Chronic kidney disease (CKD) imposes a substantial burden on the healthcare system and society. Mortality rate in patients with end-stage kidney disease is high but more patients with CKD die from cardiovascular disease (CVD) than from end-stage kidney disease. As the occurrence of CVD is associated with increased risk of CKD progression, prevention of CVD in patients with CKD is crucial. Despite the well-established relationship between CKD and CVD, there are currently no simulation models available to quantify the long-term impacts of early CVD intervention in patients with CKD on both health and economic outcomes. Since the evidence on the effectiveness of the interventions obtained from clinical trials are only available over a limited duration, assessing the full impact of an intervention requires the development of an individual-level simulation model which can integrate and extrapolate existing evidence. This study aims to develop an individual-level simulation model to estimate lifetime disease progression and project lifetime incidence of major fatal and non-fatal CVD events, dialysis, kidney transplant and mortality in patients with early, moderate and advanced CKD. We will use the Clinical Practice Research Datalink (CPRD), a population-based primary care database, and advanced modelling techniques to develop the risk equations to predict the CVD events, renal replacement therapies and mortality based on a wide range of risk factors. This model is envisaged to serve as a tool for clinical and economic evaluation of CVD preventive strategies in patients with CKD in many countries.
Chronic kidney disease (CKD) imposes a substantial burden on the healthcare system and society. Mortality in patients with end-stage kidney disease is high but more patients with CKD die from cardiovascular disease (CVD) than from end-stage kidney disease. Quantification of the long-term economic and health impact of CVD prevention in patients with CKD is crucial to inform decisions in healthcare budget planning and to maximise treatment efficiency. This study aims to develop an individual-level simulation model to estimate lifetime disease progression and project lifetime incidence of major fatal and non-fatal CVD events, dialysis, kidney transplant and mortality in patients with CKD. We will use the Clinical Practice Research Datalink (CPRD) and the linked hospital and death records for model development. Parametric proportional hazards models will be used to develop equations for risks of CVD and renal replacement events. Progression of risk factors for CKD and CVD will be modelled using linear mixed-effects models. The estimated risk and risk factor progression equations will be integrated into a discrete-time simulation model with annual cycles. The simulation will involve the use of risk factor progression equations to predict changes in risk factor levels and the use of risk equations to estimate the probabilities of the events, which will be compared with random numbers to determine whether the events occur. During the simulation, changes in risk factor values and event histories will be updated in each cycle and their values at the beginning of the cycle used to predict occurrence of events in that cycle as well as risk factor values in the next cycle. Model outputs will be annual incidence of CVD events, dialysis, kidney transplant and death, and annual values of risk factors over the simulation time. Uncertainty in the model outcomes will be handled using Monte-Carlo simulation and bootstrapping.
Occurrence of major cardiovascular (CV) events and dialysis, kidney transplant, and mortality in patients with chronic kidney disease. The following CV events will be included in the simulation model: myocardial infarction (MI), ischemic stroke, haemorrhagic stroke, heart failure, percutaneous coronary intervention (PCI), coronary artery bypass grafting (CABG), peripheral vascular disease and hospitalisation due to angina. These events will be identified in the CPRD GOLD using the medical codes and in the Hospital Episode Statistics (HES) Admitted Patient Care (APC) using the ICD-10 and OPCS codes (see the enclosed appendices for the codes of these events).
An Tran-Duy - Chief Investigator - University of Melbourne
An Tran-Duy - Corresponding Applicant - University of Melbourne
Craig Nelson - Collaborator - Western Health
Dennis LA - Collaborator - University of Melbourne
Jo-Anne Manski-Nankervis - Collaborator - University of Melbourne
Paul Amores - Collaborator - University of Melbourne
Philip Clarke - Collaborator - University of Oxford
Ting Zhao - Collaborator - University of Melbourne
HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation