A Scottish health board has deployed AI on patients with a chronic lung condition in a bid to ease “winter pressures” on hospitals.
NHS Lothian has used the technology to work out how many of the 25,000 patients with chronic obstructive pulmonary disease (COPD) in its region are at risk of deterioration and potential hospital admission.
Using a machine learning tool developed by Edinburgh digital health company Lenus Health, it has identified 5,000 patients in a high risk category who will be contacted for a review by 100 local GP practices over the winter period.
It is hoped the intervention will prevent unnecessary hospital admissions, easing the pressure on the NHS and supporting the Scottish Government’s winter preparedness plan. The intervention began in recent weeks and initial feedback has been positive.
Jenny Long, director of primary care for NHS Lothian, pictured, said: “We recognise the role that data and machine learning can play in unlocking valuable insights to assist GPs in continuing to provide excellent care to their patients. Our work with Lenus Health is an exciting step in unlocking the potential of data to better understand our patients and the risks they face.”
Dr Jeremy Chowings, deputy medical director primary care for NHS Lothian and GP and partner at the Leith Surgery, said: “The initial outputs show significant promise and have already informed the approach we are taking in advance of winter. This work underlines the importance of data and accurate data coding to allow us to maximise the benefits that AI approaches such as these can bring.”
Lenus Health used its trademarked Stratify® models to calculate the risk of each of the 25,000 COPD patients – based on routinely collected medical data.
It was granted access to the health data repository ‘DataLoch’ – a joint venture between NHS Lothian and the University of Edinburgh – to deploy its algorithms, albeit no personally identifiable data was shared with the company.
Instead, each patient was assigned a unique user code, which could then be rematched with the original data stored in DataLoch, allowing doctors to identify which patients had been scored with the highest risk.
In total, the AI tool was used to calculate risk scores based on 100 different routinely collected medical datapoints, including hospital visits, lab results, blood tests, prescribing information and even local deprivation factors.
Andy Simpson, Lenus Health product director, said: “We are delighted to be working with a team who understands the importance of preventative healthcare models. Proven, ethical and evidence-based AI models are essential to help re-orientate care for chronic conditions from what is currently a reactive, high-cost approach to one that is proactive and preventative. By doing so, we not only help patients keep out of hospital but reduce the burden that long-term conditions such as COPD has on the NHS, particularly over winter.”
Shane Burns, head of data science, added: “We are thrilled to see AI transforming patient care in the UK. I’m especially proud of the innovative work by the team at Lenus and how it’s making a meaningful difference in the lives of patients.”
The use of Lenus Stratify® in NHS Lothian is a potentially groundbreaking advancement in the proactive treatment of COPD, aiming to break the winter crisis cycle. The models were initially developed in Glasgow with leading respiratory specialist, Professor Chris Carlin – and they are now being rolled out in England to proactively improve patient outcomes.
The models have been accredited to national ISO 13485 and ISO 27001 quality and security standards and are registered with the appropriate regulatory bodies.
COPD is a respiratory long-term condition associated with smoking and poor air quality. While COPD cannot be cured, patients’ symptoms can be managed effectively with access to appropriate care. However, patients with a COPD diagnosis account for 1 in 8 of all hospital admissions in the UK and COPD care is estimated to costs the NHS over £700 million per annum. A recent Asthma & Lung UK report identified that respiratory conditions such as COPD were a leading driver for winter pressures in the NHS.