King's Technology Evaluation Centre (KiTEC), a United Kingdom- based health technology assessment consultancy, was tasked with identifying a specific group of heart failure patients who had repeat readmissions in order to accurately identify the potential market for an innovative device designed to diagnose heart failure as a way to avoid costly and avoidable hospital readmissions. The device enables clinicians to remotely diagnose heart failure and appropriate medication can be administered instead of a hospital visit. Our methodology describes an accurate way to quantify the at risk population without the need for a costly trial.

Using big data from national registries – the heart failure specific National Institute for Cardiovascular Outcomes Research (NICOR) database and the national Hospital Episodes Statistics for the National Health Service (HES) – KiTEC has devised a methodology of linking the two datasets in order to (i) accurately identify patients with repeat readmissions over a 5-year period and (ii) calculate the risk factors for readmissions. Data is linked using a common field, meaning information from both databases can be analyzed at patient level (it is pseudo-anonymized before KiTEC receives it). This allows for unprecedented granularity, as we are able to exploit the heart failure specific detail of NICOR alongside the wealth of admissions data available in HES.

There are significant challenges surrounding the use of registry data, especially in the enormous size of the datasets and in privacy legislation aimed at protecting personally identifying data. The usual regulatory approvals for health research are also more complex when linked datasets are proposed. These are important considerations, especially when linking two complementary databases.

The use of real world data has the potential to paint a true and accurate picture of a patient population, while avoiding many of the biases inherent in typically research studies. However, there are other important challenges to overcome, namely difficulties analyzing huge datasets and navigating complex legislation to access patient data.