A paradigm shift in patient satisfaction assessment.

The authors investigated the relationships between patients' reactions to health care attributes and their overall satisfaction with primary care. The study found the following: (1) patients' overall satisfaction levels are disproportionately influenced by low levels of their reactions (less satisfied) to the primary care attribute, rather than simply averaged out among attribute reactions. This is a noncompensatory relationship. (2) The marginal impact of primary care attributes on overall satisfaction decreases at higher levels of patients' reactions (more satisfied) to primary care attributes, indicating a nonlinear relationship. Patients combine their reactions to the health care attributes by means of noncompensatory and nonlinear models to form their overall satisfaction. Decision makers should selectively concentrate training resources on those areas of attributes showing high dissatisfaction rather than attempt to improve an attribute that showed the largest parameter estimate. This approach would not only save resources but result in better outcomes of patient satisfaction.