Among the many applications of artificial intelligence there were probably few who thought it could save lives by shifting patients around hospital wards.
But the pandemic has proved otherwise, with an AI system invented by a former National Health Service operations manager allowing hospitals to efficiently reorganise beds to cope with the influx of Covid patients.
The past two years have brought many tremendous medical advances, but Orlando Agrippa, 37, had little idea that his innovation now used in Britain, the UAE and Australia would be tested in such extreme circumstances.
“We can predict the type and volume of patients who might appear at the front door of the hospital, which allows our healthcare colleagues to be prepared for that surge in demand,” the RwHealth founder and chief executive told The National.
His system allows clinicians to predict if a hospital is in danger of being swamped and whether they should expand bed numbers to meet demand.
Its algorithms also tell doctors and nurses how long a patient is likely to remain in hospital and whether they can be safely discharged for care at home.
The NHS is the world’s fifth-largest employer. With 1.7 million staff who see more than a million patients on an average day, AI will play an increasingly important role in helping to streamline the operation.
The data science platform was developed by RwHealth, formerly Draper & Dash, a year before the pandemic hit. It is now proving to be crucial in the fight against Covid-19 by providing real-time information on patient flows.
Mr Agrippa devised the system when he became fed up with weekly flights between London and a Scottish hospital to assess patient flows.
“I thought that there must be a way to automate an advanced technology that replicates the stuff that I do, I pressed on with it and the rest is history,” the software developer said.
As an operational manager for the National Health Service for more than 15 years, he understood the benefits of creating a system that absorbed live data from medical systems.
After inputting millions of patient pathways, DSP views all the evidence from months or years of hospital admissions and gives a profile for the patients and complexity of their illness.
With all the data absorbed, the platform can predict how long someone will be in a hospital bed and the resources they might consume.
“Teams on the ground use that insight and analytics to look at ways in which they can accelerate the discharge of patients who don’t need to be in a hospital bed and can be managed within the community,” he said.
Freeing up beds has been vital in the pandemic with hospitals at times being overwhelmed.
The knowledge of underlying conditions, such as cancer or pulmonary hypertension, also allows the DSP to categorise risk.
“This capacity is driven by the algorithm, which is capable of looking at millions of rows of data of patients who have previously presented,” he said.
“We then leverage these insights to determine how the triage process can best be supported.”
But while the AI provides visibility and suggestions, it is ultimately down to the “clinical voice to make the clinical decision”.
The AI comes down to a clarified management of numbers and need. “The simple question that we want to provide clinicians, operational managers and management teams with is ‘How many patients do we need to see, how frequently do we need to see them and what’s the complexity of those patients?’”
While the system currently predicts Covid admissions in Britain will tail off by the end of January, it suggests that come spring there will be major pressure for operations.
“We’re likely to see a massive issue in the spring with a lot more people who have deteriorated over the Christmas holidays because they’ve not been able to have elective procedures.”
DSP has already helped sift through the huge number of people attending hospitals. NHS figures show each day 50,000 visit accident and emergency, an additional 94,000 are admitted for emergency treatment and there are 49,000 outpatient and 830,000 GP (community doctor) appointments.
Come the spring with coronavirus potentially waning, the software will play an important role in streaming the flow of the 36,000 people who attend hospital each day for planned treatment.
While Mr Agrippa does not believe that there will be a sudden rise in mortalities among those six million awaiting operations, he urges healthcare leaders to be “mindful of the potential for many to come to harm”. “What I mean by that is people who have had chronic pain or deteriorating conditions that should have been operated on, I think other things will start to happen to them because their quality of life has gone considerably down.”
The AI system is showing that patients who have been waiting for at least two years for operations postponed due to the pandemic are deteriorating significantly.
This is where the technology is now helping health planners “optimise pathways so that they can be treated”.
Two years ago NHSX, the organisation driving the digital transformation of health care, announced a £250 million ($340.7 million) investment to bring in more AI technology into Britain’s hospitals. This is a relatively small amount of the NHS annual budget of £179 billion, the second highest in the UK after social welfare and pensions.
Mr Agrippa also predicts that there are three situations that might overwhelm certain health systems around the world. Staff absence through sickness, combined with an influx of Covid patients and frail people in urgent need of operations might have a grim outcome.
Developing countries such as those in South America are a big concern for capacity constraints exposed by waves of infections. “The unvaccinated part of the population is enormous, the clinical workforce is low and the ability to flex capacity doesn’t exist in such places as Brazil.”
While DSP already provides services to Abu Dhabi’s main health system, Mr Agrippa is looking to do “some really creative things” to develop technology that eases patient flow and capacity.
With a recent £6 million cash injection and more venture capital funding on the horizon RwHealth is looking to develop future AI system that can predict diseases by using data from current sufferers.
“We’re working on a current innovation that will help identify patients with undiagnosed conditions. This capability will not only help expedite their diagnoses but will also ensure that they receive the care they need faster.”
As a data analyst, Mr Agrippa, born in Guyana, central America, could not resist taking a DNA sequencing test. It showed he was 10 per cent Chinese and 8 per cent Indian, with the rest from Ghana and Nigeria.
It’s a “pretty interesting mix”, he believes, one that has not only given the entrepreneur considerable success but also streamlined hospital admissions, potentially saving thousands of lives.
