Smart hospitals: principles and tech trends of tomorrow’s healthcare

In his first interaction with the protagonist of the hit show Scrubs, the cold-hearted and cynical chief physician Bob Kelso candidly says "Dr. Dorian, do you not realize that you're nothing but a large pair of scrubs to me? The only reason I carry this chart around is so I can pretend to remember your damn names!". He also points out that uninsured patients should be kicked out of the hospital, failing miserably to keep the Hippocratic oath that every  doctor is supposed to make at the beginning of their career.

If you are wondering what a smart hospital should be like, well, it's basically a 180-degree turn from the aforementioned scenario. Paper documentation to coordinate human resources? Poor communication between management and staff? Patient neglect? These are vestiges of a past where decision-makers in the healthcare industry were reluctant to embrace medical device software development and, even when they did, were not necessarily able to blend and harmonize such tools with standard clinical processes.

However, things are changing rapidly under the pressure of several socio-economic dynamics, which was effectively pointed out by McKinsey in its 2019 Finding the future of care provision report:

• The transition from pure disease treatment to health management, a more comprehensive discipline covering prevention through regular screening, well-being (which includes a healthy diet), and rehabilitation.

• Greater education, awareness and proactivity of patients, which translate into their willingness to be involved and better informed about treatment or hospitalization options.

• A consistent effort to improve treatment accuracy with robotics and analytics and therefore minimize the possibility of misdiagnosis or even healthcare-related infections.
• The process of healthcare retailization, which replaces stand-alone structures with networks of geographically distributed but interconnected facilities (such as community clinics, pharmacies, or rapid test laboratories within retail stores) to delegate parts of the hospital workload.

• Increasing attention to healthcare's financial aspects to avoid growing deficits and the resulting adoption of technologies to streamline clinical processes (via EHRs, automation, etc.) and cut costs without compromising on the quality of services.

One of the most promising solutions to better address these trends (not to mention the effects of the global pandemic) implies the gradual shift from conventional to smart hospitals. Let's better define their nature and find out what's under their hood, both from a technological and operational point of view.

A guide to making your hospital smart

Since we just talked about investments, let's take a look at the key technologies healthcare institutions will need to invest in when shifting from old-school facilities to smart hospitals, along with some real-life examples of their actual deployment on the ground.

1. AI for intelligent healthcare

We'll start this overview with artificial intelligence, which is arguably the most transformative technology for healthcare, according to Deloitte.

But really, there is no surprise, since AI is a boundless realm, vaster than doctor Kelso's misanthropy and encompassing or directly affecting a broad range of sub-branches and related disciplines. Therefore, its applications are as numerous as they are impactful. Here are some of the major ones in a smart hospital scenario.

• Computer vision: Combined with machine learning-based anomaly detection and pattern recognition, AI-powered computer vision can easily identify signs of potential health complications from radiological images or other clinical sources and complement physicians’ expertise for more accurate diagnoses. In this regard, the Cambridge University Hospitals partnered with Microsoft's InnerEye team to develop a computer vision system that automates the tracing of tumors and streamlines radiotherapy planning.
• Natural language processing: NLP is another subfield of artificial intelligence worth deploying in smart hospitals in a variety of AI use cases. It's commonly used to turn medical documents into a digital format and store them into clinical databases, but also to recognize specific concepts in written texts via NER (named entity recognition) and therefore categorize such content for easier cataloging. The Australian e-Health Research Center, for example, created an NLP tool to digitize free-text medical data from pathology reports and better monitor cancer incidence trends.
• Chatbots: These represent a (very talkative) embodiment of AI strictly related to the previous point, as they leverage healthcare-specific NLP and machine learning algorithms to mimic human communication and interact with clinical staff and patients. Chatbots and virtual assistants can support smart hospitals' employees in performing several tasks (Providence St. Joseph Health in Seattle uses them to streamline call center routing) or provide non-stop care to patients in need (such as the virtual nurse Molly, deployed by the British NHS to monitor and assist people with chronic diseases).
• Robotics: Physical robots are chatbots' counterpart in the real world. While robots are not necessarily powered by artificial intelligence (such as human-driven machines for remote surgery), AI-related cognitive technologies like NLP and computer vision have vastly expanded their interaction capabilities. Nowadays, smart hospitals can count on AI-powered robotic surgeons for unassisted operations and robotic helpers to perform routine tasks such as lab sample delivery or room cleaning. For example, have a look at Moxi, a tireless and quite expressive assistant deployed at Texas Health Presbyterian Hospital in Dallas.

2. Less paperwork with RPA

RPA (robotic process automation) bots might be less chatty than the aforementioned robots. However, they have still proved extremely useful in a smart hospital setting, as they can be programmed to replicate (or even learn by themselves, when powered by AI) human interactions with software applications and therefore replace or assist the clinical staff in a broad range of time-consuming clerical duties. The applications of RPA in healthcare include patient appointment scheduling, medical records updating, self-service triage, claims management, and more.

Max Healthcare, a major clinical network in North India, deployed this technology in its 14 hospitals to speed up claims processing and data reconciliation, leading to 50% reduction in turnaround time. Their RPA solution can gather data from emails and PDF files, turn it into CSV format, and enter it into their databases.

3. Data analytics is physicians’ second sight

We mentioned the role of machine learning in fueling cognitive technologies that allow medical devices to see, hear and speak (almost) like humans. But this powerful branch of AI also has a lot to offer in terms of healthcare analytics.

Smart hospitals rely on machine learning algorithms, specifically on their pattern recognition and anomaly detection capabilities, to constantly oversee both patients' health conditions and medical equipment's operation, identify risk factors, and prescribe fully personalized therapies or targeted interventions. The medical datasets required to fuel such algorithms can come from several sources, including EHRs previously stored in the hospital databases, lab tests, PGHD (patient-generated health data), and health wearables, previously collected and processed with a data lake or a data fabric solution.

An example of data analytics in a clinical scenario comes from Saratoga Hospital in New York, which implemented healthcare predictive modeling with a solution monitoring patients’ vital signs and identifying those whose conditions are deteriorating. Since its adoption in 2015, this system has reduced patient transfers to intensive care units by 63%.

4. The internet of (medical) things

Considering that, as described above, two cornerstones of smart hospitals are their data-centric approach to medical care and the "sparse" nature of their clinical workflows, the importance of a technology like the Internet of things shouldn't come as a surprise. While data analytics systems are the brain that processes the data, healthcare IoT devices represent the hands collecting data on the ground (typically paired with cloud computing for data storage and sharing).

Health wearables equipped with sensors, GPS asthma inhalers, video conferencing tools, or even smartphones with mobile health apps can easily collect clinical data and patient feedback, allowing smart hospitals to provide remote patient monitoring and consulting 24/7 - both within their facilities and outside their standard geographic reach. In this regard, the Johns Hopkins Hospital at Home, Baltimore, developed an advanced telehealth program designed for elderly patients at risk for nosocomial infections, ensuring better clinical outcomes and cost savings of up to 30% over standard hospital care.

IoT-based telemedicine, however, can also be applied for psychological support. Built by Itransition's team, a remote patient monitoring suite allows expert nurses to communicate with sexual assault victims living in rural areas and perform forensic examination via high-resolution smartphone cameras.

Moving from IoT for patients to IoT for technical equipment, it's worth mentioning that this technology has also been implemented in several smart facilities (for example, at OLVG hospital in Amsterdam) to streamline asset management with Real-Time Location Systems (RTLS). These solutions allow the clinical staff to keep track of medical devices via radio frequency identification after placing location-aware tags on such items, remarkably reducing equipment search time.

5. Virtual tools to tackle real challenges

The last smart hospital tech trend in our brief round-up is arguably the most picturesque as well, namely augmented and virtual reality. From a physician's perspective, AR and VR can be exceptional tools to enhance medical training with highly interactive case-based simulations. Not to mention the opportunities opened up by AR-guided surgery in the operating room, where smart glasses can overlay 2D or 3D projections over the patient body to better focus on their anatomy without having to pay attention to different monitors.

On the patient side, instead, AR and VR represent effective training solutions to exercise in immersive virtual scenarios during rehabilitation. Furthermore, they proved to be a valuable source of distraction to relieve chronic pains, as shown by a 2019 study conducted at Cedars-Sinai Medical Center in Los Angeles.