Cover: How Many Patients Are Needed to Provide Reliable Evaluations of Individual Clinicians?

How Many Patients Are Needed to Provide Reliable Evaluations of Individual Clinicians?

Published in: Medical Care, v. 42, no. 3, Mar. 2004, p. 259-266

by Eugene C. Nelson, Mary A. Gentry, Kathryn H. Mook, Karen Spritzer, John H. Higgins, Ron D. Hays

Read More

Access further information on this document at www.lww-medicalcare.com

This study was published in a peer-reviewed scholarly journal. The full text of the study can be found at the link above.

Abstract

PURPOSE: The purpose of this study was to determine how many patients are needed to provide reliable patient ratings of care at the individual clinician level. SETTING AND SOURCES OF DATA: The study was conducted in an academic medical center and was based on analysis of 34,985 patients who completed a 50-item survey rating the care received during a recent outpatient visit to a physician or midlevel provider. STUDY DESIGN: Analyses of patient satisfaction surveys was done to: 1) confirm the dimensions of satisfaction with outpatient care in an existing measure, and 2) determine the number of patients required to provide reliable estimates of clinician care for single items and an 11-item composite scale. PRINCIPAL FINDINGS: Factor analysis showed that the survey measured 2 dimensions of satisfaction: 1) clinician care, and 2) features of visiting the office. The 11-item clinician care scale had high reliability (Cronbach's alpha = 0.97). The number of patients needed to achieve reliability of 0.80 at the clinician level was 66 for the 11-item scale and ranged from 52 to 91 for individual items. For primary care physicians only, the comparable number of patients per clinician was 77 for the 11-item scale and ranged from 50 to 147 across items. CONCLUSIONS: For the survey items that the authors analyzed, the answer to the question How many patients are needed to obtain useful and reliable feedback? is at least 50, but varies by item type (global vs. specific) and by number of items (composite scale or single-item rating) and by the conditions of use (for self-assessment and learning or reward and punishment).

Research conducted by

This report is part of the RAND Corporation external publication series. Many RAND studies are published in peer-reviewed scholarly journals, as chapters in commercial books, or as documents published by other organizations.

The RAND Corporation is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.