Over the past decade, consumer-oriented health care has emerged as the preeminent model for service delivery. This model requires an effective blend of clinical, managerial, and consumer input.1 The emerging model should reflect a melding of professional provider and consumer perspectives. At the union of these perspectives is a demand for quality health care services. However, while health care providers are in a position to judge the biomedical and technical aspects of care, it has been argued that consumers do not have the capacity to do so.2

The consumer, for the most part, has not been involved in defining health services quality. This is true primarily because the core of the quality definition in health care has been thought of as a balance of health benefits and health harms and therefore judgments regarding the essential elements of care are best left to the providers.1 Attempts to rectify this evaluative imbalance in judging health care quality no doubt have contributed to the growth of a vast and burgeoning patient satisfaction measurement industry that purports to give patients a voice in expressing their views about the care they received.

An examination of the patient satisfaction literature produced over the past decade reflects work that, unlike the consumer satisfaction studies found in the general marketing literature, generally consists of substantively shallow, empirical reports that address circumscribed, idiosyncratic, or context-specific questions. Much of the work comprising this literature is based on simple descriptive and correlational analyses that appear devoid of any theoretically or substantively compelling framework. Examples of notable exceptions to this criticism are the works of Ware and Davies,3 Bowers, Swan, and Koehler,4 Cleary and McNeil,2 and Taylor and Cronin.5

The Otani, Kurz, Burroughs, and Waterman article included in this issue is a paper that also represents the type of research that can move us in a direction toward understanding where patient satisfaction fits within health care. The authors employ well-reasoned and empirically sophisticated models to examine how patients cognitively process their experiences with hospital attributes. Specifically, they contrast a compensatory model where patients mathematically average appraisals of hospital attributes with a noncompensatory model in which an overall evaluation is weighted disproportionately by positive or negative perceptions of health care attributes. In addition, they examine the functional relationship (linear and nonlinear) of patients' attribute perceptions with overall satisfaction and behavioral intentions.

Generally, the findings are somewhat equivocal. With respect to the effects on overall satisfaction, it was shown that patients tend to use a noncompensatory strategy in which negative appraisals are given more predictive value than positive ones. However, in the prediction of behavioral intentions, it was observed that patients tend to use a compensatory mechanism in which the effects of negatively viewed attributes can be reduced (compensated) or balanced out by positive appraisals. Equally interesting are the results from an analysis of a curvilinear model that demonstrate a diminishing marginal return as a function of increasingly positive reactions to appraised attributes such as nursing care, compassion to family and friends, and pleasantness of surroundings. The authors speculate that for both outcomes, efforts to increase patient satisfaction levels beyond a threshold are not likely to be associated with further increases in preferred outcomes. From a managerial perspective, the results of this study generally suggest that desired outcomes are likely to be obtained when efforts are designed to identify and address negative reactions to health care attributes. This would suggest that benefits might accrue from proactive problem-solving approaches as opposed to efforts intended to increase satisfaction globally. The thoughtfully innovative approach used by these authors is reminiscent of work by Tversky and Simonson6 and clearly differentiates itself from the standard patient satisfaction study. Notwithstanding some of the methodological limitations noted by the authors themselves, their article clearly provides readers with much to consider.

Our own review of the state of patient satisfaction literature yielded two major observations. First, there is a large industry that has emerged to offer assessment and prescriptive services intended to measure and improve patient satisfaction. Second, the patient satisfaction literature yielded literally hundreds of small-scale studies conducted in the empiricist tradition. For the most part, these studies employed simple descriptive and correlational analyses and were conducted in as many different health care settings and clinical/service delivery areas as one could envision. To put it crudely, much of this work seemed idiosyncratic and domain specific and dealt with "satisfaction with this and with that." Our perception is that health care management practitioners often make inferences that inform strategic decisions. This may create problems because the resulting practice decisions are often based on evidence that is weighted beyond its merits.

Fundamentally, most of the studies we reviewed were based on one or two assumptions: (a) patient satisfaction was viewed as a desired objective exclusive of any other objectives or (b) patient satisfaction was specified as a predictor of outcomes that ought to be of value to health care organizations. It is difficult to find any fault with the notion that patient satisfaction is something that health care organizations should seek to achieve. Certainly, satisfied patients are preferable to dissatisfied ones. Moreover, it would appear that an emphasis on patient-centered care7 and the development and enhancement of patient-oriented organizational cultures must consider patient satisfaction. But the real question that has not, in our opinion, been addressed is why? What are the real benefits of patient satisfaction? What effects on other outcomes might derive from it? Press, for example, has argued that patient satisfaction will result in higher quality of care, higher satisfaction of employees, lower employee turn over, improved financial health, strengthened competitive position, enhanced placebo effect, and better risk management.8

Confusion between patient perceptions/ratings of service quality and the patient satisfaction concepts is apparently widespread. This confounding of constructs, or interchanging of terms, is seen commonly in the literature and among health care management practitioners. The relationship between satisfaction and service quality has been extensively modeled and analyzed by Steven Taylor and J. Joseph Cronin.5 One of their key findings is support for a model depicting a nonrecursive relationship [mutual causality] between service quality and satisfaction. In other words, increases in patient satisfaction lead to higher service quality perceptions and vice versa. They noted that the discovery of the nonrecursive relationship between these constructs may provide greater insight into the "continuing mystery about the causal order of these important constructs."5^(p.42) Perhaps what this indicates is that our efforts should be directed at improving the technical/clinical and functional quality [and patient-centered ethos] of our services as opposed to a slavish devotion to patient satisfaction measures. A focus on the latter might be akin to climbing the signposts instead of following the road.

The literature provides numerous inferences and extrapolations regarding the influence of patient satisfaction on various performance outcomes such as patient retention and profitability. The implied relationships, however, are not always based on solid empirical support. For example, evidence for the value relevance of customer satisfaction on financial performance across various industry sectors is at best equivocal. Consistent with the diminishing marginal utility finding observed by Otani et al., others have suggested that continuing efforts required to improve customer satisfaction beyond a certain threshold may produce negative effects for the firm.9 Clearly, there are also important mediating or moderating covariates (e.g., case mix, age, payer mix) that must be considered in order to derive a thorough understanding of the mechanism by which patient satisfaction influences outcomes. Fundamentally, patients' consumption of health care services is a very complex phenomenon that is likely influenced by many factors beyond a patient's control or even awareness. Some of these exogenous factors could include convenience of location, physicians' admitting privileges, referral practices and constraints of primary care physicians; organizational ownership and size; local marker characteristics such as concentration and competition among health care organizations; employee satisfaction, turnover, job tenure, training, and levels of service orientation; patient health status, acuity level, and socioeconomic status. In addition, the randomness of medical events must be considered as should the idea that a patient may never have need (nor desire) to return again to utilize a specific health care procedure or service.

In addition to various exogenous factors that may confound the relationship between patient satisfaction and elements of organizational performance, there are other questions that should be addressed before we can be confident that prescriptive efforts to enhance patient satisfaction will pay dividends. For example, answers must be found to questions concerning the nature of the cause and effect relationship and the small amount of observed covariation between patient satisfaction and measured outcomes. Furthermore, questions and problems concerning the rather dubious psychometric qualities of patient satisfaction instruments must be resolved. The responses to patient satisfaction questions tend to be non-normally distributed with striking ceiling effects and tend to produce modest differentiation in scores across units of analysis.

In summary, it appears that patient satisfaction is not a well-understood phenomenon. There is some evidence to indicate that patient satisfaction is likely related to some patient-centered aspects of care. It also appears that patient satisfaction is an embedded feature of health care quality. It is perhaps not surprising, given the complexities of modern health care delivery and patient behavior, that many issues concerning patient satisfaction have yet to be resolved. In all likelihood, there is something of substance to patient satisfaction. However, rigorous systematic studies framed within meaningful theoretical models should be undertaken to understand the actual role patient satisfaction plays in the delivery of patient-centered health care.

REFERENCES