BACKGROUND: Although many studies of heart failure therapies test improvements of patient condition in terms of mean changes of quality of life (QoL) or exercise tolerance (ET) measures, it is of increasing interest to quantify the proportion of patients that "respond" to therapy and understand factors predicting response. These questions can be address through the use of a "responder analysis," in which the proportion of patients in whom a measure of QoL or ET improves by a minimum amount is determined. Here, we review the principles of a "responder analysis." METHODS AND RESULTS: We used data from published studies of cardiac resynchronization therapy to model the results of a responder analysis and original data from a recent study of cardiac contractility modulation to illustrate the many facets of such an analysis that need to be understood and investigated further. Some of these areas include: understanding how to choose criteria for response; how to deal with differing results obtained with different measures of response; and how to deal with potentially conflicting information provided by a responder analysis and the more standard comparison of mean changes. CONCLUSIONS: Additional prospective studies will help advance understanding the optimal way to use responder analyses in heart failure trials.
BACKGROUND: Although many studies of heart failure therapies test improvements of patient condition in terms of mean changes of quality of life (QoL) or exercise tolerance (ET) measures, it is of increasing interest to quantify the proportion of patients that "respond" to therapy and understand factors predicting response. These questions can be address through the use of a "responder analysis," in which the proportion of patients in whom a measure of QoL or ET improves by a minimum amount is determined. Here, we review the principles of a "responder analysis." METHODS AND RESULTS: We used data from published studies of cardiac resynchronization therapy to model the results of a responder analysis and original data from a recent study of cardiac contractility modulation to illustrate the many facets of such an analysis that need to be understood and investigated further. Some of these areas include: understanding how to choose criteria for response; how to deal with differing results obtained with different measures of response; and how to deal with potentially conflicting information provided by a responder analysis and the more standard comparison of mean changes. CONCLUSIONS: Additional prospective studies will help advance understanding the optimal way to use responder analyses in heart failure trials.
Authors: William T Abraham; JoAnn Lindenfeld; Vivek Y Reddy; Gerd Hasenfuss; Karl-Heinz Kuck; John Boscardin; Robert Gibbons; Daniel Burkhoff Journal: J Card Fail Date: 2014-10-05 Impact factor: 5.712