Observer error in grading performance status in cancer patients

Abstract

To assess the Karnofsky Performance Status (KPS) scale and the Eastern Cooperative Oncology Group (ECOG) scale with respect to interobserver reliability and interobserver difference between the two scales, 100 consecutive patients in a medical oncology unit were assessed using both scales, which were then given to three independent raters and also to the patients themselves. There was a high level of agreement between most paired assessors. There were three exceptions, which were the RMO/patient and nurse/patient pairs on the KPS scale and the RMO/patient pair on the ECOG scale. The level of agreement was better on the ECOG scale. For individual raters there is no statistical difference between the ECOG and Karnofsky scales. There was good agreement between all raters for both scales. Interobserver variability was less using the ECOG scale. We conclude that either scale could be used with good interobserver reliability. The ECOG scale minimises differences between observers.