REDDIE RWD Dashboard

Jonah Thomas is a Research Associate at the University of Leicester. He is a part of the team leading the tasks related to Work Package 2 - Identifying diabetes databases and metadata contained in them. In this REDDIE Insight, he explains the tasks within the work package.

Work Package 2 (WP2) of the REDDIE project focuses on real-world data (RWD) databases. RWD has previously been defined as routinely collected health datasets (such as electronic medical records, wearable device data, and prescription claims databases). It is known that there are many sources of RWD available in the field of diabetes. The overall aim of this work package is to identify and evaluate different sources of RWD which contain data relevant to diabetes therapies and outcomes. This will help researchers fully utilise the data in these databases.

Identifying relevant databases The first step of WP2 was to identify RWD databases in the field of diabetes and gain a better understanding of what variables they have collected. To achieve this, we have completed a systematic review where we searched previously published studies to identify the databases which exist around the world. I will now briefly cover what we found.

Describing the data contained in the databases

We identified over 13,000 scientific publications which utilised RWD databases to answer questions about diabetes. Within these studies, we found 515 unique databases. Most of these databases were collected in North America, Europe, and Asia. Data on over 382 million individuals was included in the databases identified (one individual may be included in multiple databases and therefore be counted twice). Prescription data was available in 431 datasets, diagnosis data in 382 datasets, and biochemistry data (cholesterol, blood glucose) in 348 datasets. Mortality data was less commonly collected (165 datasets) as was the use of technology (continuous glucose monitors, insulin pumps) in only 31 datasets. Of the datasets identified, 257 focused on individuals living with type 2 diabetes, 19 on individuals living with type 1 diabetes, and 236 on either type of diabetes. We have created an online dashboard to help people navigate through the different data contained in different datasets e. The dashboard can be accessed via the REDDIE website: https://www.reddie-diabetes.eu/research/rwd-dashboard or directly: https://jjct.quarto.pub/real-world-databases-in-diabetes/.

Developing a tool to identify databases that are fit for purpose

Each research study answers a novel question, and a database that may be right for one research question may not be right for another. We aim to connect with stakeholders and develop a fit-for-purpose assessment so researchers can understand which databases may be most suited to answer a given research question.