Letter

Satisfaction as an Outcome Measure: Reply

In Reply: I appreciate the comments of Drs. Awad and Voruganti. They appear to agree with the two points I tried to make in the brief commentary: use of a single outcome indicator alone (such as a global score of patient satisfaction) can lead to misguided policy decisions and a global measure of satisfaction fails to capture the complexity of the concept. Such short commentaries without references preclude extensive elaboration on a topic. Unfortunately, Drs. Awad and Voruganti read more into the commentary than was intended. I did not indicate that people with mental illness are incapable of expressing their "inner feelings." In fact, I stated that measures of satisfaction are "essential to understanding clients' preferences and providing feedback to professionals and managers." I've been involved in quality-of-care research for more than 20 years and have advocated since the beginning for broader measures of quality of care, especially those that go beyond measures of clinical symptoms alone and incorporate client perspectives ( 1 ).

Although Drs. Awad and Voruganti agree that more work is needed to improve satisfaction measures, they seem to believe that there is enough existing research to convince us that we can use the measures without concern. Perhaps I have misread their letter, but if they do believe this then they are mistaken. Much progress has been made in the area of satisfaction measures but much more work needs to be done before we can feel confident using them as outcome measures (especially alone) to indicate where to intervene in improving quality of care. This is true not only in mental health but across the rest of health care, including oncology. In fact, given the problems with satisfaction ratings, the Institute of Medicine chose not to include them in its influential report on quality ( 2 ). Others have provided very thoughtful assessments of the current state of such measures in the mental health area and have concluded that more research is needed ( 3 ). This doesn't mean that satisfaction can't be determined; it means we need better ways of measuring the concept than exist currently. This includes going beyond traditional survey methods and incorporating better qualitative techniques.

The goal is to improve the quality of the care we deliver so that the lives of people with mental illness will be better. To do so, we must have reliable and valid measures of the various components of quality to know where to make strategic changes in the delivery of care ( 4 ). To assume that we have enough information to use current measures of satisfaction alone as adequate outcome indicators of quality would not only set us back far more than 30 years but delay any further improvement in health care. This could result in the worst outcome—people who are neither contented nor better.