How EHR Adoption is Becoming a 'Moral Imperative' for I/DD Providers

In my previous column, I discussed facets of Arthur Webb's 2022 book, "Dangling on a String: The Future of Public Policy for the Field of Intellectual and Developmental Disabilities in New York State." I predominantly focused on his broader thesis that the support net of services for those individuals with intellectual and developmental disabilities (I/DD) and autism are "dangling on a string." 

In this column, I want to further discuss two areas Webb focuses on in his book: COVID-19 and managed care. More specifically, I will explain how the adoption of electronic health records (EHR), coupled with the use of artificial intelligence (AI), is becoming more of a moral imperative if we hope to keep that string from breaking. 

Why the COVID-19 Experience Fuels the Need for Electronic Records 

For my discussion about COVID-19, I'll highlight a portion of Webb's chapter on the topic, which he titles "The Virus." In the chapter, Webb speaks about the New York Integrated Network for Persons With Developmental Disabilities (NYIN). NYIN is a specialty insurance company for people with I/DD. What NYIN did very early on in the pandemic was launch a data project. This project was intended to monitor and measure how the COVID-19 virus was affecting the individuals the network supports. 

Webb cites that, at the time he was writing the chapter, the rate of COVID-19 infection for those in New York City group homes was five times the general population, the rate of hospitalization was almost three times that of New York City in general, and the rate of death was twice the rate of New York City. More than 200 homes had experienced COVID-19 cases, which represented 40% of all the residential settings. (Note: As New York City was experiencing these incredible effects, I was managing a company called Chimes International serving 24,000 unique individuals from Virginia to New Jersey in the mid-Atlantic. Our job was to keep people alive. COVID-19 was a challenge of a lifetime. It is impossible to describe the tensions that were omnipresent for more than a year as this virus progressed.) 

How do these figures from New York point to a need for further use and adoption of EHRs in the I/DD sector? It's important to acknowledge that general physicians are not often as familiar with treating people with I/DD and autism. These are highly specialized areas, and there are very few people — unfortunately, too few — who dedicate their lives to these populations. 

Now it's important to picture what happened as these group homes experienced their rapid rise in COVID-19 cases. I remember speaking with a colleague at the time who was running the first chapter of The Arc in the country known as AHRC. He was telling me about how he was trying to help those adults with I/DD who caught COVID-19 get admitted to hospitals by sending nurses to accompany these sick individuals. But the flooded hospitals were flat out refusing to admit these sick adults, even with the nurses willing to help with admission and care. 

I recall a powerful video featuring ADAPT Community Network (formerly United Cerebral Palsy of New York City) CEO Edward Matthews. He went on the record stating that when COVID-19 hit and the healthcare system was overwhelmed, his organization's clients essentially wouldn't be seen. This was despite the fact that those with I/DD were dying at a much higher rate than others. 

Why did those with I/DD struggle so much to receive care? While it's important to acknowledge that part of the challenge here was that hospitals were overflowing with patients, this wasn't the only factor. Another significant contributor goes back to what I said earlier about the lack of physicians familiar with treating those with I/DD.  

During the pandemic, it wasn't feasible to have a nurse, other healthcare professional, or caregiver arrive to the hospital with 20-pound binders full of papers gathered over the past 25 years of an I/DD individual's life — papers that could detail the patient's correlated or comorbid health conditions, medications, treatment history, and more — and present that to a care team for review prior to administering treatment.  

Hospital care teams lacked the time and resources to review and digest the breadth of paper records associated with I/DD individuals. If I'm a physician who has an individual come in who has Down syndrome, is 55 years old (average life expectancy of individuals with Down syndrome is 60 years), and has been on various medications to control his behavior for years, and I cannot easily and quickly access that information, how do I know how to best treat this patient? Unfortunately, during a pandemic or incident affecting numerous people, such a physician may be inclined to lower priority on treating this adult with I/DD and instead move on to first care for the patient where they believe they are more likely to find success administering appropriate treatment.  

If this individual with I/DD came from a home with an electronic health record that could provide the hospital access or at least enable a supporting physician to download some of the reports from the record, the barriers to supporting this patient — and doing so effectively and efficiently — would be greatly reduced.  

This isn't just the case for a pandemic. This is for all healthcare encounters, and especially those where time is of the essence, such as emergent and urgent care. If you're using paper records, it's unreasonable to expect that you could quickly provide the necessary background information and details on an individual with I/DD to help expedite appropriately care. 

Why the Future of Managed Care Needs Electronic Records  

In Webb's chapter on managed care, he notes that in the world of government payment systems, everything progresses to the mean — an observation I share with him. It's simply easier to deal with averages. What do I mean by this? The tendency of government is to give preference to commonness not differences. As an administrator, it's less work for me to deal with the majority who fit into the bell curve and push away those on the outside who may require more work or may represent greater uncertainty concerning treatment success.  

Such a mentality should raise a flag for everyone involved in our healthcare system. The problem with this approach to care is the people that we're serving in the I/DD space do not fit within the large percentile who fall into the center of the bell curve. Those with I/DD are on the outside and very often require specialized care. 

Unfortunately, we've seen this extensively in the managed care space. There are very few positives associated with managed care and I/DD. For every dollar a managed care company gets to treat somebody with I/DD, it costs them between $1.05 to perhaps $1.15. That may not sound like a lot until you scale it up to millions of people.  

Traditionally, managed care companies have not focused on the I/DD population. They may think an opportunity exists to profit on care for them by introducing new efficiencies and economies of scale. The residential cost per individual with I/DD, per bed, per year in New York State is very expensive. These costs take into account all the costs associated with staffing, real estate, insurances, utilities, etc. Managed care payers may look at those high numbers associated with residential care and believe they can find ways to save lots of money, which will generate profit.  

The problem with this mindset: My experience shows me that those costs have been squeezed, squeezed, and squeezed some more for more than a decade, probably close to two decades. If you can save even 2% at this point, I would consider that impressive since providers have been trying everything they can to bring down costs for a long, long time.  

For managed care to work, and Webb refers to this in his book, it needs to move away from flat fee-for-service reimbursement and toward quality. In other words, I pay you more because you've done an incredible job helping my kid with developmental disability to be more independent, lose weight, be healthier, and achieve other health goals. You earn this value-based payment premium because you reduce the need for more costly treatment. Webb says any efforts to reduce fragmentation and encourage integrated systems are good for people with special needs. That is true of any system built around supporting and delivering quality care. 

This movement toward value-based payment has thus far generated little traction, despite its apparent benefits. I believe a big reason why the system is so stressed is that it can't entertain the shift in management and metrics that would be required to administer value-based payments. However, if we ever hope to see momentum build toward such payments and a system that's more inclusive of all patients, not exclusive of some, we'll need the ability to better manage care and then track and report the quality measures that serve as the basis for the premium payments. These are simply not achievable on paper records and further point to an essential need for greater EHR adoption within I/DD. 

Why Artificial Intelligence May Help Solve Some of the Biggest I/DD Challenges 

I want to conclude this column by speaking about the use of artificial intelligence within an EHR. We're beginning to see some eye-opening advancements in this area that may greatly benefit our field. I recently had a chance to view a demo of some AI functionality and wanted to speak to two applications that I think would make a lot of sense for I/DD and autism. 

The biggest crisis in our field is staffing. It's safe to say there's a staff vacancy rate across our system of around 35%. We know it's worse in some areas, with up to 50% vacancies that become so difficult to overcome that agencies are forced to shut down. The best agencies are getting by with a 20% vacancy.  

It should come as no surprise that we're struggling to recruit and retain staff. I think back to my time at Chimes International when I was working in Fairfax County, Va. We were paying staff around $11 an hour for direct care service, and in comparison, down the road, a fast-food restaurant was paying $17.  

For $11 an hour, we were asking a person with limited skills, often limited English proficiency to come in, work in direct care, and meet the state's huge number of rules and regulatory requirements. The expectations of the position were very high when compared to the compensation. The new direct service staff needs to capture data like they were an analyst, write approaching John Steinbeck to explain a client's progress, and do all this work with effective use of B.F. Skinner's behavioral knowledge and understanding — all for $11 an hour!  

The good news is AI can help overcome some of these seemingly insurmountable challenges. One area concerns verbal interaction. Many of our staff have limited English proficiency and health literacy. However, they usually can say things like, "Michael had a great day and completed his goal on learning how to brush his teeth." This is an important I/DD data point, and an AI component can capture, track, and integrate it into the treatment plan to indicate progress.  

Use of speech-to-text functionality has great promise. Combined with AI, we can help reduce the documentation requirements for I/DD staff so that they can use the skills they have to succeed at the job and motivate them to want to continue working. 

Another application for AI that I find fascinating as a clinician concerns the concept in the treatment of children with autism called "natural environment teaching," or NET. One of the challenges I've always seen with NET is the learner paces the teacher. In your traditional school, a teacher paces a classroom of 30 kids, with the class moving only as slowly as the last few ships in the 30-kid convoy. The teacher needs to adjust the speed to accommodate the entire class, but that generally occurs naturally throughout a school year.  

With services for autism, the learner paces the teacher — a flip of the traditional classroom scenario. To succeed here, you must be a really strong teacher. But there's a major challenge: With NET, you want to continue prompting a child as they're demonstrating skills to motivate them to build on their success, and you do not want to interrupt that flow and momentum.  

Since applied behavior analysis (ABA) is a database science, you typically need to interrupt that flow to capture data. You may be able to provide 5 or 10 minutes of instruction, but then you need to stop and record what took place. That's the nature of the science.  

However, with AI, you can now input several targets as goals for the child: "I want the child to be able to identify water. I want the child to request a cup. I want the child to have a back-and-forth exchange with me a few times." As you work through these goals, AI can capture and document the progress — all without interrupting the flow of the instruction. To me, that capability has the potential to help transform ABA and significantly strengthen our ability to help those with autism. 

I am optimistic about the ways technology like EHRs and AI will strengthen Webb's "string" and hopefully help bring us back from the brink.