Patients' preferences for technical versus interpersonal quality when selecting a primary care physician

Health Services Research, August, 2005 by Constance H. Fung, Marc N. Elliott, Ron D. Hays, Katherine L. Kahn, David E. Kanouse, Elizabeth A. McGlynn, Mark D. Spranca, Paul G. Shekelle

For example, we enabled participants to control the amount of information that they would view at any given time to decrease cognitive overload (Figure 1) (Vaiana and McGlynn 2002). We chose to present the report cards on computer so that we could employ technology that people frequently encounter on websites, including existing health care report card sites (The Pacific Business Group on Health 2003; United States Office of Personnel Management 2003a): click on an underlined word or phrase with the computer mouse to access more information about the word or phrase on a new web page. This feature enabled us to provide both summaries and details about the quality of each physician. The ratings on each set of web pages were the arithmetic mean, rounded to the nearest integer, of more detailed ratings found on the next set of web pages. For instance, Dr. Amber received a three-star rating and two four-star ratings for his care of high blood pressure (Figure 1). Since the arithmetic mean of these ratings after rounding equals four, his overall rating for high blood pressure care, which is displayed next to the label "high blood pressure," was represented with four stars.

[FIGURE 1 OMITTED]

Furthermore, we tried to improve the evaluability of our report cards by using techniques suggested in prior studies (Hibbard et al. 2001). For instance, we provided a consistent number of subheadings for technical and interpersonal quality on the top layer of the report card. In addition, we enabled participants to select the presentation format that they preferred: stars, bars, letter grades, or numbers, because it is not clear that one presentation format facilitates decision making more than others, and allowing participants to select the format they prefer could make the report cards more useful (Vaiana and McGlynn 2002). Each participant viewed all technical and interpersonal quality ratings in the format he or she selected so that visual cues were consistent for both types of quality information.

Moreover, a health literacy consultant applied the Fry Readability Formula (1) to sections of the introductory and report card pages that had sufficient text to yield valid results, identified words and concepts that might be difficult for some readers, and provided overall comments based on the consultant's expertise. The Fry Readibility Formula applied to the introduction and to the section titled "Details about Today's Session" yielded an eighth grade reading level. Furthermore, in two pilot sessions participants provided verbal and written feedback on the introductory and report card pages, indicating areas that were confusing. The majority of pilot participants found the report cards "not at all" confusing. Of the three who found the report cards "somewhat" or "a lot" confusing, two participants stated that the source of their confusion was that they were forced to make difficult tradeoffs, which demonstrates comprehension of the purpose of the study. In the main sessions, we asked respondents to describe in their own words on paper the meaning of all six subheadings under "Technical Evaluation" and "Patient Experiences." Ninety-four percent of respondents' answers were judged appropriate by two independent reviewers ([kappa] = 0.56).