more useful assessment of the collaborative relationships.