Javier Llorca1, Miguel Delgado-Rodríguez. 1. Division of Preventive Medicine and Public Health, School of Medicine, University of Cantabria, Avda. Cardenal Herrera Oria s/n, 39011 Santander, Spain. llorcaj@unican.es
Abstract
BACKGROUND AND OBJECTIVE: Survival-agreement plots have been suggested as a new graphical approach to assess agreement in quantitative variables. We propose that survival analytical techniques can complement this method, providing a new analytical insight for agreement. METHODS: Two survival-agreement plots are used to detect the bias between to measurements of the same variable. The presence of bias is tested with log-rank test, and its magnitude with Cox regression. RESULTS: An example on C-reactive protein determinations shows how survival analytical methods would be interpreted in the context of assessing agreement. CONCLUSION: Log-rank test, Cox regression, or other analytical methods could be used to assess agreement in quantitative variables; correct interpretations require good clinical sense.
BACKGROUND AND OBJECTIVE: Survival-agreement plots have been suggested as a new graphical approach to assess agreement in quantitative variables. We propose that survival analytical techniques can complement this method, providing a new analytical insight for agreement. METHODS: Two survival-agreement plots are used to detect the bias between to measurements of the same variable. The presence of bias is tested with log-rank test, and its magnitude with Cox regression. RESULTS: An example on C-reactive protein determinations shows how survival analytical methods would be interpreted in the context of assessing agreement. CONCLUSION: Log-rank test, Cox regression, or other analytical methods could be used to assess agreement in quantitative variables; correct interpretations require good clinical sense.
Authors: Alvaro Alonso; Juan José Beunza; Miguel Delgado-Rodríguez; Miguel Angel Martínez-González Journal: BMC Public Health Date: 2005-09-12 Impact factor: 3.295