Measurement of Patient Satisfaction

In Response: We appreciate the interest and comments made by Lemay et al. A systematic review of the literature in patient satisfaction is difficult because investigations into this area are inconsistently defined, disunified, and frustratingly unfocused in both intent and methodology. To acquire the 20 studies we reported, several different search strategies of MEDLINE were needed. We deliberately limited targets to those studies that had measured patient satisfaction in relatively large patient samples because these were most likely to have measured patient satisfaction as a primary outcome and used instruments that would be most informative to the issue of patient satisfaction measurement in anesthesia. Because we are investigators in the field of anesthesia research, we included existing studies despite the variable quality, as these are the only studies to look at patient satisfaction with anesthesia care. We also chose to include large anesthesia surveys that used questions that we knew only indirectly measured patient satisfaction (e.g., asking whether patients would return to have the same anesthetic again or would undergo day surgery again). These questions help to demonstrate the existing latitude and corresponding ambiguity inherent in past measurements of patient satisfaction and reinforce the common limitations seen in all studies. Excluding them would not have changed the message. Social desirability is one of a wide range of potential factors that has been proposed as potential confounders of patient satisfaction [1]. We chose to highlight a few major confounders that undermine the validity of current measurements in anesthesia care; more comprehensive lists of confounders are available elsewhere [1,2]. Unfortunately, even in other healthcare domains, only a few confounders have been directly studied. Until they are, their influence is uncertain. For example, although preexisting expectations had great theoretical influence on patient satisfaction, their confounding effect when measured has, so far, been small [2]. One of the major tasks facing future anesthesia investigators in patient satisfaction will be to decide which factors to control. Finally, we did not mean to suggest that the studies of Whitty et al. [3] or Dexter et al. [4] be viewed as models of rigorous psychometric methodology. Rather, as we stated in the review [5], these studies were the first to use psychometric methodology of any degree of rigor and, even so, provide evidence that greater methodological rigor will yield instruments of greater value to anesthesiologists. For examples of truly rigorous use of psychometric methods in questionnaire development in healthcare, one should turn to the efforts of Ware et al. [6] or Meterko et al. [7] in ambulatory and hospital care.