HN Predict
HN Predict works by applying machine learning to routinely collected healthcare data, turning existing information into a powerful source for practically applied preventive healthcare. It is scientifically proven to identify 8 out of 10 patients requiring immediate preventive support to avoid unplanned hospitalisation and premature deaths.
One percent of the population - mostly elderly patients with multiple chronic conditions - consume 53 percent of all hospital bed capacity for unplanned care. Their conditions progress and ultimately, if nothing is done, they have little alternative to acute care. HN Predict is designed to particularly benefit this cohort of patients, and the healthcare providers who care for them.
Our data driven approach is designed for equitable health and can overcome recognised biases and inequalities built into traditional case finding and patient referral models. Having timely access to precise information about individual patients enables healthcare providers to better guide clinical decision-making to optimise secondary prevention. This generates better patient outcomes and releases pressure on frontline clinical staff and system.
As we developed HN Predict, more risks and future patient outcomes have been incorporated into the platform. Today, HN Predict can provide highly meaningful predictive information, on individual patient level, about a number of events that guide clinical decision making and preventative actions at point of care.
- Rapid frailty progression
- Unplanned hospital admissions
- Hospital readmissions
- Long hospital admissions
- Falls leading to care needs
- Delirium and confusion
- End-of life care
- Sepsis
- Acute Kidney Injury (AKI)
- Health care-associated infections (HAI)
Benefits of HN Predict
Makes unplanned care predictable and preventable!
Guides clinical decision making for highly targeted secondary prevention, triaging and clinical prioritisation
Takes a holistic approach to clinical risk and prevention, considering social factors
Significantly reduces health inequalities as patients who would ordinarily present late in their disease progression can be identified and contacted through outreach
Quick and easy to implement
Frees up capacity to support elective recovery
Validated for safety, efficacy and ethical fairness through HRA-approved clinical trial and government-funded R&D programme
Documented cost savings across the health economy
How it works
HN Predict collects and processes unstructured and structured healthcare data through an IG-compliant, secure cloud-based AI engine.
The AI analyses each individual patient’s data by using validated algorithms and provides a list of highlighted patients with elevated risks for certain events such as unplanned hospitalisation, rapid frailty progression or falls. This is information that can be used to better prioritise or to actively reach out to those at high risk for targeted prevention.
HN Predict can run on different primary, secondary or integrated datasets and feedback securely to be accessed at point of care through integration with existing EPRs or/and separate dashboards.
It is fully configurable enabling clinicians to define the most important events, outcomes, and patient target groups to predict.
Our evidence
HN’s research shows that many unplanned clinical events – including deaths in patients with chronic conditions – are to a large part predictable and preventable events. This means they can be anticipated well in advance and prevented.
In clinical trials, HN Predict demonstrated that it can reduce unplanned admissions and deaths while increasing Patient Reported Outcomes Metrics (PROMs). The positive impact was most pronounced in socioeconomic vulnerable patient groups.
Eight out of ten patients who will experience an unplanned hospitalisation can be identified while they are still at home, and while there is time to provide preventive support.
HN Predict meets the NHS DTAC standard and the Cyber Essentials Plus requirements. Elements of the HN Predict technology is undergoing PCT patenting.