Extrapolation of event rates

Limited follow-up in clinical trials is a major challenge to economic evaluation where a lifetime perspective is often required. Observational data is commonly available, at least for one of the comparators, and offers valuable additional information on event rates. The use of observational data to assess the plausibility of event rates extrapolated from trial data or to directly inform extrapolations has been recommended but limited guidance is available to translate this advice into practice. Our research examines how additional data, typically from large observational datasets is best exploited when extrapolating event rates. Recent analysis has drawn on data from the National Joint Register of England, Wales and Northern Ireland and Hospital Episode Statistics to extrapolate the failure rates of artificial hip and knee prostheses for comparative analysis.

Further work within this theme will utilise the PULSE database held by the National Health Service Blood & Transplant (NHSBT) to examine the impact of the interval between donation requests on donation rates and retention of blood donors. The results will inform a cost-effectiveness analysis of alternative methods to increase the supply of donated blood (the HEMO study).

Key people: Mark Pennington, Zia Sadique, Richard Grieve

Comments are closed.