This week has seen a flurry of headlines about a worrying new drug-resistant fungal infection, Candida auris, which has been detected in 20 separate NHS trusts and independent hospitals in the UK.
Candida auris is a relatively uncommon infection found in many countries, which has become a healthcare concern due to its potential to cause outbreaks in hospitals, especially intensive care units. Affected hospitals in England have used a multi-pronged approach to identify, prevent and control these infections.
It might be surprising to learn that search engine technologies being developed at the Centre for Translational Informatics (CTI) are part of this effort.
One of the difficulties hospitals face in containing an infectious disease outbreak is rapid identification and isolation of affected patients to curb spread to other vulnerable patients.
Traditionally, a review of individual patient records has been used as a method of identifying cases. However, this is a very slow process which doesn’t suit the rapid response required during a disease outbreak.
That’s where CogStack – the CTI’s next-generation information retrieval and extraction platform – comes in.
CogStack uses advanced search, analytics and visualisation technologies to automatically extract and analyse the information in health records. Its natural language processing capabilities mean that it can take into account normal variations in spelling, description and language to rapidly and accurately identify specific queries.
We used CogStack to undertake a retrospective search to identify the first identified case of Candida auris in one of our NHS Trust partners, a major London acute hospital. We were also able to retrospectively identify patients with other Candida species that are commonly misidentified by the hospital laboratory. This gave us the assurance that we did not have cases before the declared outbreak period.
CogStack’s ability to search free text in health records when structured formats have not yet been established for new emerging pathogens is unique, and proved to be a powerful tool to help manage and track the outbreak.
Our ongoing evaluation of the programme means that we soon hope to have evidence for the positive effects of the technology in helping to manage disease outbreaks like this.
While advanced search technologies alone aren’t sufficient to solve the problem of drug resistant infections, we expect them to provide an increasingly important means to improve the way that we track and control disease outbreaks.
Richard Dobson is Professor of Medical Informatics at King’s College London. Research on CogStack is funded by the National Institute for Health Research (NIHR) Maudsley Biomedical Research Centre at King’s College London.