Volume 43 | Number 5p1 | October 2008

Abstract List

Sarah L. Clever, Lei Jin, Wendy Levinson, David O. Meltzer


To determine the relationship between physicians' communication behaviors and patients' overall satisfaction with hospital care using a novel instrumental variable to address possible confounding of this association by patient attributes.

Data Sources/Study Setting

Administrative records and postdischarge survey data were obtained from patients discharged from the General Medicine service at an urban tertiary‐care academic hospital between July 1, 1997 and June 30, 2000. Administrative data included comorbidities, demographic data, and payer status. In the discharge survey, patients rated their attending physician on four communication behaviors, other aspects of their hospital stay, and their overall hospital care.

Study Design

The primary outcome was patients' ratings of their overall satisfaction with hospital care, and the primary independent variable was patients' ratings of their physicians' communication behaviors. To remove possible confounding of the association between patient ratings of physician communication and overall satisfaction by other patient‐specific attributes, we created an instrumental variable (IV) in a two‐stage linear regression. The IV was the mean of the communication ratings given to each physician by the other patients cared for by that physician.

Principle Findings/Conclusions

Three thousand one hundred and twenty‐three patients were included in the analysis. In the ordinary least squares regression, there was a significant positive relationship between overall satisfaction and overall ratings of attendings' communication behaviors, with an increase in overall satisfaction of 0.58 points on a 5‐point scale for each 1‐point increase in overall attendings' communication behaviors, <.001. This relationship was maintained but attenuated in the IV regression, with a coefficient of 0.40, =.046. Although we find that the relationship between patient communication ratings and overall patient satisfaction may be confounded by patient‐level factors, we nevertheless continue to find evidence of a statistically significant and sizable relationship between physicians' communication behaviors and overall patient satisfaction after controlling for such factors.