Big Data Computing in Healthcare: Opportunities, Obstacles, and Outputs

At every healthcare encounter in the UK, data are collected and stored in electronic healthcare records.  There is potential to link data from general practice records with hospital records (outpatients, accident and emergency and inpatient hospital stays) and with other sources, such as cancer and death registrations.  We have access to thousands of observations on millions of people, over decades of years.

This raises the opportunity to harness the power and granularity of these data for medical research and ultimately to improve clinical care and health outcomes.  Use of these resources have altered the scope, and the speed, of the epidemiology that may be conducted.  Novel analytics, including the application of artificial intelligence are increasing.  The combination of the availability of data and the techniques are rapidly advancing our knowledge.

However, the use of these data also presents some obstacles.  We must remember that this information is not collected for the purposes of medical research. It is not collected in a standardised way, is not available at regular time intervals, nor from everyone and is largely based on coded medical information.  Wrangling these records into an analysable form is a specialist task.

Furthermore, the Government’s initiative to introduce a daily download of our medical data (which is separate from the current research data available) may lead to increasing opt-outs.  Other considerations include the sheer volume of these data, leading to computational and statistical implications.

This talk will be illustrated with examples of the research conducted using routinely collected healthcare data, including studies of rare diseases, understanding disease risk factors, clinical prediction rules, and monitoring of drug and vaccine effectiveness and side effects.  Mention will also be made about how research findings have been used and integrated into clinical care.

Clare Bankhead is an Associate Professor in the Nuffield Department of Primary Care Health Sciences, University of Oxford and a Fellow of Kellogg College She is a senior epidemiologist with a special interest in research design and statistics.   Her main research focuses are on the use of large routine databases and digital health in medical research.  Research areas include health service utilization, diagnosis and monitoring of chronic diseases, and cancer diagnosis.

After 2 years for running Branch meetings via Zoom, our next event, on 14 April, “Big Data Computing in Healthcare” will be held both “in person” and “via Zoom”. (If this is successful, and virus cases do not escalate, we hope that subsequent meetings will also be of this hybrid type.)

“In Person” Logistics

The “in person” meeting will be held at our usual meeting place: OeRC, University of Oxford, 6 Keble Road, OX1 3QG. The meeting will commence at 7:30pm, and light refreshments will be available beforehand. Please register for the event as usual using the Eventbrite link on the website or via MeetUp, where you will be able to indicate whether you are attending via Zoom or in person.

Covid Guidelines

As the Covid infection is still with us , we would request that “in person” attendees follow some sensible guidelines, summarised below:

• please don’t attend if you have any Covid symptoms or test positive
• respect other attendees by maintaining social distancing both in the lecture theatre, and when socialising before the event
• please wear masks when not eating
• we would recommend talking a Lateral Flow Test before attending the event

We will be maximising the ventilation on the lecture theatre, so attendees should be prepared for the lecture theatre to be cooler than usual.

We therefore look forward to welcoming you on 14 April, whether attending in person or via Zoom, for what should be an interesting talk.