Longitudinal studies in patients with chronic kidney disease (CKD) often include both repeatedly measured longitudinal outcomes and clinical event data. These two types of outcome data have usually been analyzed separately in the medical literature. However, analyses focused singly on either the longitudinal or clinical event components of the data are constrained by important limitations, including a) loss of insight due to fragmented presentation of longitudinal and event outcomes, b) loss of statistical power due to incomplete use of the available data, c) increased risk of bias due to informative dropout, and d) logical contradictions in models used to analyze change over time periods with heavy attrition from ESRD and death. These issues have long complicated analyses of CKD cohorts, but are now assuming new importance due to the increasing availability of CKD cohorts with extended follow-up, including the 11-year African American Study of Kidney Disease and Hypertension (AASK), and the Chronic Renal Insufficiency Cohort (CRIC) Study. We propose to draw upon recent methodological advances in the joint analysis of longitudinal and event data and in the foundations of causal inference to address these limitations through three analytic aims: 1) To develop practical methods to relate exposures to the evolution of cohorts as characterized by changes in multi-state distributions defined by joint longitudinal/time-to-event outcomes;2) To develop practical methods to estimate subject-specific trajectories under flexible joint models and relate these trajectories to treatments, exposures, and clinical events;3) To apply the causal modeling framework of principal stratification to provide unbiased and conceptually coherent analyses of effects of exposures and treatments on a longitudinal outcome while accounting for attrition due to death and ESRD. Additionally, a central focus of this research, expressed in our fourth specific aim, is the dissemination of the methods developed in Aims 1-3 for routine use in analyses of CKD cohorts. This will be accomplished by publishing applications of joint modeling methods to the AASK and CRIC studies in nephrology journals, and by developing a publically available library in the R-statistical package with programs to implement the proposed methods.

Public Health Relevance

This application extends new advances in statistical methods to improve the ability of clinical researchers in nephrology to investigate effects of treatments and risk factors on patient outcomes. By better integrating analyses which are now usually done separately, researchers will be able to gain greater insights from longitudinal studies, especially those in which many patients die or reach end stage kidney disease.

Agency
National Institute of Health (NIH)
Institute
National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK)
Type
Research Project (R01)
Project #
5R01DK090046-04
Application #
8665412
Study Section
Special Emphasis Panel (ZDK1-GRB-D (O1))
Program Officer
Flessner, Michael Francis
Project Start
2011-06-20
Project End
2015-05-31
Budget Start
2014-06-01
Budget End
2015-05-31
Support Year
4
Fiscal Year
2014
Total Cost
$441,908
Indirect Cost
$42,148
Name
University of Utah
Department
Internal Medicine/Medicine
Type
Schools of Medicine
DUNS #
009095365
City
Salt Lake City
State
UT
Country
United States
Zip Code
84112
Grams, Morgan E; Li, Liang; Greene, Tom H et al. (2015) Estimating time to ESRD using kidney failure risk equations: results from the African American Study of Kidney Disease and Hypertension (AASK). Am J Kidney Dis 65:394-402
Li, Liang; Chang, Alexander; Rostand, Stephen G et al. (2014) A within-patient analysis for time-varying risk factors of CKD progression. J Am Soc Nephrol 25:606-13
Hu, Bo; Gadegbeku, Crystal; Lipkowitz, Michael S et al. (2012) Kidney function can improve in patients with hypertensive CKD. J Am Soc Nephrol 23:706-13
Li, Liang; Astor, Brad C; Lewis, Julia et al. (2012) Longitudinal progression trajectory of GFR among patients with CKD. Am J Kidney Dis 59:504-12
Joffe, Marshall (2011) Principal stratification and attribution prohibition: good ideas taken too far. Int J Biostat 7:Article 35