Mikkel Brabrand1, Susanne Hosbond, Lars Folkestad. 1. aDepartment of Medicine, Sydvestjysk Sygehus, Esbjerg bDepartment of Cardiology, Odense University Hospital, Odense, Denmark.
Abstract
OBJECTIVES: The interobserver variability of capillary refill time (CRT) has been questioned. Earlier studies of interobserver variability of CRT have been on a large number of patients but with few observers. The objective of our study was to investigate how a large group of nurses and nurse assistants would grade CRT. METHODS: We recorded a video of the index finger of six medical patients and these were shown to nurses and nurse assistants. They were asked to record the CRT and whether they found this value to be normal. The data were analyzed using the Fleiss Kappa Coefficient Analysis and graded according to the Landis and Koch correlation. Correlation between the exact numbers was evaluated using interclass correlation. RESULTS: Nine nurse assistants and 37 nurses participated. The patients were aged between 44 and 87 years. All but one patient had a systolic blood pressure reading above 130 mmHg. All had arterial blood oxygen saturation above 92% and all but one had normal body temperature. The κ value for normality was 0.56. The interclass correlation of measurement of CRT was 0.62. CONCLUSION: This is the largest interobserver study of CRT when looking at the number of observers. We found an only moderate agreement for the exact value of CRT and a moderate agreement for normality. We believe that CRT should be used with caution in clinical practice.
OBJECTIVES: The interobserver variability of capillary refill time (CRT) has been questioned. Earlier studies of interobserver variability of CRT have been on a large number of patients but with few observers. The objective of our study was to investigate how a large group of nurses and nurse assistants would grade CRT. METHODS: We recorded a video of the index finger of six medical patients and these were shown to nurses and nurse assistants. They were asked to record the CRT and whether they found this value to be normal. The data were analyzed using the Fleiss Kappa Coefficient Analysis and graded according to the Landis and Koch correlation. Correlation between the exact numbers was evaluated using interclass correlation. RESULTS: Nine nurse assistants and 37 nurses participated. The patients were aged between 44 and 87 years. All but one patient had a systolic blood pressure reading above 130 mmHg. All had arterial blood oxygen saturation above 92% and all but one had normal body temperature. The κ value for normality was 0.56. The interclass correlation of measurement of CRT was 0.62. CONCLUSION: This is the largest interobserver study of CRT when looking at the number of observers. We found an only moderate agreement for the exact value of CRT and a moderate agreement for normality. We believe that CRT should be used with caution in clinical practice.
Authors: Barbara Lara; Luis Enberg; Marcos Ortega; Paula Leon; Cristobal Kripper; Pablo Aguilera; Eduardo Kattan; Ricardo Castro; Jan Bakker; Glenn Hernandez Journal: PLoS One Date: 2017-11-27 Impact factor: 3.240