Joint Modeling of Longitudinal and Time-To-Event Data: GFR and Hospitalization

Studies that collect longitudinal data and time-to-event (survival / failure) data can analyze each separately if there is no association between the longitudinal data and the time-to-event data. However, if there is an association, then a joint analysis is necessary to avoid biased results. This study addresses whether there is a relationship between GFR measurements collected over time in CKD patients, and time to first hospitalization following a GFR value less than 60. Separate analyses are done for change in GFR over time and time to hospitalization followed by a joint analysis to assess whether there is a statistically significant relationship between the two.
Dr. Kolm is Director of Biostatistics at MedStar Cardiovascular Research Network, MedStar Heart and Vascular Institution. He has over 30 years of experience in consulting with principal investigators in the design and analysis of clinical trials, retrospective and observational studies, and large patient registries. Dr. Kolm has a wide range of statistical expertise including hierarchical general and generalized linear models, classification and tree regression, time-to-event analysis, cost-effectiveness analysis and multiple imputation methods for missing data.

[@jonastverring6925]

Thank you for a very interesting and clear example on how to use jm! Is it possible to get access to your stata code? I am especially interested in how you went from the jm model to the vizualisation of the results on time, GFR and hazard? Regards, Jonas Tverring from Sweden