Background: The statistical measure used to quantify the degree of agreement or congruence between two research subjects has been the intraclass or the Pearson correlation coefficient; however, the concordance correlation coefficient (CCC) is another measure of agreement used to examine agreement between two observers or raters.
Objectives: The aims of this study were to (a) highlight the differences among three statistical measures used to quantify the degree of agreement or congruence, (b) demonstrate the use of the CCC in examining agreement between heart failure (HF) patients and their family members, and (c) provide nurse researchers another method for evaluating agreement.
Methods: Symptom evaluation scores obtained from HF patients and their family members were used in the analysis of this study. To explain the use of the CCC in this analysis, a distinction between Pearson correlation coefficient and intraclass correlation coefficient is discussed. The CCC calculation is then described in detail.
Results: The HF patients in this sample were 71 ± 9.6 years in age, 40% male, and 41.4% African American. Most (75%) family members were female. There were several different categories of family members, but most were spouses. The CCC results indicated that no symptom achieved good agreement, and 8 of 14 symptoms were in moderate agreement (.4 ≤ CCC ≤ .7) within the dyads. Of the six symptoms with poor agreement (0≤ CCC < .4), HF patients and their family members agreed least on worsening cough (CCC = .152, 95% confidence interval = −.134 to .413) and bloated abdomen (CCC = .055, 95% confidence interval = −.224 to .325).
Discussion: Applying the CCC to dyadic data from HF patients and family members, symptoms in which the patient and family member had the most and least agreement were identified. The six symptoms with poor agreement were symptoms that can show HF decline and may be important when examining future nursing interventions. Further study is needed using the CCC with dyadic data along with other family factors that influence agreement.
Christina Quinn, DNS, RN, is Associate Professor, Gordon College, Barnesville, Georgia.
Michael J. Haber, PhD, is Professor; and YiPan, MS, is Graduate Student, Department of Biostatistics, Rollins School of Public Health, Emory University, Atlanta, Georgia.
Editor's Note Materials documenting the review process for this article are posted at http://www.nursing-research-editor.com
Accepted for publication May 14, 2009.
This study was funded by the Center for the Study of Symptoms, Symptom Interactions, and Health Outcomes NIH NINR P20 NR007798 and Postdoctoral Training Grant from NIH NINR F32 NR009888-01A1 2007-2008.
Corresponding author: Christina Quinn, DNS, RN, Division of Nursing and Health Sciences, Gordon College, 419 College Drive, Barnesville, GA 30204 (e-mail: email@example.com).