The Relationship Between Two Performance Scales: New York Heart Association Classification and Karnofsky Performance Status Scale

Abstract

ContextPerformance status is used to quantify the well-being and functional status of people with illness. Clinicians and researchers from differing fields may not instinctively understand the scales, typically disease specific, used in other disciplines. ObjectivesTo provide a preliminary description of the relationship between the Karnofsky Performance Status Scale (KPS) and the New York Heart Association Classification (NYHA) and to stimulate discussion in research and clinical practice. MethodsSimultaneous KPS and NYHA data (172 observations) from three studies of people with chronic heart failure were pooled. Linear regression was used to predict the mean KPS from NYHA. The strength of association between the scales was investigated using a Kendall's Tau-b correlation coefficient. The agreement between the predicted and observed KPS scores was investigated using weighted kappa with quadratic weights. ResultsLinear regression demonstrated a relationship between KPS and NYHA (P < 0.0001; R2 = 0.3). Predicted KPS from NYHA class rounded to the nearest 10 gave the following values: Class I, predicted KPS 90%; Class II, predicted KPS 80%; Class III, predicted KPS 70%; and Class IV, predicted KPS 60%. A moderate strength of association between KPS and NYHA (Kendall's Tau-b correlation coefficient of −0.49; P < 0.0001) and agreement between observed and predicted KPS (kappa coefficient = 0.52) was shown. ConclusionWe suggest that the NYHA discriminates poorly between clinically important performance states in people with advanced disease (NYHA III and IV; KPS <50%). The KPS, used in conjunction, would provide useful additional information in research and clinical practice.