Electronic Health Record Research using DExtER

Add to your list(s) Download to your calendar using vCal

• Professor Krish Nirantharakumar, Institute of Applied Health Research
• Wednesday 03 November 2021, 12:00-13:00
• Haworth Building Room 320A and Online.

If you have a question about this talk, please contact Jordan McCormick.

The NHS provides healthcare to 66 million people, with over a million people utilising NHS services every day. Each of these encounters is an opportunity to learn, but this data is often poorly accessible and not in research-ready formats. By bringing together experts in epidemiology, data science, and software engineering, we created DExtER (PMID: 32856160), an automated system for efficient, transparent and reproducible research. Studies which used to take months can now be undertaken within days. For example, DExtER enabled analysis of >70,000 serum testosterone measurements in young women, identifying increased risk of non-alcoholic fatty liver disease and diabetes in women with androgen excess (PMID:29590099). This led to a major new experimental medicine study (DAISy-PCOS) stratifying risk and developing novel treatments to prevent metabolic complications in women with polycystic ovary syndrome, a lifelong condition affecting 1 in 10 women worldwide.

The tool also enables representative recruitment for mechanistic studies by rapidly generating eligible participants based on inclusion and exclusion criteria by searching through millions of electronic patient records within minutes. This is supporting the NIHR Therapies for Long COVID study, which will recruit thousands of diverse participants to help characterise the symptoms, health impacts, and underlying causes of Long COVID syndromes in non-hospitalised patients, providing invaluable insight and co-producing targeted interventions with those patients, tailored to individual need. It is also underpinning data-driven clinical trials, such as RADIANT , testing new ways of running trials at NHS General Practices. Such trials use already-collected NHS health information to reduce the time taken for research, both for patients and NHS staff, with both recruitment and outcomes driven electronically. This allows new and larger groups of patients to benefit from new treatments, opening the possibility of research in populations previously ignored.

Importantly, DExtER also supports better clinical decision-making. Publications arising from DExtER have supported clinical decision-making across varied health conditions including thyroid disease, identifying that concentrations of thyroid-stimulating hormone outside normal range contribute to adverse health outcomes (PMID:31481394). Furthermore, in a UK-first study we have demonstrated that DExtER can answer clinical questions that arise during patient consultations within hours: termed an ‘informatics consult’, a tailored approach particularly important in settings where medication for one disease might have adverse impacts on co-morbidities. This has informed a large grant to create an AI tool (OPTIMaL) to provide individualised therapies for people with multimorbidity.

This talk is part of the CCB seminars series.

This talk is included in these lists:

• Haworth Building Room 320A and Online

Note that ex-directory lists are not shown.