How can artificial intelligence and machine learning radically transform readmissions?

By Orlando Agrippa, Draper & Dash Predictive HealthcareAnalytics

AI continues to offer a huge opportunity for sectors from pharmaceuticals and trials to factories, healthcare and more. Over 40 per cent of companies believe that their current models today will cease to exist in the next five years. Though many healthcare leaders and providers are concerned about the future of their hospitals and health systems, only half are considering any options around AI/ML. From speaking to a number of my NHS colleagues I understand that part of the reticence is a result of a lack of understanding of what AI entails.

In the simplest terms, I have always imagined hospitals to be similar to a hybrid of factories or hotels. And like with most factories and hotels, the issue of flow and process is vital to delivery of customer satisfaction/outcomes.

As far as numbers go the NHS will roughly see over 1 million patients each day with currently over 4.3 million patients waiting for treatment. In an earlier article, I shared my view on the issue of stranded patient, with there being an enormous focus on the NHS to release thousand beds due to patients being kept in hospitals longer than the necessary length of stay. One of the only ways to deal with this, like any hotel or factory, is to discharge patients faster and at times this can be done prematurely. However, of the thousands of patients discharged from hospitals each day, many of these patients are then re-admitted to NHS hospitals, some with serious complications resulting in further strain on services.

This issue could mean a lot of things for health systems, however if we think of this in terms of the metaphor of hotels or factories, this would be similar to the recall of a line of cars because they were delivered without being fully assembled and tested leading to poor safety for drivers or poor experience for owners. Similarly, with the example of hotels it would be similar to checking people out swiftly resulting in a situation where customers feel rushed to leave which results in a great trip becoming less memorable. Both of these examples are good comparatives to a readmission where the outcome and experience for patients can be extraordinarily poor due to premature discharge.

With the issue of readmissions AI offers some what a simple set of assistance to clinicians, operations teams and leadership. With hospitals having some of the richest patient data, leveraging advanced algorithms to better predict some of the variables already known to healthcare practitioner is now more than ever available as part of patient flow and management. AI offers limitless possibilities in helping teams to better know which patients are at a greater risk of being recalled or readmitted to hospitals by better understanding the variables driving the discharge process.

In my opinion, nothing trumps the clinical experience and prompt intervention for patients that are provided by a clinician’s years of experience in assessing patient’s fitness however harnessing data through the use of AI offers a great companion for seasoned veterans and recent graduates in healthcare to have more insights into the patient journey.

Some hospitals have begun this process by simply developing strategic partnerships with firms which have AI capabilities, developing a basic strategy, provide the challenges you face around readmissions as a use case to the firm to repurpose their capabilities to solve this challenge.

D&D is has extensively focused on supporting hospitals with solving readmission issues, for those of you wanting to explore this further please reach out to for a free demo of our supported predictive discharge and readmissions tools.