When Physicians and Patients Think Alike: Patient-Centered Beliefs and Their Impact on Satisfaction and Trust

Patient Sampling and Data Collection

We identified the English-speaking adult patients of the participating physicians who could complete the questionnaires with minimal assistance. Because of the larger study’s interest in patient expectations and requests, selected patients had all indicated that they had a new or worsening problem, or that they were at least “somewhat concerned” about a serious undiagnosed condition. Contact with randomly selected patients was made through the physicians’ appointment lists 1 to -2 days in advance of the visit. During the 11- month patient enrollment period, 2606 telephone contacts were made; 677 patients declined to participate, and another 737 were deemed ineligible (69% of these because they had no significant health concern). Of the 1332 eligible consenting patients, 1071 completed screening forms, and 909 completed questionnaires at the scheduled visit.

Eligible patients filled out the Sharing subscale of the PPOS at the end of the screening interview. The instructions for the patients were the same as those for the physicians, and the items and response scale were identical to those filled out by the physicians. Immediately before their office visits, patients filled out the Trust in Physician Scale²⁶ a 9-item instrument asking patients how much confidence they have in their physicians about specific issues (eg, to always tell the truth, to put your medical needs above all other considerations, including cost). This scale could only be completed by patients who had seen the physician at least once before (n=714).

At the end of the visit, patients provided basic demographic and personal information and evaluated the physician and their visit. Visit evaluation was measured using the sum of 5 items assessing satisfaction with care received (ie, amount of time the doctor spent with you today, explanation of what was done for you, personal manner of the doctor, technical skill of the doctor, and overall satisfaction; a=.88). Using a 5-point Likert scale (from strongly agree to strongly disagree), patients additionally evaluated their physician on 3 items (ie, I would make a special effort to see this doctor in the future; I intend to follow the advice of this doctor; and I would highly recommend this doctor to a friend). These were summed to form an Endorsement of Physician Scale a=0.90). All of the evaluative instruments were scored so that a higher score indicated a more positive evaluation.

Statistical Analysis

We first analyzed the data separately for patients and physicians using 1-way analysis of variance to determine the relationship between patient-centered beliefs and personal characteristics. Then analyses were conducted to determine the relationship of (1) the patient’s beliefs, (2) the physician’s beliefs, and (3) the difference between patient’s and physician’s beliefs (“belief congruence”) to patients’ evaluations. Because patients were clustered within-physicians, these analyses were conducted using multivariate generalized estimating equations (GEE) analysis using the Stata 6.0 software (Stata Corporation; College Station, Tex) xtgee procedure. This procedure accounts for within-physician correlation, thereby ensuring that standard errors are not overestimated. In those cases where the GEE analyses indicated significant relationships, analysis of covariance was conducted to determine whether the results remained significant after controlling for potential confounding variables.

Results

Since physicians’ and patients’ attitudes toward sharing power and information were both measured using the Sharing subscale of the PPOS, we were able to compare the scores of each group. Patients’ scores covered the full possible range of the scale, while the range for the 45 physicians was somewhat more constricted (from 2.7 to 5.7). Physicians’ mean scores were significantly higher than those of the patients (4.5 vs 4.2, P <.04), indicating a stronger belief in sharing power and information. The correlation between the score of a given patient with his or her physician scores was extremely small (r = 0.03), and the observed difference scores for each pair (patient score minus physician score) ranged from 3.33 to -3.78.

Patient Characteristics and Beliefs in Sharing

The patient sample contained somewhat more women than men (56% women) and had a mean age of 57 years. More than three fourths (77%) of the sample had completed at least some college, and 30.2% had at least a bachelor’s degree; median income was in the $40,000 to $60,000 range. The vast majority of the patients were white (81.4%), with a small representation of Latinos (6.6%), African Americans (5.4%), Asian/Pacific Islanders (3.1%), and Native Americans/Alaskans (1.9%). Of the patients, 40% were being seen in internal medicine, 37% by family physicians, and 23% in cardiology.

Women were significantly more patient-centered in their beliefs, as were patients who were younger, more educated, and had a higher income Table 1. The scores of patients aged 18 to 39 years, 40 to 49 years, and 50 to 59 years were homogeneous, and as a whole they were significantly more patient-centered than those of patients between ages 60 and 69, and those 70 years and older (using post hoc-tests, the Student-Newman-Keuls statistic). Similarly, those who had completed high school or less were less patient centered than those with some college, who differed from those with at least a bachelor’s degree. Income differences were noted between those who reported $20,000 or less versus compared with those in the $40,000 to $80,000 range compared with those of $80,000 or more. Overall, white patients were more patient-centered than nonwhites; and although the numbers of Latinos, African Americans, Asians, and Native Americans were too small for meaningful statistical comparisons, the scores of African Americans were almost identical to those of the white patients, while Latinos’ and Asian/Pacific Islanders’ scores were somewhat lower and closer to one another. The cardiology patients were less patient centered than those who were being seen in internal medicine or family practice; however, these differences may be explained by the fact that the cardiology patients were significantly older than the other 2 patient groups (mean age = 64.2 years vs 56.2 years for internal medicine and 53.4 years for family practice; F=35.80, df=2, P <.001).