The Physician’s Role in Educating Patients: A Comparison of Mailed Versus Physician-Delivered Patient Education

Sample

Using data from a survey for assessing patient satisfaction with their physicians,³¹ we calculated that a sample size of 1000 patients per group would be needed to detect a difference of 2 points on a summary satisfaction scale (a=0.05; power=0.80).

We used an automated scheduling system to identify eligible patients who arranged an office visit with a physician in family practice or general internal medicine. In the 3 clinics that have both internal medicine and family practice physicians, we stratified the sample so there was an equal distribution between the 2 specialties. The fourth clinic had internal medicine physicians only. The sample population resides primarily in the first and second ring suburban communities of a major metropolitan area. People in these communities are predominantly white, well-educated, employed, and have health insurance.

Patients were assigned to 1 of 3 groups, depending on the week that they had an office visit to a particular clinic. The patients in the physician-delivered information group received the Well-Advised book from their physicians during an office visit. Physicians were instructed to explain to patients that the book was based on medical guidelines approved by Park Nicollet Clinic and that using it could prevent unnecessary clinic visits. The physicians encouraged patients to use the book for taking care of common health problems. The patients in the consumer-direct group received Well-Advised at home through the mail, with a letter encouraging them to use the book and the nurse health information telephone line. The patients in the usual-care group did not receive Well-Advised and served as a control group.

Training

Before initiating any interventions, physicians were instructed to introduce the book Well-Advised as an opportunity for patients to play an active role in medical decision making. Health educators introduced our study at staff meetings and provided a written recommended script for physicians to use when offering the self-care book to their patients. Physicians who could not attend the self-care in-service meetings were given the written script by their department’s administrator or head nurse.

Study Design

The research design was a nonrandomized controlled study with 2 intervention groups. We used a postintervention questionnaire to test group differences. The 4 test clinics comprised 3 study sites. We considered the 2 smaller clinics one site for purposes of rotating the interventions. For each study period, we designated patients from one site as the physician-delivered self-care education group, a second site as the direct-mail group, and the third site as the usual-care group. The intervention at each site rotated so that each clinic participated in both interventions and the control. The goal was to enroll 1000 patients per study site. The intervention period lasted 1 to 3 weeks depending on site size, with a 1-week break between interventions.

Within 2 weeks of the patients’ visits, questionnaires were mailed to patients’ homes to assess general satisfaction with provider communication and specific satisfaction with the self-care materials. The questionnaire was designed to compare the level of satisfaction with variables such as provision of information, physician listening, question answering, explaining, and reviewing of educational materials. Patients were also asked about the value and utility of the Well-Advised book, and their intentions to use it. We combined questionnaire items for the analysis of findings. For categorical variables, we used chisquare to test for significant differences between groups. For continuous variables, we used analysis of variance (ANOVA) to test for differences between groups. Scales that summarized comparable items were developed that combined items related to satisfaction with the physician into one category and satisfaction with the self-care book into another category.

Results

We enrolled a total of 2954 patients in our study. Of these, 2140 (72.4%) returned surveys. Table 1 shows the demographics of respondents compared with nonrespondents. As is typical for patient satisfaction surveys, respondents were more likely to be women and older. Younger men were underrepresented. There were no significant differences in response rates across study sites or between departments.

Patients who received self-care information from their physicians were significantly more satisfied with their care and with their physician-patient communication experience than those in either the direct-mail group or the control group. There were no significant differences between the control group and the direct-mail group. The physician-delivered information group also rated the overall quality of care significantly higher than the other groups. In the majority of measures, the physician-delivered information group was significantly more satisfied with the clinical visit. Using a 5-point scale (where 1=strongly disagree; 5=strongly agree), respondents in the physician-delivered information group were significantly more likely to agree that the physician spent enough time with them, explained things clearly, and listened to what they were saying. Those in the physician-delivered information group were also significantly more likely to agree that the physician provided information so they could make decisions about their own care. Receiving the book from the physician also increased patients’ perception that the physician answered their questions and was concerned about them as patients. Not surprisingly, this group was significantly more likely to agree that the physician reviewed educational material with them. Table 2 shows individual comparisons of patient satisfaction with the physician interaction. There was no significant difference among the groups as to whether the patients believed the Well-Advised book offered a credible resource for making health decisions; the large majority agreed that the book was a credible source.