Re-evaluating the clinical and cost effectiveness of implantable cardioverter defibrillator (ICD) or cardiac resynchronisation therapy with defibrillation (CRT-D) for heart failure using real world data from the Clinical Practice Research Datalink.

Date of Approval: 
2020-11-19 00:00:00
Lay Summary: 
About 2% of the National Health Service (NHS) budget is used annually to care for patients with heart failure (HF), a condition in which the heart does not pump properly leading to breathlessness, fatigue, fluid retention and death. Patients with HF are initially treated with drugs such as angiotensin-converting-enzyme inhibitors, beta-blockers and diuretics to improve heart function and relieve symptoms. However, drug treatment may become less effective as heart function worsens. Clinical studies known as randomised controlled trials (RCTs) have shown that medical devices that improve heart function such as electronic implantable cardioverter defibrillator or cardiac resynchronisation therapy (ICD/CRT-D) are clinical and cost-effective treatments for HF. It is however unclear whether the benefits from these therapies are also realised in everyday clinical practice. This study will compare treatment with ICD/CRT-D and HF drug therapy alone to see if this decreases the risk of death, hospitalisation, complications, use of health services and associated costs among patients with HF. We will use a range of research methods to answer these questions to overcome some of the challenges of using data collected from routine clinical practice. These will provide more accurate estimates of how effective these alternative treatments are for individual patients. We will use our results in economic analysis to re-assess whether ICD/CRT-D is indeed a cost-effective treatment for HF. These results will provide valuable insights on the extent to which clinical data from everyday practice could be used to inform policy decisions about the introduction of HF treatment in the NHS
Technical Summary: 
Current recommendations on ICD/CRT-D for treating heart failure in the English NHS are largely based on clinical and cost effectiveness evidence from RCTs. However, RCTs may not be representative of HF patients treated in routine clinical practice or reflect real world health service delivery. We will assess the comparative effectiveness of outcomes among patients with HF treated with ICD/CRT-D versus HF drug therapy alone to inform decision analysis and stochastic cost effectiveness models. Patients 18 years and older with incident heart failure recorded in primary care or hospital during 01/01/2008-31/12/2018 will be extracted from the Clinical Practice Research Datalink. Index date will be defined as the earliest date of HF drug therapy following HF diagnosis. Patients will be followed at 3,6,12,24 months up to 5 years following index date for all-cause mortality (primary outcome), HF or cardiovascular death, any-cause hospitalisation, complication rates, use of health care services and associated costs. Treatment adherence will be explored as a driver of outcomes. Instrumental variable, propensity score matched analysis and conventional statistical methods will be used to assess relative treatment effectiveness. Survival models such as Cox proportional hazard or other flexible models will be used to estimate time to mortality; generalised linear models such as Poisson regression models will be used to assess rates of hospitalisation and complications. Lifetime costs will be estimated using a flexible two-part model employing logistic regression and generalised linear models with a gamma distribution. The cost per quality-adjusted-life year gained and incremental net health benefit over a lifetime horizon, from the NHS perspective, will be estimated in cost effectiveness analysis. Results from economic models will provide insights on the extent to which clinical data from everyday practice could be used to inform policy decisions about the introduction of cost-effective HF treatment in the NHS.
Health Outcomes to be Measured: 
- HF treatment patterns (prescription rate, combination therapy, optimal therapy) - HF treatment adherence - HF complication rates - Health care resource use and cost (any cause, and HF-related) - Hospitalisations (any cause, and HF-related) - Mortality (All cause, Cardiovascular- related, HF-related) - Combined end point of hospitalisation or mortality
Application Number: 

Puja Myles - Chief Investigator - CPRD
Tarita Murray-Thomas - Corresponding Applicant - CPRD
- Collaborator -
Alex Bottle - Collaborator - Imperial College London
Reza Skandari - Collaborator - Imperial College London

HES Accident and Emergency;HES Admitted Patient Care;HES Outpatient;ONS Death Registration Data;Patient Level Index of Multiple Deprivation