Jean Campbell, Ph.D.

Missouri Institute of Mental Health

5400 Arsenal Street

St. Louis, MO 63139

The Desire of Happiness is not absolutely secure of satisfaction, but only upon condition.

Beatitudes (Norris,1694)

All within Elinor's breast was satisfaction, silent and strong.

Sense and Sensibility (Austen, 1798)

The savage thinks only of present satisfaction, and leaves future satisfactions uncared for.

Contemporary Review (Spencer, 1883)

Flankers...was therefore begun to be governed by forreiners, with such jealousies,...which engendered those ill humors, and gave that bad satisfaction, which was the rise of the civil war that insued.

Political Touchstone (Momn, 1656)

the ecology of quality and performance assessment in health care has been restructured to emphasize consumer values, community care, and broadened measurements of outcomes (Campbell, 1997). This has led to the use of consumer satisfaction to evaluate psychiatric services as an important measure of quality for policy-makers, mental health professionals, administrators, advocates, and service recipients alike, creating opportunities for partnerships in accountability (Campbell, 1996). The emphasis on satisfaction is consistent with the broader trend towards consumerism which holds those that control and provide services are accountable to their customers in ways that go beyond the operations of the marketplace.

Consumer satisfaction is generally defined as the consumer's view of services received and the results of the treatment. It has been used by program evaluators to enhance health care providers' ability to render services that meet consumers' needs (Sheppard, 1993). However, important methodological concerns have limited the value of consumer satisfaction assessments (Elbeck, 1992). For instance, most satisfaction instruments are simple ad hoc measures. They fail to address satisfaction at any level of specificity, i.e., in terms of particular service recipient groups, treatments, or types of facilities (Lebow, 1983c). It is equally significant that they also do not measure consumer dissatisfaction. In their construction, the focus has often been on the kind of issues that mental health staff members feel are important, rather than issues that form the major areas of consumer discontent. They generally fail to take into account the relative (weighted) importance of such concerns. In addition, only recently has research emerged that examines how coercion and the subsequent consequences such as anger, fear, sadness, and confusion, are related to consumer satisfaction (Gardner, Hoge, Bennett, Roth, Lidz, Monahan, & Mulvey, 1993; Hiday, 1994).

Accordingly, even with its popularity, utilization of satisfaction as a measure of quality of services delivery and clinical effectiveness has been somewhat problematical. In particular, the high satisfaction rates reported in most studies (Lebow, 1982), is contradicted by qualitative data sources that stand as a robust critique of mental health service delivery systems and treatments (Chamberlin, 1978; Leete, 1988; Campbell & Schraiber, 1989; Susko, 1991). In the wake of satisfaction data that is consistent with the expressed views of consumers, recipients of mental health services have begun to develop new methods for assessing satisfaction and outcomes and using the information for system reform. To the extent that mental health consumers and professionals have embraced the concept that consumer involvement in quality of care monitoring is useful, cost-effective, and a consumer's right, technologies have been developed to facilitate the process.

Consumers have identified performance indicators (Consumer/Survivor Mental Health Research and Policy Work Group, 1992; Trochim, Dumont & Campbell, 1993; The American College of Mental Health Administration, 1997), collaborated in the development of a mental health report card (CMHS releases, 1996; Teague et al., 1997) and designed and conducted consumer self-report surveys (Virginia Mental Health Consumer Association, 1992; Belcher & Johnson, 1996; Belcher & Campbell, 1998). Key activities of consumers in evaluating the quality of psychiatric care include the following: (1) the formulation of assessment teams (Fricks, 1995; Ferry, 1996; Bluebird, 1997; LA Office of Mental Health et al., 1998); (2) the articulation of the consumer perspective through dialogues (Loder & Glover, 1992; Blanch et al., 1993; Bluebird, 1997), focus groups (Abramczyk, 1995; Carpinello, Kinight & Jatulis, 1992; Midgley et al., 1994) and concept-mapping pilots (Trochim, Dumont & Campbell, 1993; Dumont & Campbell, 1994; Campbell & McChesney, 1998); and, (3) research on self-help (Chamberlin, Rogers & Ellison, 1996; Kaufmann, Ward-Colasante & Farmer, 1993), well-being and coercion (Campbell & Schraiber, 1989; Campbell, 1992), empowerment (Segal, Silverman & Temkin, 1995; Rogers et al., 1997), and recovery (Dumont, 1995; Ralph, Lambric & Steele, 1996; Markowitz et al., 1996). The central characteristic of most mental health consumer evaluations is that they are grounded in the values of the consumer movement, and proceed consensually from a shared vision of what is quality in mental health services, and what should be measured (Campbell, 1997).

The importance of patient satisfaction has had a long history of debate, beginning over two milleniums ago in ancient Rome. Plato suggested in The Statesman that since the doctor "cuts us up, and orders us to bring him money...as if he were exacting tribute...he should be put under rigid control," and that this could be done by calling an assembly of the people and inviting opinions about "disease and how drugs and surgical instruments should be applied to patients."

Further, he proposed that the people "elect our physicians from among our number for one year terms, and severely penalize them if they fail to carry out the letter of the law."

In the past 35 years, the steady increase in the use of satisfaction assessments in health care evaluations is evidenced by the growth of empirical and theoretical publications regarding satisfaction that have appeared. In this review over 400 theoretical and empirical articles and reports were identified in the published literature from a computerized library search by crossing the key words satisfaction, consumer satisfaction, client satisfaction, report cards, and evaluation with health care and mental health services. To date, the work of consumers in evaluation and outcome assessment has largely gone unnoticed by professional evaluators because little has been published in peer-reviewed journals (Sherman & Kaufmann, 1995). An additional 40 citations on mental health satisfaction that involved consumers in evaluation projects were also collected from predominately "fugitive literature", conference presentations, and agency reports. Moreover, Psychological Abstracts (1982 to 1998) was systematically consulted and references found were classified into empirical studies (either health or mental health), and reviews and chapters. A considerable effort was made to collect all relevant published studies; however, the varied locations in which satisfaction materials appear and the inconsistent labeling of studies limits the comprehensiveness of this review.

About half of the references are drawn from the field of general health, with the remainder focusing on mental health services exclusively. Lebow (1983) compared the literature assessing consumer satisfaction with health care and mental health treatment and found that the similarities far outweighed the differences, although the mental and physical health fields had been relatively independent. Since evaluation methods in the health and mental health have converged even more in recent years, studies in mental health satisfaction are reviewed as a subset of the general category of health care.

The literature was analyzed with the following goals in mind: (1) to define the concept of consumer satisfaction and the a priori assumptions that underlie its use and implementation; (2) to identify its major dimensions from the content of available survey instruments and interview protocols; (3) to evaluate the state of the art of measuring consumer satisfaction, focusing on issues of reliability and validity of reported instruments, and identifying exemplary practices; (4) to assess the usefulness of consumer satisfaction as a continuous improvement technology for quality, appropriateness, utilization, and effectiveness of mental health care; and, (5) to articulate, where information is available, some "best practices" to assess consumer satisfaction.

To guide the review, a standardized approach adapted from Ware et al. (1978) was used to determine the following:

• whether the article is about health or mental health;
• the role of consumers;
• whether data is qualitative or quantitative;
• whether original data were reported;
• the intended use of the measures;
• design of the study;
• methods used to gather data (administration);
• whether verbatim questionnaire items were published or described;
• the type of scale construction methods used, if any;
• whether reliability data were reported;
• whether data relevant to validity were reported;
• variables that were studied in relation to satisfaction scores;
• privacy and confidentiality protections;
• and, major findings.

Nearly two-thirds of the publications (291 of 440) were empirical, including previously unreported analysis of satisfaction. Trained interviewers were used to gather data in 85 studies; questionnaires were self-administered in 113 studies; data-gathering methods could not be determined from the remaining articles. In 72 studies, single-item measures of satisfaction were used; 41 defined satisfaction by combining one or more scores into a scale. In most cases, the number of satisfaction dimensions tended to be small, or this topic was not addressed in the article. One fourth (10 of 41) of those constructing scales computed only one score, with the rest between 2 to 20 scores. In all but a few instances, where only single-item measures were analyzed, correlations among satisfaction scores were not reported.

From a content analysis of questionnaire items and responses to open-ended questions that were reported in articles on mental health services, 15 distinguishable dimensions that constituted the major areas of satisfaction and dissatisfaction assessment with psychiatric care were identified. These included overall satisfaction, participation in treatment, recognition of needs, respectfulness and empathy of care providers, amount of information provided and communication skills of clinicians, technical quality of care, understanding of problems, accessibility/availability/convenience, finances, physical environment, appropriateness of services, coercion/restrictiveness, continuity, cultural competence, and outcomes/improvement. Inconsistencies of satisfaction constructs prevailed in the literature. Not all investigators explicitly asked about satisfaction with providers or services, and terms such as "attitudes", "beliefs", and "perceptions" were used to describe the measures.

The term "consumer satisfaction" in health care evaluation has been used to describe a extensive range of research based on both broad and narrow definitions. Studies based on a broad definition use measures of self-perception of improvement, records of grievances, complaints, and praise for treatment, questionnaires assessing suggestions for improving treatment, and inquires into what is found to be either helpful or harmful. Studies based on a narrow definition are "...all inquiries into both the felt adequacy of treatment and of surrounding milieu: specific aspects may include reactions to the quality of care, to its helpfulness, its cost and continuity, the availability and accessibility of the practitioner, and the reaction to supporting services" (Lebow, 1983a).

Satisfaction research has been criticized as having little theory to guide the variables chosen for study or the hypothesis being tested, if any. Ambiguity has led to serious theoretical and methodological reservations about the validity of consumer satisfaction studies dating back almost a quarter century (Lebow, 1974). There appears to be an underlying belief in the literature that "consumer satisfaction" exists in the population simply awaiting measurement. Davis and Horobin (maintain that the picture of the consumer presented in survey research is "...frozen, temporally fixed and devoid of any sense of cognitive emergence." Locker and Dunt (1978) in an early review of the literature stated that

What do people means when they say they are satisfied with a particular aspect of medical treatment? One must know why people believe what they do and how they arrived at that view. There continue to be great limitations to understanding the process whereby consumers arrive at assessments of the services they receive. While expressions of satisfaction and dissatisfaction are the end product of such a process, little is known of the procedures, criteria and standards consumers employ.

In a thoughtful critique, Williams (1994) interrogated the a priori assumptions about the nature and meaning of expressions of satisfaction. He argued that in order for satisfaction assessments to be useful for system reform or quality control, survey results must be interpreted,

Value-Expectancy Models and Alternatives

The first assumption is that expressions of satisfaction are of conditional utility: they are important because they are assumed to be a function of some prior process-the meeting of client values and expectations. Therefore, it is believed that a substantial link between satisfaction and the fulfillment of expectations exists. The most well-known value-expectancy model was proposed by Linder-Pelz (1982a) and was based on the job satisfaction research by Fishbein and Azjen (1975). Linder-Pelz characterizes patient satisfaction as a positive attitude that is related to both his/her beliefs that the care possess certain attributes, and his/her evaluations of those attributes. Attributes are seen as separate dimensions of health care such as access, cost, convenience, and appropriateness. Three alternatives to the value-expectancy models have been identified (Lawler, 1973): discrepancy theory, fulfillment theory and equity theory. Discrepancy theory posits that satisfaction is the result of the perceived difference between what an individual desires and what the individual experiences. "Desires" is variously used as meaning expectations, what is important, or what should be. Most satisfaction studies are implicitly guided by a discrepancy approach (Greene et al., 1990; Fox et al., 1981; Zastowney et al., 1983; Pope, 1978). Fulfillment theory suggests that satisfaction is the difference between desired rewards and those received (Noyes et al., 1974; Larson & Rootman, 1976) or the simple difference between what occurred and what was expected. Equity theories propose that satisfaction is the perceived balance between inputs and outputs.

Although widely accepted, these models have failed to substantially explain satisfaction when tested, and there is little evidence that satisfaction is largely the result of fulfilled expectations and values. The inability to explain more than 20% of the variation in levels of satisfaction from a factor analysis of patient perception of quality of nursing care by Thompson (1984), led him to conclude that "other important considerations form the set of unmeasured attributes must provide the bulk of explanation." Linder-Perlz tested the hypotheses of Fishbein and Ajzen and the discrepancy and fulfillment models (Linder-Perlz, 1982, 1982a). Analysis revealed that no support could be found for the Fishbein and Ajzen theory. Further, satisfaction was unrelated to fulfillment but inversely correlated with discrepancy. In other words, the effect of expectation was shown to be significant, but it still could only explain 8% of the variance in satisfaction. Values and perceived occurrences were also found to have independent effects, but taken together with expectations, it only amounted to 10% of the variance. Therefore one may conclude that while expectations and values are involved in evaluation, they are not related to satisfaction in a simplistic, straightforward manner. In fact, it is reasonable to believe that expectations are modified through the process of receiving care.

Satisfaction could imply that certain attributes of an aspect of care have been approved or affirmed service recipients. However, the Linder-Pelz study raised the possibility that evaluations may be independent of the attributes of the aspect of care being evaluated, in that expectations were found to have an independent effect on satisfaction irrespective of their fulfillment. Linder-Pelz (1982) suggests that

Therefore, since evaluations of care could have little to do with the individual episode of care itself, the utility of expressions of satisfaction are undermined as currently applied to evaluation of quality of care. In a review of satisfaction with social services, Shaw (1984) points out that

Vuori (1972) found that hospital patients took technical competence for granted and detected defects in the behavioral aspects of care and communication, while the patients in general practice took empathy for granted but were critical of technical quality. This means that the "taking for granted" is an important factor in understanding satisfaction. It implies either the nonexistence of expectations or a more passive role adopted by service recipients than suggested by the concept of satisfaction. When and under what conditions do people take the quality of certain aspects of service for granted? The assumption that satisfaction is solely the product of evaluation may interfere with the accurate interpretation of satisfaction survey results.

The role that consumers assume regarding their relationship to the health care system has significant meaning for understanding satisfaction. For example, the Parsonian sick role is based on the passive nature of the patient. Following this logic, a person's expectations may be that s/he remain passive and defer to the paternalistic role of professionals. In fact, passive patients are often regarded as "good" patients." Therefore, satisfaction could be a reflection of the role the recipient of care adopts in relation to health care professionals, rather than a measure of the quality of care. It is also possible that the existence of values and expectations is context-driven, and there could be situations where expectations might not even exist. Consider the mind-set of a person who is encountering a service delivery system for the first time; s/he expectations based on past experience have not been formed. The technical nature of many treatments may also deter service recipients from believing in the legitimacy of their own expectations. Even though medical technology is an important determinant of clinical outcomes, studies have found that a person's evaluation of the technical aspects of health care explain little of the variance in global satisfaction (Cleary & McNeil, 1988). Rather than satisfaction being the result of evaluation, it may well be a product of the confidence of the health care recipient in medical procedures. Another possible conclusion is that satisfaction is really an expression of the thankfulness of the person for being treated (Campbell, Ho & Evenson, 1994). Therefore, in some situations, expectations may exist but not in ways that support the assumed meaning of satisfaction.

There are also problems in measuring satisfaction because satisfaction instruments have been developed on the basis of an unclear concept of satisfaction or on a broad definition of satisfaction (Ruggeri, 1994). In addition to conceptual issues, there is a lack of standardization regarding the methods of measuring consumer satisfaction (Larsen, Attkisson, Hargreaves, & Nguyen, 1979). Most satisfaction instruments are usually simple ad hoc measures (Larsen et al., 1979; Pascoe, 1983; Ruggeri, 1994; Sorensen, Kantor, Margolis, Galano, 1979; Westbrook & Oliver, 1981) which lead to a lack of generalizable results from one study to another (Kalman, 1983). Further, the measures of consumer satisfaction are criticized as being relatively insensitive to dissatisfaction while being very sensitive to satisfaction (Attkisson & Zwick, 1982; Nguyen, Attkisson, & Stegner, 1983; Ruggeri, 1994). Accordingly, both the definition and measures of client satisfaction are still vague and not consistent throughout the literature.

While it is generally agreed that consumer satisfaction is a latent concept with multiple dimensions (Greenfield & Attkisson, 1989; Tanner, 1982; Ware, 1981; Ware, Davies-Avery, Stewart, 1978), the majority of studies in the mental health field have found a multidimensional structure. For instance, MacDonald, Sibbald, & Hoare (1988) developed a 42-item patient satisfaction questionnaire with a yes/no response format. They identified eight factors using factor analysis. These eight factors are fearfulness, isolation and apathy, lack of individualization, unsatisfactory surroundings, lack of autonomy, unsatisfactory personal hygiene facilities, lack of status and recognition, and restriction of action. However, in most factor analyses a large general factor accounts for much of the total variance (Lebow, 1983c). Some widely used satisfaction measures, such as the Client Satisfaction Questionnaire (CSQ), have a limit to their sensitivity and content validity due to their monodimensionality (Ruggeri, 1994).

Most of the studies investigating satisfaction have reported generally high satisfaction or strongly positive attitudes (Kalman, 1983; Tanner & Stacy, 1985). It is reported that the levels of reported satisfaction often range from 70 to 100% (Baker & Intagliata, 1982; Kalman, 1983). A review by Lebow (1983) listed the results of over 50 surveys-the average percentage of satisfied health care recipients was 77.5%. Respondents with high levels of satisfaction tend to be older, women, less well educated, or in poorer health than respondents with low levels of satisfaction (Ross, Steward, & Sinacore, 1995). These high levels of client satisfaction are thought to be related to the influence of respondents' acquiescence response biases. Acquiescence refers to the tendency of the person to answer questions affirmatively regardless of content (Gudjonsson, 1990). It is usually related to several factors, such as lack of real anonymity during surveys (Ross, et al., 1995). This biased result can not be avoided if confidentiality is not carefully pursued during the survey. Further, Ware (1978) noted that much of the variance in the satisfaction study could be explained by acquiescence to the wording of items. The role of confounding factors in satisfaction research requires additional study. Gudjonsson (1990) also indicated that persons with mental illness seem particularly prone to acquiesce and special precautions need to be taken when they are being interviewed.

The high levels of client satisfaction are also thought to be related to selection bias (Gillig, Grubb, Kruger, Johnson, Hillard, Tucker, 1990; Larsen et al., 1979; Ruggeri, 1994). Overall, patients who agreed to participate in the study were satisfied with the treatment received, while satisfaction among clients who had dropped out of the treatment programs has been overlooked. It is also believed that dissatisfied patients tend not to return questionnaires (Kalman, 1983). Involuntary patients often communicate their dissatisfaction passively by refusing to cooperate, and consumer satisfaction studies which do not report the frequency of this type of participation refusal are difficult to interpret (Gillig et al., 1990).

Measures of satisfaction can be collected in a wide variety of settings, are easy to collect, inexpensive, and understandable to the lay population. In instruments ranging from homemade tests that have been thrown together to lengthy, psychometrically sophisticated devices, satisfaction is often used as a primary indicator of service outcomes (Lunnen & Ogles, 1997). It has been on the basis of a presumed influence of satisfaction upon behavior that much of the research on satisfaction with customers, in general, and health consumers, in particular, that has been promoted. The assumption has been that improvement in satisfaction will be reflected through treatment compliance, higher quality of services, and better outcomes. Conversely, lack of satisfaction can adversely affect the therapeutic alliance or become a lightening rod for system reform. Therefore, changes in satisfaction could be reflected through increased effectiveness of services and improved profit-and-loss statements. Early satisfaction hypotheses stated that satisfaction on the job was related directly to productivity-that the happy worker was the productive worker. It soon became apparent that such a simple formulation was inadequate, and no simple relationship between satisfaction and productivity has been established (Brayfield & Crockett, 1955; Herzberg et al., 1957; Kahn, 1960; Opsahl & Dunnette, 1966; Smith & Cranny, 1968).

Besides simply monitoring overall norms for service, in some cases satisfaction measures have been included in multi-evaluation studies (Marquis et al., 1982), thus providing the opportunity to study satisfaction in relation to a matrix of process and outcome variables. Such studies are important because they can address substantive questions and can test the convergent and discriminant validity of satisfaction measures (Lebow, 1983). Perhaps of even more importance are satisfaction studies in the health care field that have been conducted to look at the relationship of satisfaction to compliance or adherance to treatment plans (Korsch & Morris, 1969; Aday & Anderson, 1974; Kincey, Bradshaw & Ley, 1975; DiMatteo et al., 1980); symptoms or functioning (Attkisson & Zwick, 1982; Lambert and Lunnen et al., 1994; Pekarik and Wolff, 1996); utilization (Roghmann, Hengst & Zastowsky, 1979, Attkisson & Zwick, 1982); accessibility of services (Doyle & Ware, 1977; Penchansky & Thomas, 1981, Dolinsky & Caputo, 1990); disenrollment from an HMO (Sorenson & Wersinger, 1981); fulfillment of service recipients' social needs; change in provider (Marquis, Davies, & Ware, 1982; and, structure of health care organization (Fleming, 1981; Greenley & Schoenherr, 1981; Ross, Weaton & Duff, 1981, Lochman, 1983).

Aday and Anderson (1974) have suggested that perceptions of satisfaction are the result of individual patient characteristics and of the medical care systems they enter. Often service recipient characteristics have been studied such as age, race, or education (Hulka et al., 1975; Dolinsky & Caputo, 1990). Professional competence and personal communication styles of physicians have been addressed as well (Hulka et al., 1975, Doyle & Ware, 1977; DiMatteo & Hayes, 1980; Lochman, 1983; Dolinsky & Caputo, 1990). Surveys of communities and of large health consumer groups have consistently found that consumer satisfaction is substantially correlated with their perceptions of physicians' conduct and interaction with service recipients, as well as with the accessibility and availability of medical services (Hulka et al., 1975; Mangelsdorff, 1975; Ware & Snyder, 1975; Doyle & Ware, 1977; DiMatteo, Prince & Taranta, 1979; DiMatteo & Hays, 1980; Greenley & Schoenherr, 1981). In a review of factors related to satisfaction with medical care, Lochman (1983) found that the clearest relationship to satisfaction include the accessibility of medical care, the organizational structure of clinics, treatment length, perceived competence of physicians, affiliative behavior, clarity and retention of physicians' communication to patients, and physicians' control. Wriglesworth and Williams (1975) suggested that satisfaction depended to a large extent on the degree of confidence people have in their physician, and that the amount of information given to a person correlated strongly with their confidence in the staff. The authors concluded that confidence and information are inextricably bound up and that when people complained about service aspects of health care, they were really expressing a lack of confidence in their physician.

While research has not found a direct correlation between satisfaction and compliance (Korsch & Morris, 1969), investigators have discovered that consumers' reports of satisfaction were positively associated with adherence to treatment plans (Kincey, Bradshae & Ley, 1975; DiMatteo et al., 1980). The specific aspects of satisfaction identified in these studies pertained to the patient-physician interaction, in particular the extent to which requests for explanations and information were met by the physician. In addition, the perception of the physician as friendly, understanding of patient concerns, and possessing good communication skills was consistently associated with increased satisfaction and compliance. Freemon et al. (1971) found that the amount of nonmedical conversation was correlated with satisfaction and compliance, as was the positive effect in the interaction. Physicians who freely volunteered information were more likely to have satisfied patients than the physicians who only responded to expressed needs for information. Svarstad (1974) described dissatisfied patients as tuned out, mumbling to themselves out of context, completely silent, or simply responding in a yes-no manner. The influence of patient satisfaction with the amount of information and communication a person receives during a hospital stay has also been related to the willingness of the person to return to that hospital for future treatment (Houston & Pasanen, 1972). A review of research designed to enhance compliance concluded that patient satisfaction was one of ten factors that subsequently affected patient compliance (Becker & Maiman, 1980).

The above studies support a link between satisfaction and aspects of health care service delivery. Factors with complicated or no apparent relationship to satisfaction were mode of payment, clarity of patients' communication physicians, physicians' personality, patients' sociodemographic characteristics, and patients' health status (Lochman, 1983). Associations between satisfaction with health care and social categories such as age or race have been found but have not been particularly consistent (Locker & Dunt, 1978). Research has also failed to establish a significant relationship between satisfaction and clinical health outcomes. A wealth of clinical research indicates that the correlations between satisfaction data and outcome variables are very low (Lebow, 1983; Lunnen & Ogles, 1997).

In mental health research, Pekarik and Wolff (1996) used clinical significance to define and separate those who were therapeutic successes from those who had failed to improve. They found little in the way of significant relationships between outcome and satisfaction. Because they used strict criteria, they concluded that this study

Lambert and Lunnen et al. (1994) conducted a similar analysis of service recipient satisfaction, adding an evaluation from the perspective of clients' spouses or significant others. They found that satisfaction levels for the "improvers" were not significantly higher than "nonchangers," but were significantly higher than the "deteriorators." From the spouse/significant other perspective, there were no differences in the level of satisfaction across the three mental health consumer groups. Both of these recent studies suggest a continued lack of concordance between symptom change and satisfaction for mental health consumers.

The author of an early review of satisfaction and medical outcomes suggested that recipients of care are more concerned or dissatisfied with the manner and means of the processes of health care delivery than with the outcome of care or competencies of health care personnel (Kelman, 1976). Both Koos (1954, 1955) and Cobb (1958) indicated that patient dissatisfaction with the consumer-provider interaction contributes to the use of marginal practioners, and Kasteler et al. (1976) identified patient dissatisfaction as a major contributer to doctor shopping. Ben-Sira (1976, 1980) stressed the importance of the experience of observing and interpreting the doctor's performance and the influence of perceptions at this stage upon subsequent judgements of satisfaction. Ben-Sira argues in his social interaction model that because a patient enters the medical interaction with anxiety about the medical prognosis, less knowledge about medical techniques than the physician, and less ability to connect treatments with improvements in health, the affective component of the physician's communication becomes a major factor in the patient's evaluations. The affective component consists of "behavior directed by the physician toward the patient as a person rather than as a case" (1980, p. 173). It is operationalized as the allocation of sufficient time to the interaction, show of interest in the patient, and demonstration of devotion to the medical problem. Tagliacozzo and Mauksch (1972) also concluded that most patients feel inadequate to judge the competency of efficiency of technological services and, consequently, develop an intense emphasis on providers' personality or "caring functions." Among the caring functions of providers identified by recipients as the most important aspect of treatment were personal contact, communication, and concern (Cartwright, 1964; Lewison, 1975; Roth, 1963; Skipper et al., 1964).

Ben-Sira's affective component is related to communication style or "the way one verbally or paraverbally interacts to signal how literal meaning should be taken, interpreted, filtered, or understood" (Norton, 1978, p. 99). A survey of the literature on consumer-provider communication (Buller & Buller, 1987) indicates two general styles displayed by physicians in medical visits:

(1) The Affiliation style establishes and maintains a positive relationship. This includes behaviors that communicate interest, friendliness, empathy, warmth, genuineness, candor, honesty, compassion, a desire to help, devotion, sympathy, authenticity, a nonjudgmental attitude, humor, and a social orientation (Ben-Sira, 1976, 1980; Collins, 1983; DiMatteo, Prince & Taranta, 1979; Freemon et al., 1971; Friedman, DiMatteo & Taranta, 1980; Jensen, 1981; King, LaGrone & Miller, 1984; Korsch et al., 1968; Korsch & Negrete, 1972; Street & Wiemann, 1987; Truax & Carkhuff, 1967; Vickery, 1983).

(2) The Control style establishes and maintains the physician's authority characterized by a difference in control, power, and status. The manifestation of power, authority, professional detachment, and status in medical interaction is widespread (Hall, Roter & Rand, 1981; Korsch & Negrete, 1972; Lane, 1983; Lavin, 1983; Milmoe et al., 1967; Street & Wiemann, 1987).

In general, research examining the link between communication style and satisfaction has found that physicians who adopt a more affiliative manner receive more favorable evaluations. Specifically, Korsch et al. (1968) found that physicians who devoted time to listening to the mothers and answering questions in pediatric visits had more satisfied parents. Korsch and Negrete (1972) reported that physician expressions of warmth were positively related to patient satisfaction; Freemon et al. (1971) showed that consumer satisfaction was associated with warm and friendly behavior by the physician. Jensen (1981) also found that caring physicians were more likely to satisfy patients; Pantell et al. (1982) reported that when doctors spent more time being friendly and discussing nonmedical subjects, the evaluations were more favorable. Finally, Street & Weimann (1987) found that involved, expressive doctors had more satisfied patients.

In contrast, doctors who are highly controlling receive less favorable evaluations. Information control through the failure to provide clear explanations about diagnosis due to the use of jargon lowered satisfaction in the Korsch et al. study (1968). Similarly, Korsch and Negrete (1972) established that people were less satisfied when the doctor's communication tended to support a controlling relationship. Negative relationship between authoritative style and satisfaction was reported by Hall et al. (1982), Lane (1983), Milmoe et al. (1967), and Street and Wiemann (1987). In a study by Bueller and Bueller (1987), satisfaction with the doctors' communication determined 70.6 percent of the variance in patient satisfaction with medical care.

For a variety of reasons effective communication does not commonly take place between health consumers and providers. The social world is intersubjective and this common accessibility is shared through language. Through communication people are able to negotiate their interests or goals. The success of, or problems with, negotiations are dependent on whether people share sociolinguistic codes and knowledge. According to Mathews (1983), problems in the consumer-provider relationship lie partially in the fact that they do not equally share medical knowledge and language, thus their goals or interests may not coincide. Freidson (1970) attributed physicians' resistance to communicate information to their view of the lay person as incompetent to comprehend or emotionally cope with medical information. Physicians also feared that informed consumers threaten professional status and control over the therapeutic situation. Some observed that medical professionals do not view people receiving health care as adults, but rather treat them as children (Freidson, 1970; Szasz & Hollender, 1956). Consequently, the exercise of professional authority by providers is thought to reinforce reluctance on the part of persons receiving care to question providers. This often serves as a justification for providers to withhold information. Further, doctors do not often conceive of the consumer-provider relationship as an integral, important part of their role, and there is a tendency to regard psychosocial needs as incidental and peripheral to medical concerns. By withholding information, physicians can avoid scenes (Davis, 1972), protect against sadness (Quint, 1972), reduce the likelihood of mismanagement or mistakes being uncovered (Skipper et al., 1964), and assume greater patient control (Roth, 1963). Provider control of information contributes to patients' misunderstanding of their role in the relationship, encouraging passivity (Cartwright, 1964). When faced with questions from patients, providers may either accept patients on equal terms and establish a partnership relationships in which negotiation of care occurs, or physicians can discourage interaction and establish conflictual relationships. The dynamics of a conflict relationship are illustrated in Roth's study (1963) in a tuberculosis hospital in which the attempts of patients to get information were countered by opposing provider strategies to safeguard information. As a result

In addition, patients who do not receive adequate communication from physicians, may be reluctant to provide information or omit information selectively in an effort to control the physician's assessment of their condition (Skipper et al., 1963).

Speeding and Rose (1985) argued that assessing satisfaction is an insufficient way to monitor the quality of the consumer-provider relationship because satisfaction research neglects to incorporate measures of consumer participation.

Passive acceptance of physician instructions is not a sign of commitment to them. Taylor (1979) and Raps et al. (1982) pointed out that people who appear to understand and accept medical instructions may merely be adapting to a social situation which they feel unable to control and unwilling to resist. Many "good patients" may actually feel helpless because they cannot personally influence the situation (Lorber, 1975; Seligman, 1975). They acquiesce only passively to the treatment plan, complying ritualistically. Hingson et al. (1981) explained that the lack of a positive relationship between understanding a physician's instructions and compliance is, in fact, due to a lack of commitment to the treatment plan. Simson (1974) showed that people evaluate physician advise against their actual experience (e.g., experience with side effects and perceptions of efficacy), the experience of others in their social network, and cultural definitions of illness and its treatments and then act accordingly. He wrote:

There is evidence that people who are oriented toward active involvement in their health care and who participate with their physician in decision-making about managing illness are more prepared to enact health promoting behaviors. When Seeman and Seeman (1983) studied the health habits and illness experiences of a probability sample of people in Los Angles County subjects were rated according to the level of their sense of personal control over their health. They found that

Encouraging people to take a more active role in managing their health problems could be the best vehicle for transforming satisfying consumer-provider relationships into effective ones if doctors elicit from patients their assessments and feelings about the proposed course of treatment (Speedling & Rose, 1985). When a health consumer and provider collaborates in producing a mutually acceptable treatment plan, both are more likely to agree on the priorities which need to be addressed and the tasks each needs to perform (Hulka et al., 1976; Starfield et al., 1981). Kassirer (1983) suggested that

Since the vulnerability of patienthood tends to foster complacency and dependency rather than autonomy, it is critical that the consumer-provider relationship support a person's sense of control over their health. In this context, the consumer-provider relationship that fosters satisfaction without encouraging participation would be seriously limited and only weakly associated with treatment outcomes. Lebow suggested that the conceptual link between satisfaction and efficacy should be even stronger in mental health than physical health because the therapeutic relationship is a key factor.

A topic of longstanding interest in the counseling literature, the quality of the therapeutic relationship has consistently been acknowledged as an essential component of mental health treatment (Gelso & Carter, 1985; Greenberg & Pinsof, 1986; Gurman, 1977; Rogers, 1957), and the psychodynamic concept of the therapeutic alliance has been utilized in studying the quality of consumer-provider relationships (Bordin, 1979; Greenberg & Pinsof, 1986; Marziali, 1984). In fact, the therapeutic alliance has been proposed as the common factor that accounts for therapy outcomes regardless of treatment approach (Hartly & Strupp, 1983; Horvath & Luborsky, 1993; Styles, Shapiro, & Elliot, 1986). The positive role of the alliance has received growing empirical support (Alexander & Luborsky, 1986; Horowitz et al., 1984; Horvath & Symonds, 1991).

In a paper first published in 1912, Freud remarked on the need for the analyst to maintain a stance of "serious interest" and "sympathetic understanding" toward the client so the healthy part of the person's ego could form a positive attachment to, and develop a collaborative relationship with the analyst. Rogers (1957) postulated that the therapist's willingness to be empathic, to be congruent, and to accept the client unconditionally were not only essential but sufficient conditions for therapeutic success. Rogers, however, virtually ignored the client's role in the relationship, presuming a fated response to the benevolent attitude of the therapist. Subsequent research reviews indicated that it is the client's perception of the therapist's attitudes that exhibits a moderate but consistent relationship positive outcome of therapy rather than the therapist's intent or behavior toward the person (Mitchell, Bozarth, & Krauft, 1997). According to Frieswyk et al. (1986), therapist activities should not be included as a component of the alliance. The consumer's collaboration or "the extent to which the patient makes active use of the treatment as a resource for constructive change" (p. 36) is the "essence of the alliance" (Colson et al., 1988, p. 264). Bordin (1979, 1980) held that a strong working alliance requires a high degree of cooperation and mutuality with respect to the goals, tasks, or bonds of therapy.

Initial assessments of the therapeutic alliance focused on observers' evaluations, but increased acknowledgement of the importance of understanding the client's view of the relationship (Barrett-Lennard, 1962; Marmar et al., 1986; Rice & Greenberg, 1984) led to the development of parallel client and therapist versions of measures (Horvath & Greenberg, 1989). However, perceptions of therapy events or constructs developed by researchers to define the process of therapy were found to be different than those clients spontaneously generated (Elliot, 1989; Patton & Jackson, 1991). Further, consumer perceptions of the relationship have proved to be more consistent predictors of improvement as compared with other evaluative perspectives (Gurman, 1977; Horwitz et al., 1984; Marziali, 1984). In a phenomenological analysis of 66 descriptive accounts of the therapeutic alliance from the perspective of the involved client, Bachelor (1995) that most alliance variables are not equally relevant for clients and that some crucial features of the perceived working relationship are not accounted for in current theory.

While there is a wealth of studies to support the importance of the therapeutic alliance, this has not necessarily been translated into building positive consumer-provider relationships in mental health service delivery. This is because on-going tensions between providers and consumers partly emanates from the perceptual frameworks which people involved in the mental health system use to define their roles, behaviors, and socio-emotive responses. Persons who experience mental illness and the people who try to help them see the world from very different perspectives (Ridgway, 1988). From the viewpoint of mental health professionals, a diagnosis indicates pathology. However, psychiatric pathology is an abstract concept that is separate from the person and his/her suffering. The professional monitors functionality, recidivism, and symptom control, but people judge their services by the impact on the quality of their lives (Campbell & Schraiber, 1989; Trochim, Dumont, & Campbell, 1993). To the individual mental health consumer, a diagnosis becomes an assault on identity and sense of self. While the mental health professional assumes authority through the delivery of services, for the consumer there is concomitant erosion of autonomy and control. Therefore, rather than a shared reality, mental illness represents two quite different realities, the meaning of one being significantly different from the other. The incongruities between the perceptual and experiential framework of those that research, manage and deliver services and those that receive services has emboldened consumers and advocates to question the appropriateness of the opinions and studies of medical professionals when contrasted with the insights and preferences of consumers and their families (Campbell, 1993).

Consumers often voice degradation, anger, and sadness when they speak about their experiences as psychiatric patients. These experiences stand as a critique of biomedical psychiatry and the mental health services delivery system measures of satisfaction (Zinman, Harp, & Budd, 1987; Chamberlin, 1978; Nide, 1990; Colom, 1981; Susko, 1991). In opposition to findings of high ratings of global satisfaction, there appears to be wide differences between consumers and professionals on the relative importance of treatment goals (Dimsdale et al., 1979), identification of problems (Mitchell et al., 1983), barriers to service to service (Lynch & Kruzich, 1986), and needs and preferences for housing and supports (Minsky et al., 1994; Ridgway & Carling, 1987; Tanzman, 1993). Mental health consumers argue that from a traditional point of view expressing anger, refusing to take medication, or maintaining a distance from ones family, is generally considered to place the person at risk of decompensation, isolation, substance abuse, or homelessness. From the point of view of the person expressing these behaviors, these same outcomes could be viewed as leading to growth, recovery, and independence (Scott, 1993, p. 4). Controlling symptoms and behavior of consumers and maintaining the effectiveness of the service system are expressed priorities of mental health professionals, while consumers ask, "What does this do for my life?" (Scott, 1993, p. 4). These differences in perspective are shared through forums such as Pioneer Dialogues (Loder & Glover, 1992) and focus groups among mental health consumer professionals and researchers (Campbell, 1997; Campbell, 1996; The Consumer/Survivor Mental Health Research and Policy Work Group, 1992).

Recognizing these differences in perspective, one needs to ask how and to what extent psychiatric discourse is transmitted through a language that reinforces professional domination. According to C. Wright Mills (1959), an individual's difficulties are almost always interconnected with structures in society, although these relationships remains hidden in the presentation of everyday life. Therefore, an important goal for the analysis of social problems is to clarify how personal troubles and social issues relate to one another. In the intimacy of the psychiatric encounter, consumers present personal issues that go beyond the individual level. In other words, interpersonal processes occur in a social context shaped by macro-level structures in society. Ideology as a macro-level structure in society impinges on consumers and providers as part of the social context of psychiatric encounters. At the micro-level of interpersonal interaction, elements of ideology appear in the consumer-provider communication. What consumers and providers say when they meet reinforces their particular ideologic conceptions about social life. Structures of society help generate the specific social context in which consumers and providers find themselves; the talk that occurs in psychiatric encounters also reinforces broader social structures. In psychiatric encounters, technical statements direct the responses of consumers to objectified symptoms, signs, and treatments. This reification shifts attention away from the totality of social relations and the social issues that are often root causes of psychiatric problems. Instead, attention is focused on individual pathology and personality. Therefore, psychiatry can reduce the potentiality of positive consumer-provider relationships when symptoms, signs, and treatment take on an aura of scientific fact, rather than being equivocal manifestations of subjectivity in a troubled social reality. Inevitably the tendency of professionals in the psychiatric encounter is to lead consumers to adjust to things as they are by encouraging coping and accommodation on the one hand, or coercing them on the other.

An important step in looking at the utility of satisfaction should be to determine if satisfaction as a concept has meaning for consumers themselves. It is clear that people may not act or evaluate services in terms of "being satisfied." Therefore, satisfaction may not be grounded in terms of people's everyday lives. In interviews with medical patients, Fritzpatrick and Hopkins (1983) "were struck by a lack of fit, in many respects, between patients' own accounts of their experiences and the assumptions about patients contained in the satisfaction research" (p. 297). When qualitative methodologies are utilized, little support is found that people think and evaluate in terms of a continuum of satisfaction. Further, Calnan (1988), and Locker and Dunt (1978) found that people are more critical of services when they are interviewed or respond to open-ended questions. This finding has been borne out in mental health by the seemingly opposing findings of quantitative satisfaction studies in which satisfaction is found to be generally very high, and the qualitative literature of patient narratives, first person accounts, newsletters, and the like (Campbell, 1998). The reductionism necessitated by quantitative methods of determining consumer satisfaction may render satisfaction in both physical and mental health care "devoid of much of the meaning they were intended to embody" (Williams, 1994, p. 514). In fact, satisfaction could be used to deny negative voices of health consumers regarding their experiences when receiving health care. Whitfield and Baker (1992) suggest that poor questionnaires are actually a form of censorship.

Although flawed as an assessment tool, the improvement of satisfaction potentially has humanitarian value--a legitimate goal which has important implications for mental health service evaluation. The study of consumer satisfaction may be necessary to establish a general theory that will serve as a basis for practical action and for future research. The desirability of establishing a general theory of satisfaction is evidenced by the wide variety of publications in this area, and reasons why this research is recognized as important are easy to find. Understanding the sources of satisfaction and dissatisfaction may also be important in themselves, whether measuring satisfaction with treatment delivery or the people who provide services, if such information has a meaningful impact on the course of someone's recovery and subsequent experiences in everyday life. In other words, the use of satisfaction information as a quality improvement tool may be of greater value than the concept of satisfaction or quantitative results of assessments.

While some contend that the health care field must first gain a clearer understanding of how consumers evaluate before their opinions can be accurately collated and interpreted (Williams, 1994), Ruggeri (1994) advocates a more pro-consumer position and supports the use of satisfaction even though its meaning has not been clarified.

If the concept of satisfaction is ambiguous and its uses problematic, why has satisfaction as a measurement of quality proliferated? Over a decade ago Wennberg (1988) suggested that medicine was beginning to experience a scientific revolution as professional uncertainty rather than consensus about the scientific basis of clinical practice was emerging as the dominating reality. The rapid growth of managed health care organizations and the displacement of public sector health through privatization and the proliferation of state Medicaid waivers has created cause for concern about the quality of patient care. The focus on controlling access in order to control costs has left the impression that treatment may be limited as consumer choice is restricted (Croze, 1994). Today challenges to the long-held assumptions of professionals who manage the health care of the nation have reached a crescendo as health care spending surpasses $1 trillion in annual costs. Payers point to physicians that pour health dollars into unneeded tests and unproductive treatments.

One approach to these problems is to keep doctors appraised of what does and does not work, monitor the responses health care recipients, and to organize the health delivery systems to reflect such knowledge (Mehlman, 1986). Successful industrial quality control methods such as Total Quality Management (TQM) and Continuous Quality Improvement (CQI) (Deming, 1982; Juran, 1988) have created the hope that similar strategies can be adapted to health care management and quality control. In particular, person-based measures that link treatment and service variables with subsequent status and service satisfaction have been embraced by both public and private mental health administrators as the means to manage quality, cost, access, appropriateness of services. Marquees et al. (1994) observed that "[P]urchasers, consumers and managed behavioral healthcare organizations have come together in their realization of the need to move beyond simple measures of utilization and cost containment when assessing the quality and effectiveness of care" (p. 28).

It is apparent that satisfaction assessments are part of the movement towards accountability and has been seen as the necessary response to the role of consumers in the marketplace. Eddy (1989) argued that proper clinical decisions should be based on measures that are important to those that receive services. Wennberg (1988) observed that in many clinical situations, the physicians' treatment and outcomes preferences unduly influence the clinical decision process, in part because so little attention is now given to the crucial task of helping people discover and choose the care they really want. Therefore, the current problems of health care as expressed in the vast literature questioning the consistency and quality of care have stimulated both the reemergence of the patient (Reiser, 1993, p. 1014), and the rise of the concept of satisfaction.

Several studies have emphasized the need for consumer involvement in the process of program evaluation to improve quality of health care (Davis & Ware, 1988; Donabedian, 1988; Glass, 1995; Gustafson, 1991; Lehr & Strosberg; 1991; Margolis, Sorensen, & Galano, 1977; Midgley, Gilliland, Rose, Livermore, Beckstein, LeBlanc, Lemoine, & Speier, 1994; Nelson & Niederberger, 1990). Nelson & Niederberger (1990) suggested that without consumer consultation or involvement in the design of satisfaction surveys, these instruments may not provide an adequate representation of consumer satisfaction. Midgley et al. (1994) indicated that important treatment outcomes such as satisfaction are associated with consumers' "voice" and perceived fairness. However, most satisfaction surveys still represent the providers' definition of quality and whether patients have a different definition is unknown (Glass, 1995). Considering the conceptual problems previously discussed in efforts to develop a general theory of satisfaction that would inform both the providers of services and the recipients, it is critical to involve the consumers in the development of the measurement tools. Still, the role of consumers-with notable exceptions-- has been minimal in the field of health care evaluation.

One reason is that information that encompasses personal and social needs is considered soft or subjective science, and, therefore, the expressed opinions of people as part of a system of evaluation has been resisted in the medical field where objective data has led the development of knowledge and practice. On the other hand, objective data ignores the meaningful human aspects of illness and treatment and all the factors that differentiate people from symptoms, brains, or molecules. Another reason is that medical professionals assume that the principles of good health care override individual treatment preferences. Historically, changes in and improvements to health services have been based on the common assumption that the expert knows best about the operating principles of the health service delivery system. Guided by the zeitgeist of paternalism, many health professionals believe that not only do they know the scientifically correct way to deliver services, but that they know the values and preferences of their clients, are disinterested and objective without preferences and biases of their own and choose for their clients what the client would choose for themselves if they had the same knowledge (Beisecker & Beisecker, 1993; Silberman, 1993; Wennberg, 1988). Eddy and Billings (1988) have cautioned that this seduction by authority inevitably influences future behavior to mimic current practice and has lead to the creation of current practice standards without knowledge of how people would compare the benefit and harm (p. 15).

The transformation of the healthcare delivery system to a managed care environment may provide mental health consumers with the opportunity to take responsibility for their own lives in the medical marketplace and to be truly seen as customers. Health consumerism is based on the assumption that persons who seek health services are consumers just as are persons who seek other types of services. The doctor is perceived to be the purveyor of a service, and the patient is viewed as the buyer (Haug & Lavin, 1983). The consumer listens to the thoughts of the provider, but ultimately makes his or her own decisions (Reeder, 1972). Consumerism implies that values derived from principles of good medical care must be interpreted and operationalized through reference to patients' personal health care values and desires (Beisecker & Beisecker, 1993, p. 53).

Although consumerism has periodically emerged as a force in American society, its current application in health care draws its roots from two marketplace trends: consumer rights protections with its concerns regarding manufacturing and product safety, and Total Quality Management (TQM) with its focus on customer satisfaction. The former is grounded in a profound distrust by individuals of the actions and motives of providers of products and services, the latter promotes collaboration between provider and consumer and seeks to answer such questions as what do customers prefer. Both trends emphasize the need for information that accommodates consumer rights and interests.

Support for patient autonomy and rights can be found in the passage of a consumer bill of rights introduced by the Kennedy Administration in the 1960s. In response to concerns that individuals were being victimized by corporations marketing unsafe products and individuals providing unreasonably priced services (Beisecker & Beisecker, 1993), the legislation included the rights of consumers to safety, to be informed, to choose, and to be heard (Cornacchia & Barrett, 1980). With health care characterized as an industry which strongly protects the profits but only weakly protects the healing and rights of people (Cornacchia & Barrett, 1980, p. 2), it could not escape the application of such rights to health consumers. As a result, it is no longer acceptable that the doctor actively control information about patients' conditions; and service recipients must be informed of their condition and, in most cases, give consent for treatment.

People are assuming a greater responsibility for their own health care and education without consulting professionals (Harris & Hamburg, 1990). Acceptance of medical consumerism is evidenced by consumer-oriented health books that are available (Boston Women's Health Book Collective, 1973; Chisari, Nakamura, & Thorup, 1976; Levin, 1975; Sehnert & Eisenberg, 1975; Vickery & Fries, 1976), the publication of textbooks on health consumerism (Cornacchia & Barrett, 1980; Hamilton, 1982; Kime, 1970), and the organization of patient advocacy groups (Demkovich, 1987). The Picker-Commonwealth survey found that Americans are eager to use survey data to help them decide where to seek medical and hospital care and to play an active role in decisions about health care delivery in general, and in their own treatment choices (Cleary et al., 1991). Further, market demands for health information has spurred such technological developments as electronic bulletin boards for persons with medical questions, educational software on medical topics, and home medical tests.

Many of the original mental health consumer/survivor self-help groups began over twenty years ago in response to perceived coercion, devaluation and the scarcity of appropriate services in the mental health system and the community (Campbell & Schraiber, 1989). Lack of understanding, respect, and trust between consumers and professionals prompted people with psychiatric disabilities to seek the company of their peers for validation and support. Essential characteristics of the consumer/survivor self-help programs include the self-definition of needs, equal power of members, mutual respect, voluntary participation by members, autonomy, and responsivity to special populations (Chamberlin, 1978; Zinman et al., 1987). This approach has guided the development of innovative programs and services. The natural evolution of consumers establishing themselves as capable, caring, and dedicated partners providing services for people with psychiatric disabilities (Van Tosh, 1993), as well as helping themselves individually or within the company of peers, has enabled people with mental illness to exercise considerable control over their psychological and emotional problems and to enrich their lives. According to Breier and Strauss (1984), the reassurance that self-help groups provide people with mental illness reduces anxiety and restores hope, thereby reinforcing a sense of belonging in a social setting. Some of the early groups, such as the Network Against Psychiatric Assault and Project Release, created the first opportunities for individuals to gather by themselves to provide each other with support and to provide a forum for community organizing.

Historically, people with mental illness and their families have fought with limited but significant success to redress perceived injustices in the delivery of public mental health services by organizing into advocacy groups. The executive director of The National Alliance for the Mentally Ill (NAMI), a family advocacy organization, wrote, "We recognized early that users of the mental health system--that is, consumers and their families--are in a real sense the only customers. Our long history of advocacy to improve services has created new attitudes and new respect in the public sector (Flynn, 1994).

Mental health consumers have built a national self-help movement within the public sector with growing numbers now participating in research and evaluation (Campbell, Ralph & Glover, 1993), and taking leadership roles in policy and administration of public mental health services (McCabe & Unzicker, 1995). Consumers have led efforts to determine housing preferences (Tanzman, 1993), to define outcome measures (Campbell, 1997; Campbell & Schraiber, 1989; Dumont & Campbell, 1994; The Consumer/Survivor Research and Policy Work Group, 1992; Trochim, Dumont, & Campbell, 1993), and to develop partnerships models with public mental health professionals (Campbell, 1996). Some professionals and policymakers have responded to consumer demands by redesigning professional roles and creating opportunities for people who receive services to provide input and perspective (Blanch et al., 1993). The proliferation of Offices of Consumer Affairs in state mental health agencies, state-wide consumer conferences, consumer-directed social centers across the nation, model consumer case management programs, and support for consumer research and policy professionals are indicators of a vibrant mental health consumer movement (Bevilacqua, 1993; Buckley, 1993; Campbell, 1998; The Consumer/Survivor Mental Health Research and Policy Work Group, 1992; Scott, 1993).

In 1989 the National Association of State Mental Health Program Directors (NASMHPD) approved a position paper that recognizes that former mental patients/mental health consumers have a unique contribution to make to the improvement of the quality of mental health services in many areas of the service delivery system. The paper goes on to recommend that consumer contribution should be valued and sought in areas of program development, policy formation, program evaluation, quality assurance, systems designs, education of mental health service providers, and the provision of direct services.

Similarly, the current administration has attempted to re-invent behavioral health policy in America by shifting away from exclusion towards inclusion; away from dependence towards independence; away from paternalism towards empowerment. The Substance Abuse and Mental Health Services Administration (SAMHSA) Strategic Plan (March, 1993) sees this as the way to achieve a better, healthier America for all its citizens. The document states that mental illnesses may be most effectively reduced when consumers of prevention and treatment services, their families, and their neighbors share in making decisions about services they receive and the communities in which they live.

As the necessary role of consumers in evaluation and quality management are acknowledged, mental health information systems are beginning to grapple with the person-centered values of service recipients. Chowanec et al. (1994) suggests that within the field of healthcare--and particularly mental healthcare--perhaps the most far-reaching revolution will be the reorientation and re-conceptualization of the role of the patient within the healthcare delivery system (p. 47). Therefore, a critical stage in the development of quality management systems must not be overlooked, that is, that a thorough examination of consumer values and priorities should be undertaken prior to the selection of measures (McGuirk et al., 1994, p. 29).

On the one hand, the perspective of consumers can serve as a force for health care reform and quality improvement. On the other hand, consumers must overcome tremendous personal and political barriers to gain a place at the table. Since, the work of consumer researchers over the past ten years in evaluation and outcome assessment has seldom conformed to the standards of experimental research design or established the psychometric properties of their instruments, little has been published in peer reviewed journals. The results of most consumer studies, both qualitative and quantitative, are found in unpublished reports, proceedings, and newsletters (Sherman & Kaufmann, 1995). Consequently, the movement towards satisfaction and report cards has depended on a literature that fails to reflect the experience and insight of consumers.

The advent of managed care in behavioral health and the recognition that quality indicators need to reflect the perspectives of the customers, has created a willingness in the field to re-examine current assessment protocols. Support of consumer focus groups and public forums launched the articulation of a consumer/survivor core data set of indicators that could be included in outcomes management systems. Based on the input of persons with physical and mental disabilities, the Accreditation Council on Services for People with Disabilities (1993) developed value-based performance measures. Outcomes included personal goals, choice, social inclusion, relationships, rights, dignity and respect, health, environment, security, and satisfaction. The Center for Mental Health Services (CMHS) supported a series of focus group sessions of national consumer leaders to explore values and outcomes in mental healthcare. According to participants, traditional mental health systems pathologize problems in living, hold low expectations of consumer achievement, are paternalistic, offer a limited range of options, and define anger as symptomatic. Some key concerns included the threat of involuntary treatments, subtle forms of coercion, lack of respect towards consumers by mental health professionals and providers, and the debilitating side-effects from medications. It was observed that researchers fail to ask questions (e.g., about the presence of personal freedom and decision-making power) that would capture detrimental effects of treatment and care. Recovery, personhood, well-being, and liberty were identified as relevant outcomes that are seldom measured or operationalized in traditional mental health research or program evaluation (The Consumer/Survivor Mental Health Research and Policy Work Group, 1992).

In the past few years, the federal government has sustained the continued investigation of service priorities and quality indicators (Campbell & Johnson, 1995; Ganju, 1995) to help service recipients select providers and monitor the quality of their care. Two major consumer pilots to identify valued mental health outcomes were funded by CMHS to provide input into the MHSIP Consumer-Oriented Mental Health Report Card (MHSIP, 1994; Teague et al., 1997) and the Consumer Component of the State Mental Health Agency Profiling System (Campbell, 1998). In these pilots, consumers brainstormed statements describing specific consumer-defined individual and system indicators and measures that they wanted in a quality management protocol. Using concept-mapping technology, domains of concern were identified and ranked. These included damaging effects of the system, consumer impact on service delivery and system development, oppression due to race and gender, healing and recovery, coercion/control and autonomy, personhood, alternatives to traditional services, citizenship, quality of life, and employment (Dumont & Campbell, 1994; Trochim, Dumont & Campbell, 1993).

In behavioral healthcare, report cards are being devised to assess the benefits and impact of services for a broad range of system stakeholders. Among the national organizations engaged in producing reports cards for behavioral health care are the Center for Mental Health Services (CMHS) through the Mental Health Statistics Improvement Program (MHSIP), the National Alliance for the Mentally Ill (NAMI), The National Committee for Quality Assurance (NCQA), the American Managed Behavioral Healthcare Association (NCQA), the Institute for Behavioral Healthcare (IBH), and the Joint Commission on Accreditation of Healthcare Organizations (JCAHO).

Ultimately, the assessment of satisfaction is based on the refusal of people to accept paternalism. Paternalism abdicates the right to evaluate. Williams (1994) sees consumerism as relying on three fundamental factors: (1) the existence of consumer opinion; (2) the consumer's belief in the legitimacy of that opinion; and (3) willingness of engage in an expression of that opinion. Satisfaction data is only useful if people actively evaluate and criticize rather than be passive. Beyond that understanding, however, it is important to identify what it means for people to act as consumers in services such as health care. There are many models of consumerism ((Winkler, 1987) with different roles defined for the service recipients. In order to articulate a theory upon which to base an evaluation protocol, it is essential to discover what rights and obligations consumers believe they have. In general, people who use satisfaction assessments have not interrogated theory as part of the development of an evaluation practice. In fact, based on the methods of interpreting satisfaction data, there appears to be a built-in assumption that all recipients are acting in the same way. This belief is, in a sense, contrary to the proclaimed concern for the empowerment and rights of service recipients, since it denies people the power to behave as consumers.

Clearly, consumer perceptions about treatment efficacy and satisfaction with the type of treatment received, as well as the way it is delivered, determine which treatments are sought or complied with and which outcomes are valued (Mirin & Namerow, 1991). However, it appears that report cards, for example, serve different stakeholders with different values, goals and priorities. The customers of health care services have not been necessarily included or had empowered roles when they were developed or implemented. As a result, systems of quality management that are specifically focused on the identification, collection, and distribution of data have only partially addressed the information needs of consumers. In a analysis of the technical assistance needs of consumer/survivor and family stakeholder groups within state mental health agencies, Campbell (1998) found that state departments of mental health were not involving mental health consumers in meaningful roles in the design and implementation of survey instruments and services research projects except as advisory board members (50%) and input through focus groups (71.7%). In only 13% of states were consumers funded to participate in research activities, and less than 40% funded consumer initiatives to monitor satisfaction. A survey of public mental health provider agencies conducted by one state department of mental health found that consumers were involved in the development of the instruments only 25% of the time; few of the agencies included consumers in the loop of information retrieval and dissemination; and only 11% of the agencies were able to track special populations; the majority of agencies were not able to measure and/or analyze results related to gender, ethnicity, age or rural residence (Campbell, 1994).

We have to follow rigorous protocols for the tools we use to measure the variables we are to study. The goal is to study the laws relating situations, personal characteristics, and policies to satisfaction and to behavior. Satisfaction measures should have the following attributes:

The measures must be applicable to a wide variety of consumers in a variety of treatment modalities and in a variety of agencies. Specifically, the verbal level of the measures should be low enough that they can be given to almost any consumer. Wide applicability implies that the words used should be simple enough to be understood by almost anybody, and that those words should be relevant to almost any person's situation. That makes long, complicated attitude statements, with many qualifying words and phrases problematic. Also, attitude scale alternatives, such as "strongly agree," "agree," etc. are subject to interpretation of the level, extent, or strength of the qualification.

They should be inexpensive in terms of time and money. That means the tool should be short, and perhaps administered in groups, eliminating the need for individual interviews. A short pencil-and-paper checklist is one useful approach.

They should be standardized, so that they are comparable from person to person, both in administration and in interpretation. Standardization requires not only clear instructions and format, but also the compilation of data for the kinds of people for which comparison must be made. For example, norms for satisfaction for a particular treatment modality should enable one to compare different consumers in similar circumstances.

They should separate the various aspects of satisfaction, so that feelings about respect of staff, for example, are not lumped with feelings about the appropriateness of services received.

They should be reliable so there is reasonable consistency from question to question and from time to time. Reliability is a complicated problem. What do we want? Do we want to have a very high internal consistency, so that the answers to every item in a scale agree with the answers to every other item? This, however, implies redundancy. On the other hand, we would like to have the answers to questions in a given scale agree reasonably well with each other, so that a person who strongly disliked his doctor could say so clearly. Do we want to have high consistency through time, so that the answers given in one year would agree with those given at a later time? This would mean that the service situation may not have changed over time, and a sensitive measure should reflect changes and not hide them by disproportionate constant variance reflecting only inflexible individual traits.

They should be valid, agreeing with other, supposedly equivalent measures, and with a generally acceptable intuitive understanding of what is meant by satisfaction. Validity originally meant soundness or strength. A valid measure could be counted on. How can one determine whether a measure of satisfaction is valid? Methods of estimating validity have proliferated, but there are two main approaches. The first is that a number of persons have agreed about the kinds of responses that define a concept, although the format of observations by respondents may differ. A measure should agree with other measures which have been designed to cover the same concepts. It must be recognized that this is a necessary but only minimal requirement for validity. It is quite possible that a number of measures will agree with each other, but that none is related to the intended concept. Usually satisfaction measures have merely been assumed to be valid, on the basis that their content "obviously" taps satisfaction. In addition, it is often casually assumed that such a satisfaction measure is comparable to others, and the necessity of demonstrating this comparability empirically is neglected.

The second meaning of validity has to do not only with individual differences, but also with the effects of changes in situations (Cronbach, 1957). If a measure is valid, it will not only predict individual differences in behavior which may be measured by other measuring instruments and under other situations, but it will also permit us to establish laws concerning differences among means which are predictable and interpretable. The measure should have "intrinsic validity" (Gulliksen, 1950) or understandability. That means that you cannot determine if you have a validated measure if you have performed only a few correlational exercises; you must also achieve a reasonable level of understanding of not only the correlations with other variables, but also the manner in which satisfaction is related to changes in the situation. Given these laws, one can then extrapolate to new situations and to new individuals. An adequate theory is prerequisite to evaluating the validity of a given measure in this more stringent sense.

Even with the difficulty of getting significant attitude-related variables, the demand for consumer-centered outcomes has been pronounced. Coleman stressed the importance of designing and implementing a valid consumer satisfaction system at a workshop at the Clinical Technologies conference. "Client satisfaction makes a difference in how closely you can match what the client wants with what you offer and it is vital to set up a system that allows clients to tell you what they want" (Assessing client satisfaction; 1998). He suggested the following steps that service agencies should take to insure a meaningful, workable system for collecting consumer input and producing relevant reports:

• Ask the right questions by deciding what information is needed, who will be providing it, and how will it be collected (e.g., self-report, behavioral indicators, or correlates).
• Ask the questions the right way. In other words, decide if you want to get feedback through a paper and pencil survey, an interactive computer program, or even via videotape. Since each method of administration has both strengths and weaknesses, cost and ease of administration should be carefully considered. Further, the voice and tone of questions is important.
• Choose the sample carefully. Since it is often impractical to survey all service recipients, a representative sample is important. Factors such as sex, age, level of income and race should be considered, as well as clinical and service variables that could bias results.
• Assess validity to make sure the data are accurate and representative or use an already published tool where results could be compared to normative data.
• Identify possible confounders when getting confusing or erroneous results. Such confounders could be self-selection of sample or by method. For example, if using a telephone interview, not everyone may have a phone depending on level of income. Also, if a walk-in sample is used, closed cases are eliminated.
• Consider the process for collecting surveys: when to administer the survey and how often.
• Often response rates are poor so some use a reward system to encourage participation. Other problems such as concerns about the confidentiality must be addressed to have good response rates.
• Technological aids such as using bar codes and the latest computer software may help to relieve the burden of collecting and using the volumes of information collected. 
• The results of satisfaction surveys should be subjected to various methods of interpretation. In other words, just reporting results is not sufficient. Methods of interpretation include comparison with published norms, prior data collected at the agency, or analysis of subgroups. Satisfaction results could also be correlated with other outcomes to determine if there is a relationship, and to particular services.
• Most important, satisfaction should be used to make a difference to people receiving services. Are consumers being heard and responded to based on the data that is collected?

The most frequent observation encountered in the present review of the literature was that the concept of satisfaction lacked a generally agreed upon definition, with authors offering a variety of often similar and overlapping definitions. This is because satisfaction is an abstract concept with no predetermined content. It is up to the user to either adopt someone else's definition, or to construct a definition of his/her own. Similar to other abstract concepts such as Quality of Life (Bergner; 1989), each investigator must think about his or her own study, the study population, and the intervention and decide what should be assessed. Once the domains or categories or variables are identified in this way, an investigator can start to review existing measures to determine appropriateness. The problem becomes particularly acute when satisfaction has to be defined across a variety of different cultures. It is as though Humpty Dumpty were speaking to Alice in Through the Looking-Glass, "When I use a word,...it means just what I choose it to mean-neither more nor less." (Carroll, 1872).

Sometimes authors simply offer a descriptive definition of what is meant by satisfaction. Seldom is a general theory mentioned. In the final analysis, it is generally the measurement instrument used that provides the specific operational definition of the concept. The operational meaning of satisfaction is no more and no less than the measure used, although a theory-linked definition may have broader implications extrapolated from the theoretical framework.

The wish to use a 'valid' satisfaction measure needs to be refined by specifying the kind of validity that is sought. There is no instrument valid for all purposes and all situations.

Another psychometric property needed is reliability. Just as with validity, several different types of reliability must be demonstrated. Not all satisfaction instruments that are reviewed can demonstrate reliable results.

Another problem that must be addressed in defining satisfaction is the content of the items chosen to measure the concept. One common distinction made in the measurement literature is that there are (1) subjective items and (2) objective items. The subjective items address how the subject feels about some aspect of his/her life. For instance, how they feel about the quality of their care or if they were treated with respect. Objective items that would classify satisfaction might include the rate of grievances and complaints at a service agency. Satisfaction scales contain only subjective items. It is frequently observed, however, that improvements in objective conditions do not always correlate with improvements in subjective. This leads to further ambiguities in how to understand the results from these two aspects of satisfaction.

Because the items chosen to define satisfaction scale are not equally important, scales in general seek to weight these items, usually based on the subjective ratings of importance provided by the subjects. This is not true of satisfaction scales at this time, and no weighted satisfaction estimate along a common measuring scale is currently available.

In addition, some scale makers argue that the contents of the scales should be uniquely determined by each subject in order to provide a meaningful measure of their satisfaction. This would provide a satisfaction estimate based on a unique set of items for each individual that could not be compared directly with another group or individual Such scales are primarily used in for longitudinal studies of change in individuals or intact group.

The perception of satisfaction by consumers and providers is often not the same. People judge satisfaction only in present time, and do not forecast the future. In addition, the criteria used by consumers may change from time to time depending upon the relationship of the person to the mental health system and to everyday life.

In summary, the tests that satisfaction scales must meet are no different than those expected of any other psychometric instrument. The problem is somewhat more complex for satisfaction measures because: (1) there is no generally agreed upon definition, (2) there is no generally agreed upon theoretical framework, (3) there is no generally agreed upon set of items of domains of measurement, and (4) there is no generally agreed upon weighting of value or importance. Consequently, none of the scales developed thus far can satisfy all the demands of a sound measure.

In view of these circumstances, the potential users of a satisfaction scale can consider three possible courses of action:

1. they can review the scales and their characteristics and chose the one that comes closest to meeting their needs;

2. they can determine to develop their own scale and commit to establishing its psychometric qualities; or

3. they can ask consumers to identify those aspects of their lives that psychiatric services are intended to influence and identify items that will help determine whether or not providers were successful in meeting their expectations.

If the choice is to pick one of the existing scales, you gain the advantage of being able to compare results. One must recognize, however, that the measure may fail to meet a variety of psychometric qualities and will probably continue to do so for the foreseeable future. The task of completing a psychometric instrument for a complex concept like satisfaction can be a lengthy one. One must also accept that because of its sensitivity to changes in subjects and situations, its characteristics of validity and reliability will also fluctuate over time.

If the choice is to develop a new scale, one can gain the advantage of having a scale that is designed to measure satisfaction as the service recipients conceive of it. You face, however, a formidable task of psychometric development, before you can even begin to examine some preliminary results. And, while a systematic development of such an instrument might bring one incrementally closer to a measurement goal, one also face a lengthy task of refining and re-calibrating their measure as programs and circumstances continue to evolve.

If the choice is to identify those elements of consumers' experiences, perceptions, and motivations one can simplify the task by focusing on a finite group of specific elements. There may already be specific instruments designed to address some of those elements. If there are not, they have the standard psychometric task ahead of them. However, the task should not be as complex as it would be for the first two strategic choices. It is less complex because they would give up the pretense that they are measuring the multidimensional, global, abstract concept of satisfaction.

No doubt the concept of satisfaction will continue to be used in popular discourse and in the professional literature. However, after 40 years of experience and thousands of attempts to shape this concept into a scientifically precise measure, it may be time to recognize that satisfaction is not a concrete thing out there that can be defined and measured with great precision, if we only try hard enough. It is, instead, an abstract concept that can be defined in a variety of reasonable ways that will always have a personal and situational meaning. Conceptually, it can never serve as a stable, concrete yardstick of some objectively determined characteristics of service delivery.

The definition and measurement of satisfaction starts not with an act of discovery of something 'out there' but with an act of creation, a vision of something we value and wish to define as 'satisfaction.' This is not an invitation to define satisfaction in any arbitrary way possible, as ??? warns against. There is still much to be gained by attending to how others have conceived of the concept, in order to reap the benefits of a cumulative development of an idea.

While accepting the abstract nature of satisfaction and the legitimate variability of its definition may seem to be a defeat, it may actually be a good thing. Consider the consequences if the reverse were true. If everyone agreed on what constituted a satisfying service episode, would everyone strive to seek and use services in precisely the same way? If everyone agreed on the dimensions of quality, would we all want to live in the same city, drive the same car, eat the same food, read the same books, attend the same movies, have the same number of children, or marry the same spouse? If we observe what people do in all those instances, we recognize that there is a fair amount of consensus on some of those matters but also a great amount of diversity. We would also recognize, that the answer at one point in our lives might differ from the answer we would give at another point.

There may be no harm in continuing to use satisfaction as a general organizing concept. However, the continued attempt to measure it as a precisely defined entity may be a self-inflicted wound that need no long preoccupy our scientific energies. If there are measures that someone chooses to call satisfaction that have sufficiently robust psychometric properties that they can measure change, characterize different groups, or reflect the impact of programs, they could serve a useful purpose. As long as we recognize that satisfaction, like ice cream, can come in many delightful colors and flavors, we would avoid the conceit of thinking we have captured a universal and generic property of the world.

In conclusion, it is important to consider how behavioral health outcome domains that have been recommended by consumers are to be included in the research and evaluation of psychiatric systems of care. Twenty years ago Prager and Tanaka (1979) stated, "Representing the consumer's perspective on the meaning of mental illness and the correlates of getting better, the process of client involvement in evaluation design and implementation is not only realistic and feasible; it is, we feel, a professional necessity whose time is overdue" (p. 51). However, without developing ways for constructive dialogue to occur and shared decision-making to take place, the focus on mental health outcomes reinforces a "kind of turf war over controlling human beings in a landscape that includes an entire array of service options and widely divergent goals and definitions of mental health and quality of life" (Scott, 1993, p. 5). The Final Report of the Task Force on the Design of Performance Indicators Derived from the MHSIP Content (MHSIP, 1993) suggests that the ideal environment for the development of a system of performance indicators is one in which (1) intents of all stakeholders are articulated and shared; there is a culture of respect for and constructive use of data; (3) changes are accomplished through participatory development; and (4) resistance is reduced through disclosure of fears and implementation of safeguards that address those fears.

Even within a collaborative framework, it cannot, however, be assumed that if data elements are refined and collected through statistically valid methods, system functions will be accepted or optimized. Most research literature on evaluation methodologies, as well as discussions of quality assurance information systems now in use in health environments, focus primarily on defining the appropriate indicators and data collection instruments to capture needed information. In fact, there appears to be a certain naive faith that if the "right" data collection tools can be developed, quality and cost effectiveness would be axiomatic. On the contrary, for research and evaluation to realize its promise of improving services and reducing costs, attention must also be paid to interrogating latent values inscribed in the social relations of measurement, and to integrating the structural components of information systems. Outcome studies need to be grounded in theory building through reflexive modes of discourse in order to respond to the epistemological politics of knowing and acting. Profound issues of power and knowledge are oftentimes translated or conceived of in terms of knowledge exchange and decision support. It is not surprising that as we enter a medical marketplace environment that the customers of mental health services are demanding a voice that matches or even surpasses all others in how research and evaluation shapes its content, its character, and its influence on society. Therefore, the processing and presentation of information as it flows through the healthcare system, and the distribution and feedback loops that would allow multi-stakeholder decision-making to take place must be redesigned to support quality improvement partnerships in policy, management, and clinical practice. Ultimately, it is at the level of people communicating with each other that human values have the greatest potential for becoming incorporated into the institutions of society. Through an open architecture of knowledge production and exchange research and evaluation can become a value-added force in the delivery of mental health services. 

Abdellah, F.G. & Levine, E., (1957). Developing a measure of patient and personnel satisfaction with nursing care. Nursing Research, 5, 100-108.

Abdellah, F.G. & Levine, E., (1958). Effect of Nurse Staffing on Satisfaction with Nursing Care (Hospital Monograph Series No. 4). American Hospital Association: Chicago, IL.

Abramowitz, S., (1987). Analyzing patient satisfaction: a multianalytic approach. QRB, April, 122-130.

Aday, L.A., (1975). Economic and noneconomic barriers to the use of needed medical services. Medical Care, 13, 447-456.

Aday, L.A. & Andersen, R.A., (1974). A framework for the study of access to medical care. Health Services Research, 9, 208-220.

Aday, L.A. & Andersen,R., (1975). Access to Medical Care. Health Administration Press: Ann Arbor, MI.

Allen, H., Baigent, B., Kent, A. & Bolton, J., (1993). Rehabilitation and staffing levels in a 'new look' hospital hostel. Psychol. Med., 23, 203-211.

Alpert, J.J., Kosa, J., Haggerty, R.J., Robertson, L.S., & Heagarty, M.C., (1970). Attitudes and satisfactions of low-income families receiving comprehensive pediatric care. American Journal of Public Health, 60, 499-506.

American Psychological Association, (1974). Standards for Educational and Psychological Tests. Washington, DC: Author.

Andersen, R.A., (1968). A Behavioral Model of Families' Use of Health Services (Research Series No. 25). Center for Health Administration Studies: Chicago, IL:

Andersen, R., Kravits, J., & Anderson, O., (1971). The public's view of the crisis on medical care: an impetus for changing delivery systems. Economic and Business Bulletin, 24, 44-52.

Anderson, J.G., (1973). Health services utilization: framework and review. Health Services Research, 1, 184-199.

Andrews, F.M. & Withey, S.B., (1976). Social Indicators of Wellbeing: American Perceptions of Life Quality. Plenum Press, New York.

Andrews, G., Teesson, M., Stewart, G. & Hoult, J., (1990). Follow-up of community placement of the chronic mentally ill in New South Wales. Hosp. Community Psychiatry, 41, 184-188.

Andrews, S., Leavy, A., DeChillo, N. & Frances, A., (1986). Patient - therapist mismatch: we would rather switch than fight. Hosp. Community Psychiatry, 37, 918-922.

Apostle, D. & Oder, F., (1967). Factors that influence the public's view of medical care. Journal of the American Medical Association, 202, 140-146.

Armor, D.J., (1974). Theta reliability and factor scaling. In Costner, H.L. (Ed.), Sociological Methodology, 1973-1974. Jossey-Bass Publishers: San Francisco

Attkisson, C.C. & Greenfield, T.K., (1994). Psychological Testing: Treatment Planning and Outcome Assessment. In Maruish, M. (Ed.), The Client Satisfaction Questionnaire-8 and the Service Satisfaction Questionnaire-30. pp. 402-420. Lawrence Erlbaum Associates: San Francisco.

Attkisson, C.C. & Greenfield, T.K., (1996). Outcomes Assessment in Clinical Practice. In Lloyd I. Sederer & Barbara Dickey (Eds.), The Client Satisfaction Questionnaire (CSQ) Scales and the Service Satisfaction Scale-30 (SSS-30) pp. 120-127. Williams & Wilkins: Baltimore, MD.

Attkisson, C.C. & Pascoe, G.C. (Eds.), (1983). Patient satisfaction in health and mental health services. Eval. Program Plann. 6(Special Issue), Pergamon Press, New York.

Attkisson,C.C. & Zwick, R., (1982). The client satisfaction questionnaire: psychometric properties and correlations with service utilization and psychotherapy outcome. Evaluation and Program Planning, 5, 233-237.

Azim H.F.A. & Joyce, A.S., (1986). The impact of data-based program modifications on the satisfaction of outpatients in group psychotherapy. Can. Journal Psychiatry, 31, 119-122.

Azrin, S.T., Coursey, R.D., & Fallot, R.D., Towards a Theory of Life Satisfaction Judgment for Persons with Serious and Persistent Mental Illness.

Bashshur, R.L., Metzern, C.A., & Worden, C., (1967). Consumer satisfaction with group practice: the Cha case. American Journal of Public Health, 57, 1991-1999.

Bass, B.M., (1956). Development and evaluation of a scale for measuring social acquiescence. Journal of Abnormal and Social Psychology,