Measuring Patient Satisfaction with Anesthesia Care

In clinical settings such as anesthesia, using patient satisfaction as an indicator to monitor the quality of clinical care has potential merit. For patients, satisfaction represents, at least in theory, an evaluation of the healthcare experience based on their own values, perceptions, and interactions with the healthcare environment. For healthcare providers, patient satisfaction can be used to assess the actual impact of healthcare processes on the patients themselves. Patients' ratings of their satisfaction can reflect many facets of care not easily examined in any other manner: compassionate bedside skills, efficient attendance to needs, participation in decision-making, and adequate communication and information 1. An ideal measure of patient satisfaction could therefore provide unique feedback on the quality of practice for medical specialties such as anesthesia. A renewed focus has been sought for patient satisfaction as a clinical end point and a quality indicator of anesthesia care. Previous studies have demonstrated the limitations of using adverse anesthetic events to monitor anesthesia care 2. Major outcomes (death and complications such as myocardial infarction, cardiac arrest, and severe brain injury) are too rare to allow useful comparisons among institutions. Assessing minor outcomes (such as pain and nausea) pose significant methodological problems of uncertain case mix, inconsistent reporting compliance, imprecise definitions, and under- and overreporting. Orkin et al. 3 emphasized the still unproven relationships between most intermediate events (for example, transient intra- and postoperative hypoxemia) and clinically significant adverse events (death, myocardial infarction, etc.), as well as the difficulties in establishing those relationships. They also suggested that, in light of the demonstrated safety of anesthesia, the field may have become a victim of its own success and that a redefinition of what constitutes important anesthetic outcomes is necessary. This is echoed in a review on quality models for anesthesia care by Eagle and Davies 4, who cite preliminary evidence of the benefits of examining positive outcomes such as patient satisfaction 5. In an accompanying editorial, Duncan 6 states that quality in anesthesia should encompass that which satisfies all of its customers, which includes surgeons, administrators, fellow anesthesiologists, and patients. In an editorial, Klock and Roizen 7 hoped to "wake up" the specialty of anesthesia to the potential value of patient satisfaction as a means to assess and improve the quality of modern anesthesia practice. Thus, researchers and authorities on anesthetic outcomes and quality have articulated the need to expand the scope of important anesthetic outcomes and have targeted patient satisfaction as a potentially useful indicator of the quality of anesthesia care. Despite the many attractive features of patient satisfaction, current measures of satisfaction in anesthesia care suffer from lack of refinement and have uncertain reliability and validity. Cross-sectional surveys using single-item questions and yes/no or Likert response formats 2,8-11 have yielded uniformly high scores (>80% satisfied or very satisfied). Unfortunately, it is unclear what these global ratings mean. Are patients truly satisfied with their anesthesia care or merely expressing their satisfaction with their surgical or hospital care? Do patients base these positive ratings on a single factor (such as intact survival) or on several criteria (friendliness, sensitivity, compassion, information, and communication)? Are reports of satisfaction biased by patients' respect, trust, confidence, and gratitude to their doctors, nurses, and healthcare in general (a so-called halo effect)? The inability to answer these types of questions limits the utility of simple measures of patient satisfaction in anesthesia. If patient satisfaction can truly help to monitor the quality of anesthesia care, a better measure of patient satisfaction that has proven reliability and validity is needed. In this article, we attempt to lay the foundation for the construction of such a measure for anesthesia. We first review the methodology of patient satisfaction measurement and discuss the inadequacies of current measures in the anesthesia literature. We then outline the early important considerations unique to the setting of anesthesia that support and guide the construction of a more reliable and valid patient satisfaction questionnaire for anesthesia care. What Is Patient Satisfaction? A frequently cited definition formulated by Pascoe 12 envisages patient satisfaction as healthcare recipients' reactions to their care, a reaction that is composed of both a cognitive evaluation and an emotional response. Pascoe's model, synthesizing concepts from psychology with those from theories of consumer satisfaction, goes on to propose the process by which patients evaluate their care. Each patient begins with a comparison standard against which care is judged. This standard can be any of the following: an ideal, a minimal expectation, an average of past experiences, or a sense of what one deserves. Furthermore, the patient can assimilate discrepancies between this expected standard of care and that which is actually experienced. A change in satisfaction occurs when the difference between actual and expected care exceeds a patient's capacity to assimilate that difference. This change in satisfaction does not necessarily mean dissatisfaction, because a patient may start out with a very low expectation of the standard of care. Actual standard of care that sufficiently exceeds this expectation can change a patient's initially negative opinion to a positive one. Put simply, patient satisfaction depends on the congruence between what is expected by the patient and what occurs to the patient 13. Although of some interest, much of this theory is speculative and incomplete and poses more questions than answers. For example, how can we measure a patient's actual mental state when they say they are satisfied? Is patient satisfaction a cognitively based attitude, an emotion, an intrinsic psychological trait (e.g., a tendency to be grateful), a cultural attitude about health and healthcare, or some combination of all of these elements 14? We also do not know what determines a patient's expectations or how different facets or dimensions of care are incorporated into a global rating of overall care. Finally, we do not know what determines a patient's capacity to assimilate a discrepancy between expected and actual care. Psychological models of satisfaction such as Pascoe's appreciate the underlying complexity, but they are too preliminary to guide the construction of a valid measure of patient satisfaction. To capture the scope of patient evaluations of broad healthcare domains (such as primary ambulatory care and hospital care), a satisfaction questionnaire must span broad attributes or components of care. Two prominent examples are the sophisticated questionnaires specifically designed to evaluate primary ambulatory care and hospital care: the Patient Satisfaction Questionnaire (PSQ) 15 and the Patient Judgments of Hospital Quality Questionnaire (PJHQ) 16. The PSQ contains eight dimensions (Table 1) that reflect components of ambulatory care (technical, interpersonal care, availability and continuity of care, etc.). The PJHQ is divided into a taxonomy of dimensions (Table 1) representing hospital services or phases of hospital care such as medical, nursing, admissions, and discharge.Table 1: Dimensions of the PSQ (ambulatory care) and the PJHQ (hospital care)When evaluating specific components of provider care, patient satisfaction seems to have dimensions that consistently separate care into technical and interpersonal components. Investigators in nursing care divide patient satisfaction into dimensions of technical care, education, and trust 13,17,18. Similarly, in medical care, patient satisfaction with the care of physicians is categorized into technical and interpersonal care dimensions, either within larger questionnaires such as the PSQ, or within smaller instruments such as the Medical Interview Satisfaction Scale, which measure satisfaction with a specific component of physician care 15,16,19. Although researchers are able to distinguish the technical and interpersonal dimensions of physician or nursing care, patients do not seem to be able to do so. Statistical analyses of satisfaction ratings in nursing care and in ambulatory care suggest that these technical and interpersonal dimensions are not always evaluated independently by patients 13,15,20. These results support the contention of most researchers 21,22 that the interpersonal characteristics of specific episodes of care (such as affect, communication, and information) are indistinguishable from the technical aspects of that care to the patient. Patient satisfaction with a provider's (such as an anesthesiologist) care may be construed as a unidimensional measure of the quality of the provider-patient interaction during an episode (or episodes) of care. Psychometric Questionnaire Construction In light of its uncertain psychological foundation and multidimensional complexity, the development of more reliable and valid instruments has followed a formal methodology of questionnaire construction first developed to measure complex psychological phenomena such as intelligence 15,16,19,23. Rather than rely exclusively on single-item ratings of patients' global satisfaction with their entire experience, a psychometric questionnaire uses multiple items to probe specific events or concerns that occurred in that experience, events that together determine patients' satisfaction with their care (or dimension of care) 24. If a global satisfaction rating is necessary, ratings of individual items can be summed or averaged. The construction of a psychometric questionnaire follows a rigorous protocol (Table 2) 25,26. Initially, a formal phase of item generation is undertaken to develop a bank of items to include all important elements of patient satisfaction. A comprehensive list of such items are gleaned either from focus groups of patients or providers or by one-on-one interviews. By directly soliciting the input of patients, investigators obtain items that represent what patients actually value; scale items will have content validity to patients. By soliciting the input of providers, investigators can better ensure that significant elements of care have not been missed. Accommodating the insight and emphasis of providers during item generation improves the face validity of the final instrument to the providers who will ultimately use it. As a result, items in the questionnaire represent the actual views and values of those who have personal experience with the domain of care being evaluated.Table 2: Steps of Psychometric Questionnaire ConstructionOnce generated, these items are then grouped into dimensions of care. The grouped items are formulated into an initial questionnaire version that is tested in a large sample of patients. Detailed analysis of patient responses to one or more pilot tests of a maturing questionnaire version reduces the original items and dimensions into a shorter, more feasibly administered final questionnaire version 15-17,27. For example, in developing the PSQ, Ware et al. 15 collected 2600 items from a literature review of existing scales and open-ended interviews with ambulatory care patients. Content analysis was used to group items into 18 dimensions of care. From these items and dimensions, study investigators ultimately constructed a 68-item pilot questionnaire (PSQ-1) containing eight dimensions 15. Statistical analysis of responses to reduce the initial pilot questionnaire into the final questionnaire version is a cornerstone of psychometric scale construction. Items that produce excessively skewed distributions or have a high rate of missing responses are eliminated. Items that truly reflect satisfaction are identified by their higher item-total or item-dimensional correlations and better homogeneity (Cronbach's alpha) of the dimensional and overall satisfaction scale that contain them 25. Reducing questionnaires to only these items yield item, dimensional, and overall summed scores with less skew, greater variability, and good to excellent internal consistency and test-retest reliability 15. The use of statistical methods alone can be misleading, as items could be grouped into dimensions that do not represent true elements of care. For example, factor analysis of patient responses to a standardized satisfaction questionnaire for hospital nursing care, the LaMonica-Oberst Patient Satisfaction Scale, yielded factor solutions that merely reflected the positive or negative direction of item wording 13,20. Dimensions are most powerfully demonstrated when the conceptual themes of care identified in item generation also emerged in factor or correlational analysis, as in the PSQ and PJHQ 15,16. The construct validity of the items is then established by demonstrating significant correlations between scores on items/dimensions and alternate measures of satisfaction, such as a patient's intention to recommend or return to care 16,28, by being able to discriminate among test sites 16 or by predicting patients who have had more patient-centered care 17 or greater perceived improvements in their care 27. Patient Satisfaction with Anesthesia Care The literature on patient satisfaction with anesthetic care is sparse and unfocused. We searched MEDLINE between 1966 and 1997 to retrieve anesthesia studies that measured the level of patient satisfaction (either directly or indirectly) in large samples of patients (n > 100) (Table 3 and Table 5) after general anesthesia 9,11,29-46. Careful scrutiny of these studies revealed common crucial limitations of traditional ratings of patient satisfaction.Table 3: Patient Satisfaction Instruments in Anesthesia According to the Type of Rating UsedTable 5: Table 3.-(Continued)The results of the single-item or multiitem ratings of patient satisfaction with overall or specific aspects of anesthesia care used by all 20 studies, taken at face value, offer some reassuring evidence that 80%-100% of patients are satisfied or very satisfied with their anesthesia care. Are these ratings of satisfaction a true indication of the care that patients had received? That they may not be is dramatically demonstrated by a woman who claimed she was satisfied with her general anesthetic despite being awake during her operation 36! Indeed, a number of patient and methodological factors may promote significant discrepancies between the actual quality of care received by the patient and the level of satisfaction reported in a single-item rating. First, in settings of real or perceived high risk, such as an operation or an anesthetic, satisfaction ratings may be dominated by a sense of relief that the was As a result, patients who state that they are satisfied may be merely expressing their and gratitude to the providers who them This could much of the of in which of patients physicians and only and their patients may be to their and This may be to patients' trust in the healthcare a that is their or a of any care. As by Pascoe these in the healthcare the validity of all satisfaction In anesthesia, these may for the high of patient satisfaction or in studies in which satisfaction ratings from patients in settings patients may be more to or on positive negative If this is then in care may be A tendency to only positive events may be one patients anesthetic complications when have a with most of the patients at least one they from a list of specific complications positive in most in single-item them to in the satisfaction of patient As a result, has been able to that or in care can to in measured patient satisfaction. For example, when et al. the impact of postoperative on patient satisfaction with the quality of their care, they the satisfaction scores in patients to or postoperative The from single-item ratings in all groups was consistently a of 5) the patients their their the of their or the of the As out in the accompanying by Klock and Roizen the results of et study more about the and of their patient satisfaction instrument than it does about the actual value of the postoperative of the multiitem scales are is more than single-item ratings scales to more patients who of satisfaction with anesthesia or care with single-item scales scores are only the items represent the of patient satisfaction most important to patients. to these during scale construction items and scales that have content validity. For example, and reported high of patient satisfaction (>80% had average in several of care information, operation information, anesthetic information, pain postoperative and Unfortunately, these of care may not to the patients their care or reflect what patients most value in their care. Furthermore, the wording of the items to what of care In items not better in the actual it is that patients not know to which specific of their anesthesia experience item In a divided her scale into dimensions and and items for reported that more patients the medical and nursing items the as or than the or items or any item in the dimension the wording of the items was not it is unclear what patients rating. in methodology the and validity of anesthesia satisfaction studies have to crucial such as the in which patients patient such as health and the case of their the characteristics of and the of questionnaire Each of these factors patient ratings In study in Table 3 and Table sought to the reliability of its patient satisfaction et al. and et al. reported the reliability of their instrument > but the reliability of their single-item ratings of patient satisfaction was not specifically reliability of patient satisfaction ratings in anesthesia is not or is the internal consistency of multiitem scales of satisfaction. As a result, any among patients in of satisfaction suggested by these scales may be to the of the not to in actual quality of care. In the ratings of patient satisfaction in anesthesia surveys may not measure satisfaction with anesthesia care. They to an uncertain dominated by patient and of gratitude and trust in the and nursing global ratings in are to actual in care. A valid and reliable patient satisfaction measurement first a multiitem questionnaire composed of items representing valid of patient satisfaction specific to anesthesia care. reliability and demonstrating validity then occurs during pilot of the questionnaire using standardized As in other healthcare the to such an instrument has been a psychometric one. a of Patient Satisfaction unique to anesthesia care support the of psychometric methodology as a of patient satisfaction measurement in anesthesia and help to guide early in psychometric questionnaire construction. First, the results of preliminary studies the state of patient about anesthesia care determine the of patients to distinguish anesthesia from aspects of their care of patients are of the and of If patients do not know what anesthesia care then they may be to distinguish their anesthetic care from that by the of the and valid ratings of satisfaction with anesthesia care a multiitem questionnaire that the of patients who may be about the and of anesthesia care the elements of their care that are both specific to anesthesia and important to surveys by in have yielded a large list of patient of their anesthesia care (Table The span the entire and are The results of these surveys that, individual patients may have of their anesthesia experience their provide a of the entire experience when in multiitem Patient of because surveys have sought the input of patients in the of the items anesthesia surveys to an uncertain the of the who constructed surveys designed to measure specific outcomes and and not to measure patient evaluations or ratings of their care. The items in these surveys may therefore be dominated by what patients actually or perceived in their care more than by what patients actually information, more communication, etc.). As a result, items representing true of patient satisfaction may items different from those in surveys These items will during formal phases of item Finally, the development of a patient satisfaction instrument for anesthesia care is more than in other healthcare a greater level within the can the of anesthetic care. An mental or emotional state as a of medical or surgical that may the of general the use of to postoperative and transient postoperative may some patients or to their Thus, initial psychometric could focus first on or more specific domains of care, such as anesthesia, care, or specific types of all of which have less which can better the of patients' satisfaction with their anesthesia care. Two studies the potential value of a rigorous psychometric to patient satisfaction studies on specific types of patients within their hospital practice. et al. reported the construction and pilot of a questionnaire to measure satisfaction with general anesthesia in patients and et al. developed an questionnaire to measure patient satisfaction with for minor surgical all anesthesia these investigators reported an phase of item generation questionnaire construction. et al. their items from formal focus groups with patients. et al. an initial of items that, in their the by patients during postoperative interviews and then these items with a of healthcare investigators also these items for them into a pilot studies provide evidence that item generation during questionnaire construction to produce better satisfaction et scale divided patient evaluations to dimensions of care and specific ratings within these dimensions that had better identified patients who less satisfied). Patient ratings more in items that patients to evaluate their anesthetic or their than their overall care. et al. directly that the summed scores of their questionnaire had excellent internal consistency and test-retest reliability a single-item rating of satisfaction test-retest reliability They also correlations of their multiitem with a single-item global rating of satisfaction and the ratings of providers tests of construct Although these studies the value of psychometric questionnaire crucial group sought to determine patient satisfaction in anesthesia is better divided into summed ratings of specific dimensions of care a global summed rating. group measured the of such as health or the and of questionnaire all of which have real but inconsistent on ratings of patient satisfaction Finally, it is uncertain the items developed and constructed by et al. for anesthesia and et al. for care are reliable and valid for other or in different such as anesthesia care or more complex anesthesia care. the questionnaires and items in the early psychometric of et al. and et al. and by other these studies will more rigorous studies that will develop and use sophisticated psychometric questionnaires and standardized methodology to measure patient satisfaction in Table then will ratings of patient satisfaction in anesthesia surveys or become and able to truly reflect the overall quality of modern anesthesia practice. Patient satisfaction with healthcare is a complex psychological The simple ratings of patient satisfaction used in most anesthesia surveys are to this Psychometric methodology has reliable and valid multidimensional patient satisfaction questionnaire instruments for and nursing care of formal item generation and statistical item during questionnaire construction and pilot ensure that ratings of care based on a final questionnaire version reflect the true values and of patients and represent satisfaction specific to anesthesia care. evidence that multiitem questionnaires specific to care and specific settings of care yield anesthesia satisfaction scores with proven reliability and demonstrated validity. The development of such psychometric satisfaction questionnaires is to the specialty of anesthesia but is to the use of patient satisfaction as an of anesthesia care.