Artificial intelligence could explain surge in antibiotic prescriptions in winter

Analysts believe that AI technology could help us identify the reasons behind the surge in antibiotic prescriptions during the winter months.

New research by the database management software company Exasol showed that while London is leading the way in reducing antibiotic prescriptions, cutting them by 18 per cent in the past five years, rates surged to 50 per cent above the summer prescribing rate during the extreme winter weather of January 2018.

Using NHS Digital Data they established that even during the milder winter of December 2015, rates were still 34 per cent above summer levels in the city.

This has prompted concerns that GPs are still overprescribing antibiotics for winter viruses, which have little effect. Antibiotics are designed to kill bacteria by disrupting their cell walls, but the external covering in viruses is almost identical to the host cell’s membranes making them very difficult to target.

Exasol believe that with more data, AI and machine learning will be able to help pinpoint the reasons for more antibiotic use in winter.

By collating data on information such as the diagnosis for each patient, to which day it was prescribed on, and what the weather was like on that particular day, they may be able to pick out correlations between datasets not noticed by humans.

From there links can be made to see which specific illnesses are most commonly prescribed antibiotics incorrectly, so doctors can work on reducing overprescription and educating the public on the issue of antibiotic resistance.

Sam Sibley, Strategic Partner and Alliances Manager at Exasol said: “Although we didn’t use AI in this project as the size of the data volume didn’t require it, if there was more data looking at other variables then AI would be really helpful.

“If you have access to the right technology, you can turn anything into a data problem and help address it. You can spot things that might have some sort of influence affecting the outcome and get completely new insights in seconds, which empower you to do things differently.

“So in our research we’ve seen Southwark has achieved the biggest reduction in antibiotic prescriptions in London, but what is causing that? By gathering more data, we might be able to see,” he said.

However, Professor Helen Stokes-Lampard, Chair of the Royal College of GPs, doubted GPs were overprescribing in the winter months:

“It is normal for GPs to see a surge in patients with infections over the colder months, and as more people develop viral infections, a proportion will also develop bacterial infections and so it may be entirely appropriate that the rate of antibiotic prescriptions also increases over this period,” she said.

“This does not mean that GPs are prescribing antibiotics for seasonal illnesses like the cold and flu – GPs are highly-trained medical professionals and are all acutely aware that antibiotics will have no effect on viral infections, only bacterial.

“We know how vital antibiotics are, which is why GPs will only ever prescribe them if we genuinely believe they will help the person sitting in front of us and when there aren’t any more suitable treatment options available.”

However Mr Sibley believes AI could help to better identify when antibiotics are genuinely needed: “These techniques are definitely going to be commonplace in the future. A robot can calculate faster and more accurately than humans can. But I want to emphasise that it is very important to look at the data, and that the data is available to build on.

“AI ultimately comes down to the reliability of the data you’re using. It’s all about governance and understanding the journey the data has been on. Healthcare is a massively regulated market, with especially sensitive data. But if you have the right data, governed in the right way with proof and lineage you can develop fantastic insights.

“This is just the beginning. There’s so much more that can be done and so many ways that data can make everyone far more efficient and save lives. The healthcare market is evolving and changing and data is the biggest asset we have. Some organisations are still unaware of where their data resides, and it can often be complex and fragmented, but if done right it can be game changing.”