AI And Machine Learning Helping Hospital Readmission Rates

Automated healthcare

Imagine you go to hospital, you get treated, you leave when you’re ready and, aside from scheduled follow up appointments, that’s it. It sounds like what you would expect from a hospital experience, but the reality can be very different. Huge pressures on healthcare systems and professionals mean some patients are discharged earlier, or later, than they should be.

The results – re-admissions, and more serious health issues for patients and blocked beds – have very serious consequences, putting yet more strain on an institution already close to breaking point.

What’s the answer? The solution can only be found within the data lakes of patient information, and the only viable way to make full use of this is with AI.

Putting pressure on the profession

Around one million patients are seen daily by the NHS. Despite these huge numbers moving through the system, an estimated 4.3 million people are still waiting for treatment. The number of beds available in hospitals has, despite a growing population, more than halved in the last thirty years, going from 299,000 to just 142,000. To compound this further, the stranded patients – those who are kept in hospital longer than they need to be – are taking up a bed that, in an ideal and more efficient world, could be put to better use for someone else.

This puts into stark perspective the immense pressure to get patients in, treat them and get them back out so the cycle can begin again. It makes hospital treatment sound rather like patients being on a conveyor belt, or even a widget in a factory, and in some ways it might be better if it was considered this way. Of course, releasing patients as quickly as possible is good for everyone when it’s done right, after all, no one wants to be in hospital for longer than necessary.

However, there can be a very serious cost – both physically for patients and financially for the NHS – if people are sent home too early and this almost inevitably leads to higher levels of readmissions. There are wider implications for healthcare professionals and hospitals when patients suffer. As well as the financial impact, this can take a toll on mental wellbeing and dent reputations, reducing confidence and trust in institutions and staff.

Making an appointment with AI

Patient flow is one area where AI and machine learning can make a very real difference and support the work of healthcare professionals at every level. It already is in many hospitals. And while there is no substitute for the expertise and experience that clinicians provide for diagnosis and intervention, AI can help provide additional data-driven support. By saving the years of knowledge and judgement calls of our most prized healthcare professionals in AI algorithms, we can also save the best of our diminishing skills base for future generations of patients.

By harnessing hospitals’ rich patient data and leveraging advanced algorithms, AI can help to better predict some of the variables already instinctively handled by healthcare practitioners, supporting them to reduce readmissions, put the right staff in the right place at the right time, and identify problems before they occur. In fact, AI offers almost limitless possibilities in helping teams to better understand which patients will be more at risk of being readmitted and providing additional insight into, and control of, the patient journey.

It is reported by the Medicare Payment Advisory Commission (MedPAC) that 75 per cent of Medicare patient readmissions can and should be avoided. ClosedLoop and SafeCare AI are two healthcare focused technology companies that use predictive analytics taken from electronic medical records to cut avoidable readmissions.

While I don’t believe for a moment that machines will ever be able to replace humans when it comes to patient care, there are so many areas where machine learning can be an invaluable asset. It is essentially an additional and impartial supporting tool that helps make informed clinical decisions. Machines can analyse years of data, predict the impact of many different circumstances, making it easier to maintain patient flow, plan resources and even preempt peaks in demand. The result? Greater efficiency, reduced readmissions, better patient outcomes and more effective use of resources. This will transform lives and will benefit the bottom line in an increasingly budget constrained system. In short, as a society can we afford not to make use of this technology? 

AI is sometimes viewed with suspicion, despite the huge opportunities and proven success across a wide range of sectors. We need to stop thinking about man vs. machine and focus more on how technology can save struggling business models or inefficient factories. There are a few forward thinking hospitals who are developing strategic partnerships with AI capable businesses, but the number is small and there is resistance to change across the wider healthcare sector. 

Building an awareness

Though most healthcare leaders and providers are concerned about the future of their hospitals and the challenges faced by the healthcare system, just half are considering any form of AI or machine learning as a viable solution. This is due to a number of real issues that are holding back progress. Chief amongst these is the limited understanding by healthcare professionals and decision makers of the truly transformational impact this technology can have on patient health and hospital costs. We need to make sure they are aware. But there are also the new funding rules which requires that ROI on technology in the NHS must be realised within a year. Imagine how many technology startups would still be around if they had to adhere to these rules? So many tech unicorns that are household names and viewed as the future of industry or commerce have yet to turn a profit. If this is not an option for health how can it innovate?

While other industries are already benefiting from being disrupted by new technology, healthcare appears to still have some way to go. There is serious potential, but for the health space the stakes are so much higher. With awareness of the benefits this technology can bring and an innovative environment, the NHS can and should be a pioneer. It has the data at scale, the institutional knowledge and expertise and needs to mobilise that for the benefit of all patients.

Readmissions aren’t just problematic for the NHS – they are a societal and economic challenge. AI services offered by the likes of ClosedLoop and SafeCare AI, though, may go some way to encouraging healthcare providers and facilities to adopt artificially intelligent technology. Across the globe, healthcare systems are searching for ways to remain viable and AI will be a critical component in ensuring their future.

For more expert insights on the future of AI and healthcare, sign up to our free weekly newsletter here