The Association of Clinical Biochemistry & Laboratory Medicine
NHS Tayside and University of Dundee have developed the Intelligent Liver Function testing programme (iLFT) which, whilst not an example of Artificial Intelligence (AI), provides an example of a stepping stone on the evolution to AI within routine NHS laboratories. How can such a programme be developed from scratch? Identification and understanding of the problem that requires to be addressed (including increasing deaths from liver disease, too many LFT results to handle, complex guidelines) are explored, along with how to get the experts to agree on a possible solution.
The formation of a multidisciplinary team of subject matter experts is paramount to the approach, including laboratory experts who fully understand the complex systems required to deliver an automated approach to the problem .i.e. how to convert the expert opinion into an automated laboratory algorithm.
Laboratory infrastructure for this ideally includes automated analysers on tracking systems, and an order communications system that can handle url links for access to in depth advice. The algorithm for iLFT was applied to historic laboratory data prior to testing in a formal step-wedged clinical trial in Primary Care. Health economic analysis of such new ways of working is recommended. Once embedded in clinical practice, ongoing audit and development ensure constant evolution through human, if not machine, learning, with iLFT now approaching its fourth iteration.