A new research community for people with long-term conditions such as diabetes or heart disease, aims to give patients a voice, a choice and control over how they share their health information with healthcare professionals and researchers who are seeking to improve personalised care of these conditions.

Patients will be invited to shape the direction of how newly developed artificial intelligence technology will use their information to understand which combinations of medical care and lifestyle choices work best for different types of people.

People who join the community will have the option to share data such as their health outcomes, and the medical and non-medical treatments they received, via their NHS care provider. This data will help thousands of similar people who have the same condition, or the same combination of conditions.

The community will be open to healthcare providers and researchers that have an interest in partnering with patients to learn what approaches work best in achieving the health outcomes that they want in the way that they want.

The move comes off the back of five years of research and development, part funded by UK Research and Innovation, that enabled Healum, the digital health software provider behind the initiative, to test and develop its community-based approach to designing and training AI healthcare solutions. The funding enabled Healum to crowdsource the inputs of thousands of patients and healthcare professionals to codesign its digital healthcare platform and to shape the way that its machine learning solutions should be built and managed. Through close collaboration with the UK’s National Institute for Health and Care Research and Vernova CIC, Healum formed a research partnership across Greater Manchester Clinical Research Network that worked to the National Institute for Health and Care Excellence’s evidence standards for digital health.

Healum’s digital health platform enables healthcare professionals working across primary and secondary care to co-create digital care plans with their patients as part of a shared decision-making process. These care plans can then be easily accessed by the patient via a mobile app.

An initial project which trialled this approach with people living with type 2 diabetes showed promising results. Across 27 practices in London and East Cheshire, patients showed a reduction of 9.5% in HBA1C within the treatment group, compared to a 2% reduction for the control group. Ninety-eight per cent of messages from healthcare professionals were opened and over 50% of people used personalized resources and goals to achieve a reduction in HBA1C.

The project enabled Healum to test and validate both its underlying AI technology to train its personalised care recommendations, as well as cementing its ethical approach to AI that focuses on trust, efficacy, transparency, equality and control.

Anuj Saboo, CTO and Cofounder of Healum, said:

“Every development decision we’ve made regarding our AI models for personalized care has been informed by patient and healthcare professional stakeholders. We know that choice and control over the role that their data plays in informing their health journey – and the journey of others – is crucial. This principle inspired us to launch our AI research health community that uses a live learning model as part of the shared decision-making process when managing long term care”.

Dr Adrian Heald, Chief Investigator of the UKRI study and Consultant Physician in Diabetes and Endocrinology said:

“A key challenge in long-term condition management is to have a sense of how the past and current health profile of a person can best inform care plan choices for the future. We often find ourselves scratching our heads to know what the next best step in long-term condition management is and any intelligence as to the best options, at the point of care, will only serve to help healthcare professionals to improve patient outcomes.”

Jonathan Abraham, CEO and Cofounder of Healum commented:

“Artificial intelligence has the potential to unlock many benefits for healthcare if it is guided by the wisdom of the people that deliver and use services. Our research to date has seen the powerful network effects of crowdsourcing data from healthcare professionals and their consenting patients to train machine learning algorithms that can identify the optimal set of choices for an individual.

“We’ve listened to the needs of people with long-term conditions and the healthcare professionals that support them to understand the daily support they need to navigate the complex set of medical and non-medical choices they face when managing different conditions, and the role of AI in assisting in that process.”