Smarter hospitals
taking healthcare from clipboards to the cloud
Entering a hospital can sometimes feel like you just stepped out of a time machine. Things are being faxed and photocopied, there’s a lot of colored papers, clipboards exist (and not in a camp counselor way). It’s rarely ever a frustration-free experience, especially if you are the patient. Unfortunately, that theme continues as you step into a hospital room that is lacking any semblance of technological efficiency – Gen Z’s worst nightmare. If you were a 90s kid, you’ll probably remember the movie Smart House. That is the kind of AI we were hoping to see and interact with in our lifetimes, even in hospitals. It’s not that far off to believe that hospitals of the future will be really smart, just like Pat. In fact, some hospitals have already rolled out smart hospital technology, which can include features like virtual sitting, telemedicine access directly from hospital rooms, increased use of remote patient monitoring devices, and ambient technology.
Smart hospital technology is intended to improve the quality of care being delivered while decreasing provider burden by automating patient check-ins. There are also smart hospital applications for operating rooms, which include the use of cameras for physician consultations during surgery (companies like Surgical Safety Technologies are working on building these tools).
Today (in most cases), when you are in a hospital room someone will come in to check on you, take your vitals, and make notes in your clinical record. At no point do any of these things happen automatically. Some of the goals of smart hospital technology are:
to make sure your vitals are constantly being monitored by remote monitoring devices and that the data is being captured in the clinical record
give you access to a clinical staff member at any time via telemedicine, which shortens response times
movement detection via ambient sensors, including fall detection
improving the patient experience using tablets and smart TVs
Of course, smart hospitals won’t appear overnight (just like how we don’t all have smart houses yet). Currently, several hospitals have started adopting virtual sitting and nursing technology, which requires the use of hardware in the form of a smart camera. There are several venture-backed companies that have set up end-to-end platforms to help hospitals become smart. Examples of these companies include Care.ai (purchased by Stryker in September 2024 after raising $27M), Andor Health (raised a Series A in 2021), LookDeep (raised an unknown amount of money in 2022), and Artisight (has raised a total of $50M+ from various investors, including Nvidia).
All of the products built to empower hospitals to get smart are using differing levels of AI. Virtual connectivity to providers does not in itself require the use of AI, but translating information discussed during patient calls and structuring unstructured conversational data does. Recording patient health information and data collected from wearables is also relatively low lift from a technical perspective, but alerting a care team that a patient might be at risk or needs attention requires a level of clinical understanding that LLMs are being trained to use. Fall detection requires a form of movement analysis and sensor utilization that can be improved by training foundational models so they’re able to more accurately know when a patient falls (Cherish Health built a device that has an ambient sensor platform to detect falls). And of course, on the tablet and smart TV front, we hope your Netflix algorithm knows that you want to watch Smart House after reading this article.
Transcription
Avid voice note senders and receivers will know that Apple’s voice note transcription feature changed the game. You now don’t have to wait until you’re in a quiet location to listen to the 5 minute and 47-second-long voice note your friend sent you about what they did last weekend, because Apple transcribes the details for you to skim at your leisure. You also don’t have to try to remember every last detail when crafting your response (“who did they say they ran into?” “what cafe did they say was good?” “how many licks does it take to get to the center of a Tootsie pop?”).
For providers, transcription offers a similar but probably more meaningful convenience. The utility of transcription services for a busy primary care doctor who is seeing 20 patients a day can have serious implications on quality of clinical data capture and even for billing purposes. Products built by transcription companies can have a wide variety of components, some of which are actually incorporating the use of advanced AI. AI doesn’t impact the quality of a recording or the speakers’ voices. Where it comes in is in the translation of the transcription into real, clinical insights. It can recognize medical terminology and create visit summaries for providers to quickly access at a later point. It can write clinical information directly into the EHR, taking away the time your PCP spends sitting on their rolling stool clicking away at an interface that looks like it was built in 2003. The AMA published an article citing time savings of ~1 hour a day for physicians using an AI scribe tool in a rollout at The Permanente Medical Group. AI transcription tools have been the most widely and rapidly adopted in healthcare to date, for good reason. Using these tools does not come without risk. Providers are still expected to review the notes generated by the technology to ensure accuracy and completeness. Even OpenAI’s transcription tool – Whisper – has been found to hallucinate, fabricating or capturing inaccurate information.
There are several companies that have created various levels of AI transcription tools. Some of the largest companies include: Abridge (Epic partnership, raised $494M), Ambience Healthcare (raised $101M), Nabla (raised $33M), Suki (raised $165M), and DeepScribe (raised $37M). Of course, Microsoft is also in the game with Nuance’s DAX Copilot. The market is quite competitive, but this is a technology that many health systems, provider groups, and even small independent practices want to use. The best products will be the ones that are the most clinically accurate and are able to seamlessly write notes back into patient medical records.
Consumer tech: do you want to see my data?
Devices that enable consumers to capture and analyze their own health data have become increasingly popular. Wearables like WHOOP (raised over $400M), Oura Ring (raised over $900M), Apple Watch, and Fitbit (acquired by Alphabet), once marketed as “fitness trackers” now allow people to access their health information continuously. These wearables not only track physical activity but also monitor sleep patterns, heart rate variability, and other health indicators, giving users a detailed overview of their health status in real-time.
Consumers are increasingly bringing their health data to the doctor’s office, but the technology has not been integrated into care workflow s yet. If synced with EMRs, data from these wearables can be incredibly beneficial to patients and providers. Clinicians would have access to dynamic and comprehensive patient profiles, and would be better able to personalize treatment plans, enhance diagnostic accuracy, and potentially predict health crises before they escalate.
I can see CLEARly now, the form is gone
When contemplating a seamless, end-to-end AI-driven experience at the hospital, the focus largely seems to be on the examination / procedure room or pre-arrival patient navigation and scheduling. What about the moment the patient actually walks through the door? Anyone who’s been to the doctor’s office has experienced the dreaded clipboard with ten pages of information and medical history required. Not to mention having to write today’s date and your signature approximately 17 times. It’s frustrating, particularly because it’s recurring, and it’s exactly the kind of thing that could and should be solved for with technology.
CLEAR, known for its presence in airports and stadium venues, has made the foray into healthcare. They have stated their mission as a single, universal health identity that can save patients and providers time during patient intake. CLEAR has partnered with University of Miami Health, Wellstar Health, Rush University System for Health, and other healthcare institutions to consolidate patients’ medical information into one digital ID, using a single sign-on, HIPAA compliant account.
CLEAR has also established partnerships with InterSystems and Epic, further pushing for fully integrated patient data management. By integrating their biometric identity technology with InterSystems Health Gateway, they intend to make medical data more easily accessible to patients and to streamline check-in and pre-registration processes. Through their Epic MyChart integration, CLEAR is expediting registration and on-site check-ins, working with Wellstar to launch check-in kiosks. The process involves patients registering with CLEAR in advance of their appointment, then checking in at the kiosk via a selfie. Other partnerships with Verato, a patient-matching platform, and b.well Connected Health signify the efforts to truly reconcile fragmented patient data.
While this sort of biometric technology integration easily fits the Smart Hospital criteria, it does raise questions about patient privacy and security. CLEAR ensures that they are HIPAA compliant and providing IAL2-verified access to electronic health records. Because only authorized users can access patient data and identity verification can be done with biometrics, CLEAR claims the risk of fraud significantly decreases. They currently have 31 million users, making it even more seamless for those patients who are already registered. Integrating CLEAR’s technology into the patient and provider experience feels like a long-awaited futuristic move for a healthcare system that relies too heavily on obsolete practices. And if it makes it easier to get into the stadium for a game or a concert, well, that’s just gravy.
Data is the new oil concrete
Just as no one would build a house without a solid foundation, implementing a smart hospital system requires careful consideration and strategy surrounding the underlying infrastructure. While things like modern networking infrastructure, power distribution and hardware availability are table stakes for this conversation, the same thoughtful design and implementation needs to be had surrounding the system’s data assets as well. Without this, there will be nothing smart about it.
Skipping these basics is like building a house on sandy ground – it might work initially, but problems will eventually surface (and they may be catastrophic).
Building a Robust Data Foundation
Okay fine, we get it; we need to focus on the data platform. But what does that even mean? What makes a good data platform for a health system to build all these shiny, smart hospital features on top of?
The platform needs to be resilient, scalable, and flexible to accommodate the continuous inflow of data from various sources such as patient wearables, EHRs (Electronic Health Records), and other IoT (Internet of Things) devices. (Oh, and it also needs to support integration with the most popular database of them all, Excel.)
Decisions need to be made surrounding where this platform ultimately lives: in the cloud, on-prem, or deployed using a hybrid model. Generally speaking, you will find the hybrid model wins out and allows you to take advantage of each of the paradigm's strengths.
Another big decision to make here is the classic buy vs. build debate. Is this platform something you will build internally from scratch? Or will you go out and purchase a piece of software to solve all your problems? There are plenty of vendors in this space, for “turnkey” solutions like Innovaccer, SG Analytics, and N1 Health. Balance is key here, and a hybrid approach of buying components and stitching them together is generally where we see folks going.
Now that you have chosen your architectural framework and begun to amass a corpus of components (and maybe even begun to build some of your own), you have a very real problem that must be solved with talent and hands-on deck: integration. This is the straw that breaks most of the camels’ backs when it comes to data platforms at a health system. Oftentimes integration is forgotten about or glossed over, when in reality, it should be top of mind throughout the whole process as no matter your decisions to this point, integration will likely be your long pole in the shed.
As smart hospitals integrate increasingly sophisticated technologies like wearable health devices and biometric identification systems, they must thoughtfully cultivate strategies surrounding data architecture.
Cost Considerations
The financial burden of implementing smart hospitals mirrors the complexities of building and maintaining a modern home (admittedly at a much different scale). Just as homeowners face significant upfront costs for construction, hospitals must make substantial initial investments in hardware, software, and licensing - essentially laying the financial foundation. Ongoing expenses, similar to home maintenance and utilities, come in the form of continuous system updates, security patches, and infrastructure upgrades along with maintaining a skilled staff to support it. Perhaps most importantly, like teaching a family to effectively use and maintain their new smart home features, hospitals must invest heavily in staff training and continuous education to ensure optimal system utilization.
These costs can be substantial, but like adding a master bathroom or redoing your kitchen, smart hospital investments often pay for themselves over time through enhanced operational efficiency, improved patient outcomes, and reduced long-term expenses. The key is developing a comprehensive financial strategy that balances immediate costs against long-term benefits, ensuring the investment yields sustainable returns for years to come.
Patient Privacy and Regulatory Compliance
Data privacy and regulatory compliance serve as the non-negotiable foundation of any smart hospital infrastructure. Like a sophisticated home security network with multiple layers of protection, hospitals must implement comprehensive safeguards that align with stringent regulations such as HIPAA in the U.S. and GDPR in Europe.
This involves implementing bank-vault level encryption for all patient data, whether it's stored locally (commonly referred to as “data at rest”) or traveling between systems (commonly referred to as “data in transit”). Access controls act as the equivalent of an advanced key card system, ensuring only authorized personnel can enter specific "rooms" of data, while regular security audits serve as the vigilant security guard, continuously monitoring for potential vulnerabilities.
This multilayered approach to privacy and compliance isn't just about following rules - it's about maintaining the sacred trust between healthcare providers and their patients, much like maintaining the security that makes a house feel like a safe home.
Using AI and Analytics Responsibly
Speaking of trust, now that the health system is beginning to amass a treasure trove of data, it is paramount that the data is used ethically. Integrating AI into smart hospitals is like adding a highly sophisticated but carefully monitored automated system to a home - while it can dramatically improve efficiency and effectiveness, it requires thoughtful implementation and constant oversight.
Just as a smart home system needs to work reliably for all family members regardless of their characteristics or preferences, AI in healthcare must be developed and deployed with careful attention to fairness and inclusivity. This means maintaining complete transparency about how AI algorithms make decisions and ensuring these systems are trained on diverse datasets that represent the full spectrum of patient populations. Like a home security system that must protect all residents equally, AI systems in healthcare must be rigorously tested and monitored to prevent biases that could lead to disparate treatment outcomes. The goal is to harness AI's powerful capabilities while maintaining the human-centric, ethical foundation that healthcare is built upon.
Conclusion
The transformation into a smart hospital isn't just about technological upgrades - it's a complete reimagining of how healthcare is delivered, experienced, and sustained. From ambient sensors to AI scribes and biometric check-ins, the tools to modernize healthcare are already here, and adoption is steadily underway. But these innovations must be built on strong, scalable infrastructure, governed by rigorous standards for privacy, integration, and responsible AI use.
Smart hospitals promise faster, safer, more personalized care, while also easing the administrative burdens on clinicians. Still, the journey requires more than just shiny tech. It demands thoughtful investment, collaboration across vendors, and an unrelenting focus on the patient. If done right, tomorrow’s hospitals won’t just feel smarter, they’ll actually be smarter, serving as a new standard in care that finally feels as modern as the rest of our connected lives.