Long Term Care IT Innovation: AI & Interoperability

We've touched on some of some of the threats we’re facing in long-term care with staffing and cybersecurity, let’s talk about some of the technology coming that could revolutionize the way we see healthcare.

AI

The big buzzword. Everybody’s talking about it, but can anyone tell us what it is? Is it Arnold Schwarzenegger in sunglasses and a leather jacket?

Is it R2D2 and C3P0 from Star Wars? 

No one ever knows. Let's narrow it down a little. AI is a general term for the theory and development of computer systems that can perform tasks that, historically, require human intelligence. Did I lose you yet?

6 subfields of AI can help us better realize what is possible and what could be coming. I will focus on 4 of them and how they could apply to healthcare. Something to remember from this is that each subfield can build upon the last; some aspects intertwine; just because something is classified as natural language processing doesn't mean that it doesn't have aspects of deep and machine learning built in.

Machine Learning

At its most superficial level, machine learning equates to algorithms and models that enable computers to analyze data and react differently depending on the data presented. Your email spam filter is an example of machine learning. The model uses machine learning to identify and classify emails as spam or not spam; it learns from specific keywords in the email to guess whether you should receive it.

Deep Learning

Deep learning is similar but much more in-depth. This is where many layers of artificial neurons are created to allow for the solving of much more complex problems. Machine learning is based on specific algorithms; deep learning is made with many layers of those algorithms to create a virtual neuron network, as you would see in a human brain. You see it in some software you use that tries to predict what will happen with your high-risk residents. It's using historical data to make an educated guess on which patients may be at risk.

Natural Language Processing

Natural language processing enables computers to understand and respond to human language. Have you ever used the "talk-to-text" feature on your iPhone? Well, imagine PCC having this feature without the typos, and you get a complete analysis back with suggested interventions based on what you say. Imagine not having to type your clinical notes; you talk, and the computer writes it out for you.

Machine Vision

This is the ability of a computer to interpret visual data. There's already been success in this area regarding retinal scans. AI has detected Alzheimer's and other conditions through a simple eye scan; it's incredible. Studies are happening now on its ability to detect hypertension, diabetes, and inflammation.

This leads us to a term coined by Dr. Eric Topol at Scripps Research: "keyboard liberation," which includes interoperability.

Keyboard Liberation

Dr. Eric Topol, currently the Executive vice president at Scripps Research, is a firm believer that these technological advances will restore the human bond in medicine. He believes that in our current state, there are so many requirements for documenting in so many ways and in so many different places that it takes out some of the human connection that used to be associated with healthcare.

Consider how much time your nurses spend not directly speaking with your residents; think about how much of their attention strays away from your residents because they need to get their charting done or whatever other documentation they need to do. Imagine if a computer could type out their charting while interacting with their residents. You talk to your resident, your resident talks back, and the laptop transcribes. We're talking about AI listening to the conversations between staff and patients and transcribing the documentation. Imagine how much more personable the care would be if we could just take care of the patients rather than worrying about data entry.

Now I know what you're thinking, 'I do NOT want everything my staff says out loud to be in the clinical record.' And I agree, that seems dangerous. There will be other challenges that could come from this. Still, the upside will far outweigh the operational changes that may need to happen. We aren't quite there yet, but when we get there, it will benefit our staff members, clinicians, and, most importantly, our residents. 

Interoperability

Interoperability is the ability of different healthcare information systems to exchange, interpret, and use data cohesively. Let’s take a second to think about the other stops a resident may take in the healthcare system.

A Resident's Journey

This is John. 

John

He’s a very energetic guy. John has his primary care physician he sees in the community, which uses its own EHR system. He, unfortunately, has a fall and ends up in the local hospital, which has another EHR system. 

John recovers from his fall in the hospital, but he now needs some short-term care, so he goes to a nursing home... with its own EHR system. 

After he’s done there, the social worker sets him up with home health for the transition and yet again, another EHR system. 

Oh, and I forgot to mention that at each stop, he’s using a different pharmacy, and possibly therapy which each have a different system… 

Now I know what you're thinking: "you're trying to make it look too complicated. We have referral systems that get us the data".” And I know that, but imagine if, through AI, all 8 of these systems could synchronize their data so that each clinician had access to all of it at each point in John's journey. Imagine not worrying about a missing medication list from a referral packet.

We're seeing the beginnings of this through systems like Allscripts and Epic. But as the tech progresses, more and more of these processes will become automated, which will give you, as operators, more time to focus on your residents rather than your paperwork.

Advancing technologies will eventually allow us to focus less on data entry and more on taking care of people. It will enable us to be more personable and precise simultaneously and help limit the threat of human error. So, we've talked about the coming pie-in-the-sky technology, but where are we now? Stay tuned next week, and I'll be going through some of the main issues we see in a facility, along with some tips to remedy them.