Trends and determinants of survival for patients on renal replacement therapy over time.

Summary
End Stage Renal Disease (ESRD) is the severe end of the spectrum of Chronic Kidney Disease, requiring some form of renal replacement therapy (RRT) to stay alive.

This work will utilize, and extend when required, advanced statistical methods for the analysis of longitudinal data collected from the UK Renal Registry. These will include multistate modelling to describe how ESRD and its management in RRT patients evolve through different stages over the patient’s lifetime, including transitions between different treatment modalities, the occurrence of other health events, hospitalisation, recovery of kidney function, or death.

Further, graphical Markov models will be utilized to build causal diagrams to describe relationships among patient factors and treatment modalities underlying longer survival time.

Background
RRT is an expensive treatment, consuming up to 2% of the NHS budget for a population of around 50,000 patients. Furthermore, the number of patients requiring RRT is growing steadily with no signs of this rise stopping.

RRT options consist of peritoneal dialysis (PD), haemodialysis (HD) or renal transplantation. Unfortunately, although transplantation is generally a better option for mortality and morbidity, patients are frequently either too frail to receive a renal transplant or have to wait a considerable period of time before receiving a transplant. Particularly for younger patients, this means that they may well transition between numerous different RRT modalities over the course of their lifetime.

Dialysis treatment remains problematic, with a 5-year survival frequently worse than many cancer diagnoses but there have been significant changes in clinical management and technology for these treatments over many years. How much these changes have impacted on survival remains unclear, with studies of renal patients, and dialysis patients in particular, frequently struggling to attract the funding required for the large randomized controlled trials necessary to determine effects upon patient survival.

In this scenario, it is particularly important to derive maximal information from observational data. In the UK all RRT patients are recorded in the national UK Renal Registry, which has been recording data since 1997 and has complete UK data from 2007. Survival analysis adjusting for age suggests that survival in these patients has been improving but with different trends over time for HD and PD patients. Relative survival analysis, effectively matching for changes in the general population survival rates, suggests that the improvements in survival for dialysis patients is primarily in the younger population, apparently contradicting the survival analysis adjusting for age. Furthermore, both of these analyses fail to account for changing rates of transplantation and the survival of the transplant function, as well as differences over time in the utilisation of the different dialysis modalities.

Research Aims
1 – How have transitions between different modalities of RRT changed over time?
2 – How has the survival of RRT patients on different modalities changed over time, taking into account changes in transitions between different modalities?
3 – What are the determinants of the changes with time demonstrated in the first 2 aims?
4 – How has the survival of RRT patients changed over time when compared to the general population?
5 – To determine the value of multistate modelling relative to the widely used Cox model in renal registry reporting

The statistical modelling will be guided by strong clinical knowledge and experience from experts in the field of nephrology at iACS so that clinical insight from the analyses can be demonstrated and communicated meaningfully. The student will also benefit from the methodological expertise provided by the iPCHS.