Clinical Natural Language Processing Research Group

Whole population automated reading of brain imaging reports in linked electronic health records (WARBLER)

We will analyse anonymised data from brain scan reports to study the effect of brain appearance on future health.

Brain scans could give important information about future health by measuring brain appearances. Scans might show brains that are smaller (‘atrophied’), brains with more damage to the small blood vessels, or brains that show damage from old strokes. These appearances might indicate a greater chance of new strokes, heart attacks, dementia or early death.

To be certain about how important brain scan appearances are for future health, we need to know the brain scan appearances of a very large number of people whose health has been followed for a long time. Studies like this are very difficult to perform, because brain scanning is expensive, and patients are frequently lost to follow up.

NHS Scotland holds a large amount of information about the appearances of brain scans that have been performed as part of routine clinical care, and about the future health of these patients. Therefore these provide an excellent opportunity to study the effect of brain appearance on future health at a low cost, with excellent follow up, and of great relevance to the patients in Scotland. We will analyse anonymised data from brain scan reports with an experienced team of clinicians, radiologists, epidemiologists and statisticians at University of Edinburgh’s Centre for Clinical Brain Sciences, School of Informatics and Usher Institute.

Website

WARBLER project details on  NHS Health Research Authority website

Research team

All group members.

Funder

Chief Scientist Office, Medical Research Council, Turing