In a retrospective case-control study, structural characteristics of hospitals that closed during the years 1976-1980 were contrasted with three comparison groups: hospitals that were acquired in a merger; hospitals that joined a multihospital system; and hospitals that remained autonomously opened, to investigate these characteristics as predictors of closure. Characteristics investigated included environmental, structural, and process variables. The independent variables were measured 5 years prior to outcome. Findings indicate that closed hospitals resemble hospitals acquired in a merger ("failure"), and likewise autonomous hospitals resemble hospitals that join a multihospital system ("success"). The most important predictors of hospital failure were the physician-to-population ratio, the East North Central and West North Central census regions, the level of diversification, low occupancy rate, location in a standard metropolitan statistical area, the chief executive officer's lack of affiliation in the American College of Hospital Administrators, profit status, bed size of less than 50, and presence in a state with a rate-setting agency. Surprisingly, this study shows the bed-to-population ratio to be unrelated to closure. In addition, the findings strongly support the open-system perspective, which, unlike the closed-system perspective, is concerned with the vulnerability of the organization to the uncontrollable and often unpredictable influences of the environment.