RT - Journal Article T1 - Joint modelling of longitudinal and time-to-event data and its application in acute kidney injury JF - RJMS YR - 2015 JO - RJMS VO - 22 IS - 132 UR - http://rjms.iums.ac.ir/article-1-3884-en.html SP - 46 EP - 55 K1 - Joint modelling survival and longitudinal data K1 - Coronary artery bypass grafting K1 - Acute kidney injury AB - Background: In many clinical trials and medical studies, the survival and longitudinal data are collected simultaneously. When these two outcomes are measured from each subject and the survival variable depends on a longitudinal biomarker, using joint modelling of survival and longitudinal outcomes is a proper choice for analyzing the available data. Methods: In this retrospective archival study, 84 patients with coronary artery bypass surgery admitted in the intensive care unit of Jamaran Hospital in Tehran were randomly selected during 2001-2008. We defined the survival event as the 25% decline in GFR from baseline and the repeated measurements of urinary output from ICU admission to time of event as the longitudinal biomarker. Results: The study showed that older age (HR=1.112), male sex (HR= 4.307), and number of bypassed grafts (HR=1.874) were significant effective factors on reducing the amount of GFR and risk stage of acute kidney injury event. Conclusion: In this study, it was concluded that decreasing urinary output can be considered as an informative medical biomarker for acute kidney injury. Moreover, joint modeling of longitudinal and survival data which considers the relationship between these two outcomes, is an efficient approach for analyzing these kinds of datasets. LA eng UL http://rjms.iums.ac.ir/article-1-3884-en.html M3 ER -