Electronic health records don’t leave much room for nuance: On paper, two patients with the same type of knee problem might score the same on a standardized walking test, and have tried the same treatment options. But once they leave the clinic, the same condition could impact each of their lives in considerably different ways.
“If disability is an interaction between a person’s health condition, and their environment, and the personal choices they make, we’ve got to understand that full context,” said Amy Houtrow, a pediatric rehabilitation doctor and professor at the University of Pittsburgh.
Technologies like natural language processing can sort through a sea of health data to extract vital information about a patient’s day-to-day experiences and challenges outside the doctor’s office that might be overlooked. These tools also show promise in helping to amplify the voice of people with disabilities, putting their own expertise about their health at the heart of their care.
STAT spoke with Houtrow about opportunities in conversations between doctors and patients and through technology to enhance how the lives of people with disabilities are represented in their medical records, and other areas where technology and disability intersect. The conversation has been condensed and lightly edited.
Can you talk about the challenges with capturing disabled people’s lived experiences in clinical encounters?
I think the first reason it’s very hard to capture it is, we’re very limited in the amount of time people can spend in clinical encounters. And we try often just to get to the point— to the piece of information that’s going to guide us to make a recommendation for treatment. We’re not often asking about all this contextual information. An example would be somebody who has come in with asthma: We really want to find out what’s going on with the symptoms of their breathing trouble, right? So we can fix those symptoms.
Asthma is actually really strongly related to things like pollution. So children, for example, who live near busy highways have higher rates of asthma than children who do not, but we aren’t often asking for all of the things that would contextualize their asthma. Also, in part, because as a doctor, I’m really not in a position to change where somebody lives. And knowing where they live may or may not have any relevance to what the treatment is that I can offer them, because treatments are somewhat limited.
… I use asthma as an example because asthma can actually be very disabling. You are really having trouble breathing, you know, your lungs aren’t working, you can’t (let’s say) go up a flight of stairs, or when you’re wheezing like that, you can’t mow the lawn or vacuum the house. It can be very limiting in terms of what you’re able to do. So even though we think of asthma as just like a temporary breathing problem that people can use an inhaler for and take medicine for, it can be, in fact, very disabling. … When you get an asthma attack, it can get worse and worse, and you get sicker and sicker if you weren’t able to get the kind of care that you need quickly. So when you’re in the site of care, when you’re in the doctor’s office, or wherever, it’s like, let’s start trying to solve this problem. The problem that we have is this breathing problem. It is contextualized by a whole bunch of other things. But that’s not the focus, so I think we often miss opportunities. …
And so asking questions about, well, what makes it easier for you? Or what sort of things would help you be able to do this better? It allows that person to talk to you about what those things might be. “Well, it would make it really easier if I had a scooter to get around instead of because I get really tired when I walk far, I don’t have enough endurance.” And then the person might not know, like, oh, a scooter would help. They might not know what the fix is, they might say, on questioning, “Well, the thing is, I can’t really go to the grocery store anymore because I get so tired when I’m there and my back hurts so badly I find myself leaning over on the shopping cart the whole time. So I’m having trouble making sure that I can have well-rounded meals,” or something like that.
And what are some of the challenges with conveying that complexity of their lives in medical records?
The more voice that we can give patients in our medical records, I think, the better. No one is a better expert on themselves than the person themselves. We often do not have enough of that expertise placed into our medical record.
If you read medical records, they’re often very dry. They don’t have a lot of nuance or explanation, and that’s for efficiency. They’re supposed to tell you what you know or need to know. And honestly, there are times when someone shows up in the emergency room with a bad asthma attack and I want to know what they had last time they had an asthma attack, what sort of medicines, what other health problems, that sort of thing. I don’t really want to spend a lot of time trying to read through a story about, you know, their home on a busy street near a highway, like I want to cut to the chase in an emergency.
So it’s like balancing what makes clinical care as efficient as possible with what helps us really understand what’s going on for individuals with disabilities. I think there’s always a place for us to figure out how we can have both, how can we be most efficient and also understand the patient with a disability’s situation better, so that we can help them be as functional as possible and do the things they desire in life.
You and your colleagues recently published an article on how we can use natural language processing technology to create a more complete picture of patients with disabilities’ experiences. What can we bring to the surface with the help of this technology?
When I take care of kids, they often tell very elaborate stories. I’m a rehabilitation doctor and might ask about, you know, how are things going with your walking and that might not show up very evidently in this like two-paragraph story that they just started telling me about what happened at the playground, so natural language processing can help us pull out information that’s relevant and more specific. And I think those stories tend to be very rich with information. Simultaneously, we don’t want to overburden the medical record system in such a way that we can’t get the information we need.
But if I want to find the issue about walking, in this case, from those stories, technologies that really help us sift through and through information in a standardized way can be very, very helpful and help us be more efficient because there’s a ton of gems in there, right? So that kid telling that story might actually be illuminating to me why it has been hard for them walking, it might be endurance or it might be that their legs get really tired after they’ve climbed up the ladder at the playground twice, that help me take better care of them. And in that illumination, if we can think about on a grander scale for many more people, using technologies to help get at that actual pertinent information in an efficient and effective way, while giving patients the voice in their records to do that, feels really remarkable to me.
I think it’s really important for people with disabilities to be understood as human beings, fully, because our society is pretty ableist, meaning it discriminates and devalues the lives of people with disabilities. As a patient, you want a doctor who you can trust, who understands what’s important to you and your values. And in a world where we have in the record more of the patient’s voice, that gives providers, doctors, and their patients a potentially much stronger bond that hopefully forms a better trusting relationship and allows us to collaborate on decision making for treatment or management plans.
What do you think the health care system sometimes misses at the intersection of disability and technology?
Technology has given people more freedom if they had access to it. … But one of the things that bothers me is, although many technologies have clear benefit to addressing disability, oftentimes we see those devices not being prescribed or utilized. For example, people unable to verbally communicate or use sign language can use a device like an iPad to communicate by pressing a button that says, “I need to use the bathroom,” or, “I would like cereal for breakfast.” Insurance might cover this program for your communication needs on the iPad, but not the iPad itself. Or a wheelchair that really allows you to function in the winter. So I’m very interested in how we navigate the advances in technology in general, how they can be optimized for people with disabilities, and then utilized, effectively covered by insurance because insurance often doesn’t cover a lot of things that seem like accessories. …
There is a ton of technology that ends up serving people without disabilities very well that was originally designed for people with disabilities. [Houtrow pointed to curb cuts, which make it easier for people using wheelchairs to cross the street but also benefited people pushing strollers], and technologies that were breakthroughs in general that ended up also being big breakthroughs for people with disabilities. Other technological advances end up being around convenience for people who don’t have disabilities but are really imperative for people with disabilities. [Houtrow highlighted environmental controls such as remote-controlled light switches or window shades, which were built in service of “smart homes” but are valuable for people with certain disabilities.]
I think we should be very excited that the technology can continue to help foster people participating as they desire in society if they have disabilities but be really cognizant of all of the barriers for that actually happening, because there are a whole bunch of societal variables and with the way things are structured, like that people with disabilities tend to be living in poverty, so it makes it really hard to buy the fancy new device. And if insurance doesn’t cover it, then you might not get it. So it might be something that could really help you that is just out of reach. …
Even something as fundamental as eyeglasses is distributed inequitably among the people who need it. And another, I think, really important example is being able to buy hearing aids over the counter. When we think about those technologies, the more advanced they get, the cost of them comes down. And as the cost comes down, it becomes more accessible to more people, and I think that’s another aspect of how we consider the advancement of technology and how it benefits people with disabilities is, as costs decrease, as technologies become more common, they become more accessible. Finally, the technology is cheap enough that it’s more readily available. But we would like not to have to wait [for] that point, right? We would want the cutting-edge technology that really helps people with disabilities to be available for them now. Not 10 years from now.