In 2020, medical research is more important than ever before. The COVID-19 pandemic has opened eyes to the flaws of our global public health mechanisms and demonstrated the need for earlier, faster diagnosis – especially in the case of infectious diseases. The opportunities for AI within healthcare are infinite, enabling drug discovery, management of medical data, clinical trialling and more. AI is being deployed in thousands of medical initiatives, each that vary in approach and focus. Wider integration of these developments into the healthcare model could optimise the system and ultimately save lives. I’ve been looking into a variety of AI startups powering healthcare and discovering their ability to transform the pace, accuracy and success with which we conduct medical research in today’s world.
To begin with: drug discovery – the first step on the path to an accredited cure. Rather than beginning with a blank canvas, drug repurposing involves analysing the medical information already available to inform the development of new pharmaceuticals for new diseases: an approach being explored by experts amid the pandemic. As anticipated, AI is playing an important role in this field.
Cambridge based Healx is a leading company on a mission ‘to accelerate treatments for rare diseases’ through the use of its cutting-edge AI platform: Healnet. Healnet extracts disease knowledge from published sources and combines this with curated biomedical data. Adopting a ‘hypothesis-free’ approach, it draws links between drugs and diseases and offers powerful insights into potential treatments and cures, ‘filling in critical gaps in our knowledge.’ These insights are visualised and prioritised in a Knowledge Graph available to expert pharmacologists and biologists to assist their work in drug development. Whilst Healx draws inferences from abundant medical sources, Switzerland based Novartis adds to the information pool from the lab. Since October 2019, they have collaborated with Microsoft to unite pharmaceutical expertise with technology in their AI Innovation lab – ‘the Novartis engine and go-to place for AI.’ Novartis is currently using machine learning to classify images of cells being treated with different experimental compounds. AI firstly makes predictions about which compounds are biologically active and worth exploring and then groups those that have a similar effect on the cells. As explained in this Digital Authority Partners article, ‘as computers are far quicker compared to traditional human analysis… new and effective drugs can be made available sooner, while also reducing the operational costs.’ These initiatives show real promise and may progress drug development at a faster, more efficient rate.
Once a drug undergoes clinical trialling, problems may still arise. A recent report by Proventa International revealed that ‘patient recruitment was ranked as the most important challenge facing clinical operations professionals in the next few years’ and ‘retention of patients… and patient engagement were serious considerations for the next decade.’ Deep6 AI reports that 86% of clinical trials fail to recruit sufficient patients’ which slows research, delays treatment and increases drug costs. To tackle this, their platform ‘applies artificial intelligence to medical records to find more, better-matching patients for clinical trials in minutes, rather than months.’ The software extracts thousands of clinical data points, such as the symptoms, genomics and lifestyle of patients, and arranges this in a graph format that can be easily matched with complex clinical trial criteria. As displayed in their corporate video, it’s clear that this technology has great potential to save time and lives.
Once selected patients are enrolled on a clinical trial, their drug adherence is crucial in ensuring both the validity of trial results and also their personal safety – nonadherence can contribute to antibiotic resistance, for example. Despite this, nonadherence in real‐world settings can exceed 50% in some populations, according to the ACCP Clinical Pharmacology Journal. AiCure offers a solution here: a smartphone app that uses the front-facing camera to visually confirm medication ingestion by the correct patient. Not only does this approach help patients maintain treatment regimens and spare them of frequent trips to the clinic, but it also provides an indicator of accuracy so that only reliable patient data ends up influencing the overall outcome.
The creativity behind AI medical initiatives is astounding: from Tempus’ record-breaking libraries of biomedical data to voice-powered digital assistants for scientific laboratories… the world is flooding with ideas. Despite this, creativity has to meet reality. As noted by David Shaywitz in this compelling Forbes article analysing a podcast interview with CEO of Novartis, Vas Narasimhan, ‘the current challenge for data science and technology (DST) in healthcare is moving beyond the “dancing bear” stage, where “the wonder is not how well the bear dances, but that he dances at all.”’ AI is exceptionally demanding, requiring huge investment and environmental backup to take the centre stage of frontline healthcare. It is of little surprise, therefore, that AI isn’t the silver bullet to solve COVID-19. But given the right space to flourish, it could be at our fingertips come a future health crisis.
Verne Global provides the best, most affordable solution in this space. Their Icelandic data centre, powered by 100% green energy, allows the most advanced technology to thrive in an optimised environment, kept in perfect condition for real-world use. Here, creative ideas can be realised as more than just ideas, bringing us much needed hope for the future.