Dr Joachim Werr and Dr Simon Swift of Health Navigator on their AI-powered solution that identifies patient risks, enabling preventative care
Prevention is the new buzzword in healthcare. Thanks to Artificial Intelligence (AI), a new era could be dawning as tools are created by data scientists and clinical teams to pinpoint problems before a crisis. Linked with huge datasets, there is the opportunity to screen the entirety of a population as part of population health management programmes.
Health Navigator has been working on these solutions for more than a decade. Its AI tool set can identify 80 per cent of patients who may be at high risk of disease progression and hospitalisation between 6-12 months before it happens. This insight has been proven – in clinical trials and real world applications – to save lives, improve clinical and patient outcomes and enhance efficiency for health systems.
Founder and Chair Dr Joachim Werr, a former A&E physician, was inspired to create a solution to prevent the same patients reappearing in the department in crisis by using the existing hospital data. In this way, clinicians could intervene before emergencies arise. “We knew that 5 per cent of patients consume more than 80 per cent of hospital resources, so it made sense to see if there was a method of identifying those most at risk”.
“Many elderly people go in and out of what you could term a clinical crisis where they need a lot of care,” he says. “Most are vulnerable older people with uncontrolled chronic diseases. Yet 8 in 10 high utilisers are new each year. The interesting thing is that the following year they won’t be high consumers, either because they are no longer with us or those particular issues are resolved. Thus, it became imperative to learn how to find out when these issues would arise in near-real time.”
Pivotal trials
Health Navigator’s data-driven solution is seeing increasing uptake across the UK & Ireland, and further afield, thanks to leveraging technology to shift focus from reactive treatment to data-driven early intervention. Since it was founded in Sweden in 2010 and then established in London in 2015, the company undertook randomised control trials in both countries. In 2018 they launched a UK trial to identify patients in real-time, and the published results in 2019 showed a strong impact from AI predictive care.
With advice from the Nuffield Trust and with the support of several NHS trusts, the trial meticulously tracked up to 2000 patient outcomes across multiple trial sites. “It revealed that patients over 75 years of age experienced a significant reduction in mortality rates—up to 50 per cent for elderly men,” says Joachim. For younger patients, while no mortality impact was observed, the intervention still delivered significant reductions in healthcare usage.
The company’s approach is built around a three-tiered system:
1. Predict Platform: The core component aggregates and analyses data to identify individuals at high risk of adverse outcomes. Whether it’s predicting emergency care needs, long term condition exacerbation or potential readmissions, the platform offers a granular view of patient risk that can be tailored to specific healthcare commissioners’ needs.
2. Engage Platform: Once high-risk individuals are identified, the Engage platform re-identifies and contacts them, inviting them to participate in a personalised care programme.
3. Proactive Platform: Trained clinical coaches, either from the company or from the healthcare provider’s own staff, interact with patients remotely, assessing their needs holistically. Unlike traditional healthcare settings where a cardiac nurse might only address heart-related issues, the proactive coaches tackle all aspects of a patient’s wellbeing—from medication adherence to social support needs.
There is plenty of flexibility in the structure so, for example, some organisations might opt for just the Predict platform to identify high-risk patients, while others might implement the full suite, including proactive coaching.
The importance of data
The main piece of work is undertaken by their platform, Predict. For data specialist and doctor Simon Swift, it begins with understanding where the data sources are and legal routes for access. “There is some lack of clarity at the moment regarding use of data in this way, the current UK legislation is a bit clunky when applied to this, as the case finding, engagement and coaching is primary purpose but the risk model training is probably secondary use. This makes information governance people a bit jumpy on occasion”.
Secondly, we look at the structure of this data, along with the data architecture in that particular region and area. “In the UK some ICB’s have integrated data sets that also have a clear interface and agreed governance for secondary use purposes, but others don’t,” he says. ‘We need to see the quality of the data and identify the different data sources to access that data, ideally on a daily basis.
“From a patient perspective, we take whatever data we can get and link all those data sources together on an individual level. Then we can understand exactly what kind of care each individual is able to access and what they need. We create a risk profile for that patient.”
Even if the data available is only hospital data, their predictions have been surprisingly successful, enabling them to work in regions with less integrated data systems.
Engaging the patient
Once the data has been gathered, it is returned to the system to identify those in need of proactive support and coaching. “It’s quite different from the norm where patients seek care. Here, we reach out to them proactively,” explains Simon who goes on to clarify that while some may initially be surprised to be contacted out of the blue, many appreciate the proactive support. “We’ve conducted extensive patient engagement studies and even have video testimonies. Around 70 per cent of patients contacted through this method engage with the service, which is a high uptake,” he adds.
During the initial meeting, the patient and coach review all pertinent medical records, discussing care contacts, medications, and any other factors that may influence their health. This holistic review often uncovers issues that might otherwise be overlooked in a traditional care setting. Whether it’s a spouse struggling to provide adequate support or a patient unable to afford their medications, the coaches work to address these root causes. The company runs an accredited coaching system that follows their own certified methodology and allows the coaches to use the platform.
The aim behind the concept is for the coaches to identify which areas of health and social care the patient consumes most and to try to instill a programme of self-care that enables them to better look after themselves. “The coaching intervention normally lasts no more than four months and delivers a total of around between 10 and 12 coaching hours from the nurse to the patient,” Joachim says. “If you look at the impact we achieved in the randomised controlled trial, it’s a very low resource, investment for the return that you get. We found that the patients who received the coaching intervention consumed 32 per cent less non-elective care events in hospital, meaning they had fewer days in hospital. It has a huge potential to decrease A&E activity and urgent care bed utilisation across the NHS by a third.”
Enabling preventative care across providers
Healthcare Navigator’s solutions is a global offering which can be structured to each region and culture through local partnerships, calibrated to the local population. “It saves money and releases capacity,” says Simon. “That’s the business case that allows us to deliver that win to patients. We can work with insurers and to public sector payers like the NHS, or we can work directly with providers because we decrease demand and release capacity. We reduce the number of people turning up at night to emergency departments and we also reduce hospital bed utilisation so, if a provider has capacity issues, we are a solution for them.”
CONTACT INFORMATION