SPONSORED CONTENT

MendelScan, one of the winners of the Artificial Intelligence in Health and Care Award, is a rare disease case-finding platform that helps find the hidden clues in patient electronic health records, aiding clinical decision-making and reducing time to diagnosis. Led by the platform’s creators, Mendelian, a highly engaged Patient and Public Involvement Group have been instrumental in shaping how the use of such technologies can be successfully communicated to the public

Who is Mendelian?

The health-tech innovation company Mendelian have been embedded in the rare disease community since 2015. Their mission is to help end the diagnostic odyssey for individuals by bringing timely patient insights to primary care. Transforming how rare disease patients are identified, diagnosed, and treated they use clinically informed and secure technology to accelerate diagnosis and achieve better patient outcomes in alignment with the vision of the UK Rare Diseases Framework. 

Harnessing the power of data and technology, combined with clinical expertise, Mendelian develops algorithms that GPs can use to scan millions of health records via their MendelScan software, flagging patients with potential rare diseases and helping to initiate the diagnostic process sooner.

What is MendelScan?

MendelScan is a software that empowers frontline healthcare professionals to uncover health insights from symptoms that might have previously gone unnoticed or unconnected MendelScan software integrates with a GP’s computer system where it carefully looks at patient health records (EHRs) to find signs and symptoms that might suggest a rare or hard-to-diagnose disease. With over 40 current filters, the system can quickly identify patterns and anomalies. Mendelian’s team of doctors then review matched records to ensure the insights are accurate and relevant, providing a robust, human quality control layer on top of the AI generated insights.

MendelScan is a registered medical device which means it is bound by strict safety standards and undergoes further rigorous quality checks to ensure accuracy and reliability. Mendelian works with GP practices under strict data agreements to ensure compliance with the relevant laws and MendelScan does not access personal identifiable data like the person’s name and address.

Patients whose data is flagged are brought to their GP’s attention with an outline report on the selection rationale and suggested next steps for referral or supplementary testing, all in line with NHS pathways. As part of the continuous improvement goal, GPs are invited to provide feedback on the usefulness of the insights in supporting their clinical decision-making.

Working closely with the NHS in the UK and as a key partner for the NHS Genomic Medicine Services Alliances (GMSAs), MendelScan has been used to scan over 10.5 million health records to date.

EHR: electronic health records are a digital version of a patient’s documented medical history. It includes test results, diagnoses, prescriptions and treatment history. Patients may choose not to share their health records outside of their GP practice for purposes other than direct care by using the NHS opt-out Service and in this case, their EHR will not be accessible by MendelScan.

NHS AI Health and Care Award

In recognition of MendelScan’s potential to improve rare disease diagnosis in the NHS, Mendelian was one of nine “promising artificial intelligence (AI) healthcare technologies” to receive a share of government funding to accelerate their research.

The aim of this 18-month evidence-generating research project was to evaluate whether MendelScan could deliver on four key aims:

1. Reduce time to diagnosis for rare disease patients
2. Increase overall disease diagnosis rates (find previously undiagnosed patients)
3. Reduce avoidable healthcare activity associated with the pre-diagnosis phase for rare disease patients
4. Be implemented practically, ethically and affordably within the NHS

To allow for MendelScan to be commissioned and deployed within the NHS they were required to achieve three layers of evidence:

• Validation: large scale retrospective validation of the algorithms on 23-million de-identified NHS records under ethics.
• Implementation: research studies on practical deployment aspects, public sentiment, clinical acceptability and usability, health economics and resource impact modelling, and NHS commissioning.
• Real-world deployment: Deployment at pilot sites followed by third-party evaluation under ethics.

Patient and Public Involvement Group

To consider the patient and public sentiment toward such technologies, it was important to Mendelian to have an engaged group that could provide feedback and act as a sounding board for all activities. After an open recruitment, a PPI Group of seven was formed, chaired by Toni Mathieson, CEO of the rare disease charity Niemann-Pick UK, and Trustee to two further international Niemann-Pick organisations. The group represented diverse voices with differing perspectives from both the rare disease community and the wider public.

Meeting quarterly, the PPI Group provided feedback on written materials and helped shape the core activities within the various aspects of the project. They provided crucial input into the design and testing of a Public Perception Survey on the use of technology for rare disease case finding—input which resulted in increased public participation. They provided a valuable role in evaluating the communication of research findings and results—advocating for clearer, patient-centric, accessible language. Their input into patient-facing communications was particularly beneficial and resulted in outputs with clearer, more compassionate language with an emphasis on emotional support and practical considerations.

By working closely with the PPI Group, Mendelian ensured that their interventions were not only technically sound but also aligned with the needs and the values of the patients and public they serve.

“As a parent of children with an ultra-rare disease, I’m acutely aware of the diagnostic odyssey that patients and families often face. This inspired my interest in AI, its potential to shorten the diagnostic process and enable timely access to effective care.

Working alongside the Mendelian team as Chair of the PPI Group has been an incredible opportunity. The development of MendelScan offers practical solutions that stand to make a real difference for the rare disease community and beyond, and it has been immensely rewarding to see how our input has helped bridge the gap between complex science and the people it ultimately serves.”

Toni A Mathieson, PPI Group Chair, and CEO, Niemann-Pick UK (NPUK)

Managing public perceptions and expectations

One of the key takeaways from the engagement work with the PPI Group was the importance of public understanding of the technology. In general, there is a real nervousness among the public regarding the use of artificial intelligence and sharing personal data, so it is vital to address these concerns fully when communicating about the use of such technologies within our healthcare. But the group also identified the need to set and manage expectations around MendelScan’s capabilities and limitations.

MendelScan’s performance is directly related to the quality of the data documented within the patient’s EHRs and the presence of relevant signs and symptoms within that documentation. This means that the tool might not detect all individuals with a condition if their symptoms are not clearly recorded. It is important for both healthcare professionals and patients to understand this.

It is important also to communicate that MendelScan is not a diagnostic tool in itself. It requires the interaction of the lead GPs at the practice to review and then action the recommendations of a timely specialist referral and/or additional testing. Only at this point can a confirmed diagnosis be made.

Another important observation made by the PPI group was the polarity between patients and healthcare professionals when considering the tool’s accuracy. In the context of rare disease diagnosis, patients prefer a tool with higher sensitivity—one that picks up more true cases, even if it flags many people who do not ultimately have the disease (false positives)—to increase the chances of a previously overlooked patient actually receiving a confirmed and accurate diagnosis. This contrasts with the healthcare sector’s typical focus on specificity, which aims to minimise false positives (patients flagged inappropriately) to reduce unnecessary interventions and cost—but which may restrict exploration.

While there are various limitations that the public must be aware of, this technology holds significant promise for enhancing healthcare systems and providing GPs with an additional tool in the quest for shortening the rare disease diagnosis journey for their patients. Early validation results have shown that implementation of MendelScan has the potential to lead to a diagnosis up to four years earlier than traditional methods and in one disease area studied, the 14% of patients flagged by MendelScan were confirmed diagnoses, but as this was conducted on data known to be incomplete, the reality could potentially be even higher.

So far MendelScan has identified over 1,000 people with a potential rare disease and this figure is likely to increase exponentially as its use is more widely rolled-out across the NHS and internationally. Within the world of rare and hard-to-diagnose, this represents an exciting step forward.

---

To learn more about Mendelian, visit: https://www.mendelian.co

To learn more about the NHS AI Health and Care Award: https://www.nihr.ac.uk/news/ground-breaking-ai-research-aims-to-improve-tests-and-treatments-for-thousands-of-patients/32852

---