Ozgur Asar, PhD student
Longitudinal and survival statistical methods with applications in renal medicine
- Quantifying the relationships between the underlying kidney function and hazards for renal replacement therapy and death
- Early detection of incipient renal failure amongst primary care patients
- Investigation of the short and long term effects of acute kidney injury on kidney function amongst chronic kidney disease patients
- Developing open-source statistical software for the developed methodologies
Funded by: HeRC
Chronic kidney disease, acute kidney injury, incipient renal failure
- The Chronic Renal Insufficiency Standards Implementation Study
- Salford Integrated Records
- Longitudinal and survival statistical methods
Benefits and outcomes
The associations between the underlying kidney function and the hazards for renal replacement therapy and death amongst chronic kidney disease patients were estimated unbiasedly. This would help the nephrologists to better manage their patients using the repeated measurements of kidney function biomarkers.
A predictive model is developed to detect progression towards end-stage renal failure using data from primary care patients. This will help the primary care pyscians in terms of sucessfullly referring a primary care patient with high risk of renal failure to secondary care. In our application, a number of patients who might well be referred to secondary care but not referred were identified. A possible next step of this project is to develop the real-time survelliance system that would be used routinely.
Epidemiology of acute kidney injury, its short and long term effects were described. We found that although acute kidney injury is a sudden event, it might be preceded by an accelerated loss of kidney function and patients have a recovery phase on average. In long-term, acute kidney injury was found to be associated with an increased loss of kidney function, specifically in our applicatio the loss was approximately doubled after an acute kidney injury occurred. To the best of our knowledge, this study is the first one investigating the subject by advanced statistical modelling. It would help the nephrologists to better manage their patients if they show signs of acute kidney injury and to be more cautious regarding the electronic records of acute kidney injury events, which are currently not well recorded.
Open source software was developed. It has been used, and we believe will continue to be used, by many applied scientists. The benefit of this project is that it enables applied scientists analysing their data using a class of advanced statistical methodologies which is freely available. The methodologies our software implements were not available in a software format before, either freely or paid.
Supervisor: Prof. Peter Diggle, Lancaster University
Medical Collaborators: Prof. Philip Kalra and Dr. James Ritchie, Salford Royal Hospital