AI technology: A game-changer for rare disease diagnosis
The route to diagnosing a rare illness can be long, difficult, and extremely stressful. Often, these conditions present with a range of complex and seemingly disconnected symptoms, meaning they can sometimes remain undetected for years - even entire lifetimes.
In the UK, for example, the average diagnosis of a rare disease takes around 5.6 years, eight clinicians (including four specialists), and four misdiagnoses before an accurate conclusion is reached. And these conditions are also far more common than people realise, collectively affecting around one in 10 people globally, with children accounting for around 50% of those.
Due to these factors, rare and hard-to-diagnose diseases place a huge burden not only on individuals and their families, but also clinicians and entire healthcare systems. Therefore, it’s fundamental to prioritise and continue to tackle this diagnostic odyssey.
To achieve this, we believe that pioneering AI technology holds the key to widespread adoption and significant results. Here are three reasons why it’s a game-changer when it comes to rare disease diagnosis:
It empowers clinicians
AI technology isn’t designed to replace clinicians - its function is to help them do their jobs at an even higher standard. For example, many general practitioners may not have even heard of certain rare diseases or be aware of the symptoms.
And this is completely understandable. Research from Rare-X suggests that there may be as many as 11,000 rare diseases – over 2,000 more than previously thought - and on top of this, five to ten new conditions are described in medical literature every week.
To combat this issue, AI technology can apply validated case finding algorithms to patient healthcare records at scale, meaning potential conditions can then be flagged to clinicians, enabling them to recommend referral and treatment pathways based on this insight.
And this isn’t the technology of the future: it’s making a difference right now. Our medical device platform ‘MendelScan’ is currently implemented in over 50 NHS primary care practices across England, using AI technology to help clinicians identify rare and hard to diagnose disease patients far more efficiently.
So far, to date, MendelScan has scanned over 800,000 patient records and, over the course of last year, feedback gathered from clinical users indicated that 54% of cases were identified as valuable to review, with 36% moved forward for further analysis and testing.
The UK Government is also investing in this technology - Mendelian recently announced that it has been awarded £1.4 million as part of The Artificial Intelligence in Health and Care Awards (AI Awards), which is run by the NHS AI Lab in partnership with the National Institute for Health and Care Research (NIHR) and the Accelerated Access Collaborative (AAC), and deploys £123 million to accelerate the testing and evaluation of the most promising AI technologies that meet the aims set out in the NHS Long Term Plan.
It can accelerate early diagnosis
Early diagnosis is key in the field of rare diseases, for a number of reasons. Not only does it ease the suffering of individuals and their families by helping them access the right care and treatment faster, it also reduces the heavy burden these conditions collectively place on healthcare services and staff.
According to our findings, it’s estimated that MendelScan can detect undiagnosed diseases 4.4 years earlier than current standard care - indicating huge potential to deliver a range of benefits. Additionally, we’re hoping to identify patients who may never have been identified without technology such as ours.
Our 40+ validated case-finding algorithms cover a wide array of diseases, from ultra-rare diseases like Fibrodysplasia Ossificans Progressiva (FOP) to more prevalent, yet still rare, diseases such as Duchenne Muscular Dystrophy (DMD).
Last year, five possible cases of X-Linked Hypophosphatemia (XLH) were identified via MendelScan technology and all of these cases were moved forward by clinical reviewers. This is a highly encouraging metric, given that early intervention in XLH has a significant positive impact on patient outcomes.
It can make healthcare more equal
AI technology can also enable clinicians to treat more patients in an increasingly consistent and equal manner, regardless of background or status.
Encouragingly, there is emerging evidence from our pilots that MendelScan could reduce the impact that health care inequalities (including socioeconomic factors) have on early rare disease detection and diagnosis. Here in the UK for example, our solution has the potential to support NHS public health core values and address key priorities of the UK Rare Disease Framework.
Taking a wider view, in this way this type of technology is set to revolutionise, not just rare disease diagnosis, but healthcare more generally. Cutting-edge advances, including big data, AI, and machine learning are likely to become the fabric of these services in the future.
The leveraging of analytics to mine significantly untapped reserves of clinical data will aid doctors and specialists in providing highly efficient, personalised medical treatment and care to increasing numbers of patients, with the ultimate outcome being to save and prolong more lives.
While separate technologies will, and already do, contribute value alone, the true potential lies in the synergy of multiple advances across the entire patient journey.
It provides a scalable solution
The more AI technology is implemented and used, the more data it has to work with and the more accurate it becomes - meaning that it’s a truly scalable, global solution.
This year, Mendelian will evolve in line with the ambition of solving the diagnostic odyssey and helping more rare and hard to diagnose disease patients, at scale, across the world. MendelScan will continue to be deployed at a regional level within the UK to deliver proactive care for millions of people, and beyond the UK’s borders, we will be preparing for a number of international projects.
The future of rare disease diagnosis
Until relatively recently, the issue of how to accelerate rare disease diagnosis seemed almost impossible to solve - the sheer number of conditions and complexity of symptoms make it one of the greatest challenges in modern healthcare. But AI technology has helped turn a corner.
This type of innovation, when implemented effectively, will enable clinicians, and therefore the system, to work smarter and far more effectively, ultimately providing better outcomes for all.
The result of this is a lasting positive impact, not just on small patient populations, but potentially across healthcare sectors around the world.
About the author
Dr Peter Fish is the CEO of Mendelian, a healthcare start-up using AI technology to accelerate rare disease diagnosis.
