Smart Hospitals: The Future of Healthcare Powered by AI and IoT

Medical centres compete to achieve 'smart hospital' status

For many of their reluctant occupants, hospitals are not seen as shiny incubators of transformative technology. In the popular imagination -- in the UK and beyond -- many are seen as remaining far too dependent on antiquated postal appointment services and outmoded IT systems, blighted by poor building maintenance, bad food and harsh strip lighting. Yet, faced with ageing and growing populations and stretched budgets, hospital managers worldwide are leaning on technology to improve patient experience and outcomes. The "smart hospital" market -- a broad term for technologies including artificial intelligence (AI), the Internet of Things (IoT) and robotics that collate and analyse reams of clinical data to improve care and increase efficiency -- will be worth $148bn by 2029, Deloitte, the consultancy, has forecast. "Smart hospital capabilities will be built around the interaction between the physical building and the patient," says Frances Cousins, a partner specialising in healthcare technology at Deloitte. "If you [are in] a hospital bed, you might be able to order your food, call your nurse, see the plan of your diagnostics [and] lower the blinds [using your smartphone]." This is already happening in some hospitals. In the UK, Nottingham University Hospitals NHS Trust is testing technologies that allow patients at a neurological rehabilitation unit to use their voice or a bedside terminal to control the temperature, blinds and lighting in their rooms. AI-powered CCTV cameras also help spot patients with cognitive impairments trying to leave the building. Patients' electronic records are the technological building blocks of many smart hospitals. Hospital IT systems can connect these records that track the flow of patients into and out of hospital beds, wards and operating rooms in real time. Electronic wristbands fitted to patients can be checked simultaneously using care management and bed planning systems, with hospital performance targets monitored via screen-filled "control rooms", alerting staff to where there are blockages in the system. Critical medical equipment can also be located more efficiently. Rachael Ellis, a programme director at Hull University Teaching Hospitals NHS Trust in England, points to the time wasted in retrieving medical equipment borrowed by another ward or searching in storerooms. "Imagine a big storeroom, and there are all kinds of little [sterile] trays . . . and they all look the same," she says. The trust now uses Radio Frequency Identification (RFID) technology software from Zebra Technologies to track more than 70,000 assets, including syringe pumps, nebulisers and machine centrifuges. Staff can see the location of equipment and check inventory using a handheld device and a "wand" for scanning medical items. The trust estimates that, across 2,500 staff, the new system saves about 35 hours per employee each year in tracking equipment. The Cleveland Clinic, a medical centre operating in the US and London, uses an AI algorithm to predict which patients are at risk from sepsis -- a life-threatening condition caused when the body overreacts to an infection. If a patient is considered at high risk, an electronic alert is sent to an expert on a specialist sepsis team. Benefits of the pilot include reduced mortality and patients spending less time in hospital due to fewer "complications" during their stay because of improved screening for sepsis infection, says Sarah Hatchett, chief information officer. Cleveland is introducing this sepsis-detection tool across its organisation, she adds. At Guy's and St Thomas' NHS Foundation Trust in London, almost all prostate, kidney and lung cancer surgery is now done with the help of a robot. Surgeons control the robot's 'arms' from a console, providing them with a 3D, high-definition view while they operate. The surgical instruments and camera are inserted through tiny incisions in the body. Benefits of robotic surgery include less time in hospital and faster recovery because the surgery is "less invasive", according to the trust. Other hospitals are using robots for more mundane tasks. For instance, the Sunshine Coast University Hospital, in Queensland, Australia, uses self-driving vehicles to deliver linen, meals, and other supplies to wards, freeing staff to spend more time with patients. In February, Oulu University Hospital in Finland claimed that it had started using Europe's first 5G mobile hospital network. This network allows for instant communication between staff and augmented-reality glasses for nurses and doctors, giving them access to patient information, including during surgery. Mark Davies, chief health officer at US tech group IBM, predicts that new digital systems will require changes in how medical facilities are built. "The design of hospital buildings will be more modular, meaning elements like floor plans and facilities will be more standardised," he says. This will allow digital systems to fit together like "Lego pieces", simplifying IT projects. However, to become "smarter," many hospitals must first address the legacy of their disjointed and ageing IT systems. In Germany, Charité -- Universitätsmedizin Berlin, one of the largest university hospitals in Europe, plans to install a new hospital information system, which will probably cost more than €100mn. According to research by Statista and Newsweek magazine, the hospital is already one of the top 10 "smartest" in the world. However, its scattered IT systems and networks can be challenging to manage. "Our hospital information system is probably 30 years old," says Henrik Andreasen, the hospital's chief information officer. "We have, over many years, implemented a lot of technologies in the different institutes . . . overall, those technologies are not speaking together." Modernising IT systems and installing new technologies such as AI and data analytics can come at a heavy cost. Mike Jones, a healthcare technology expert at the Gartner consultancy research company, suggests that to achieve "smart hospital" status, a typical medical centre might need to almost double the proportion of its total operating budget spent on IT. He estimates that for a large UK hospital with a total annual operating budget of between £800mn and £1bn, increasing IT spending from about 2.5 to 4.5 per cent could cost an additional £15mn per year. Such cost demands may partly explain why fewer than one in 10 hospitals worldwide are now classified as smart hospitals, Jones adds. Still, most experts believe smart hospitals will become more common over the next decade as health providers seek ways to improve care and operate more efficiently. "[Becoming a smart hospital] can feel at times pretty slow going," says Hatchett of Cleveland Clinic. "But ultimately, I think we've seen some very promising early results that I think are . . . worth continuing to invest in."

Networked devices help head off medical woes and speed recovery

From ultrasound machines and ventilators to interactive, data-sharing knee implants and blood-pressure monitors, connecting medical devices over hospital networks and the internet is improving patient care and medical research. The increase in digitally connected devices has created a buoyant technology market known as the "Medical Internet of Things". By 2029 this software market will be worth approximately $134bn, up from $93bn in 2025, predicts Statista, a research company. It is difficult to determine the exact number of medical devices connected to hospital networks worldwide but Gregg Pessin, a healthcare technology expert at research group Gartner, estimates the figure to be between 2.2mn and 3.3mn. One example is a device developed by Impedimed, an Australian technology supplier, that can detect if someone is at risk of breast cancer-related lymphoedema -- a common and debilitating side effect of treatment. The swelling, often in the arms and legs, can be painful and restrict movement. The technology requires patients to stand barefoot on the device, which resembles a digital upright weighing scale, and place their hands on a platform. The machine sends a low-level electrical current through the patient to measure body fluid and composition. It can detect lymphoedema in under one minute, using data analytics and cloud-based software to analyse, store and transfer the data. Test results appear on a web portal and are fed automatically into the patient's electronic health record, allowing for earlier intervention. Monmouth Medical Center, in New Jersey, is one of the hospitals using the device, which the American College of Surgeons has accredited. Manpreet Kohli, the hospital's director of breast surgery, says that previously, a clinician at the medical centre would use a tape measure to check for early signs of swelling in a patient's arms -- a possible symptom of lymphoedema. However, slight differences in how each clinician measured, for example, the circumference of a patient's arm meant it was a somewhat subjective diagnostic tool, says Kohli. "In the past, patients would come in with a swollen arm, and we would do measurements and . . . by that point [the lymphoedema] is usually already quite organised and it's difficult to reverse," she says. Now, using the Impedimed machine, a clinician can detect an increase in fluid in a patient's arm as small as "two and half tablespoons". And that is helping the hospital catch more patients with early-stage lymphoedema -- who do not have any visible symptoms -- and treat them sooner. Other healthcare technology providers are using AI and cloud computing to improve the clarity of scan images and make medical devices easier to maintain. GE Healthcare uses AI software to optimise medical scan images taken by its ultrasound machines. St Luke's University Health Network, a network of health clinics in Pennsylvania, US, is one of its customers. Lauren Fazzolari, an ultrasound specialist at St Luke's, says the technology has helped it improve patient care. "You can make the [scan] image drastically different depending on how you set the machine and the general way the [sonographer] scans. By implementing standardisation, you're able to control some of those variables and that's going to increase your quality." And because every ultrasound machine can now be updated once centrally, staff no longer have to travel to different health campuses to update each one. Some connected devices help doctors remotely monitor a patient's recovery after an operation. Canary Medical, a medical data company, makes sensors used in "smart" knee-replacement implants developed by Zimmer Biomet, another medical technology company. The implants incorporate sensors anchored in the shin bone that monitor a patient's gait, movement and activity. That information is sent online via the cloud to the patient's doctor, who can monitor the patient's post-operative recovery. Doctors can compare their patient's recovery -- based on age, gender and time since surgery -- to that of other patients who have had the same operation, using data pooled and analysed by Canary's AI software. According to one US surgeon, Jacob Ziegler, who has implanted Canary's replacement knee into a patient, one of the biggest benefits is an objective measure of a patient's recovery after an operation. "Historically, we had to do it very, very subjectively," says Dr Ziegler who is based in Mankato, Minnesota. "And patients didn't have any way of knowing what normal should be or what to compare it to." Yet despite evidence that connected medical devices can improve patient care, the resulting increase in sensitive medical data stored online also creates risks, experts warn. When Princess Alexandra Hospital NHS Trust in the UK used cyber security software from Armis to review its hospital IT network, it found more than 9,000 internet-connected devices -- ranging from medical devices to staff PlayStation game consoles and electric cars. Jeffery Wood, deputy director of information and communications technology at the trust, was concerned that a higher-than-expected number of connected devices could increase the risk of cyber security breaches. To minimise security risks, the trust created four separate computer networks -- including the main hospital IT network, another for connected medical devices and another for staff's own equipment, such as mobile phones and laptops. The new system has improved cyber security while giving hospital staff easy access to data on medical devices, Wood says. Despite such security concerns and some regulatory hurdles, experts predict that connected medical devices will become ubiquitous and improve healthcare. But they also caution that many medical devices are not yet used widely enough or receiving sufficient data to achieve their full potential for improving patient care. As Gartner's Pessin puts it: "We are collecting really rudimentary . . . [medical] data today. It's as if we're in first grade in calculus [and] we're having a lot of difficulty just adding one, two and three . . . but we'll get there."