Since the invention of the transistor, we’ve been living in the Information Age. Which sounds a bit like we’re enjoying a sci-fi utopia but mostly means that we all have 20,000 unread messages in our Gmail inboxes.
Now as those transistors keep getting smaller, we get devices that can process more and more data, and memory devices that store more data, which follows the well-known Moore’s Law. Every 18 months, we can process twice as much information, or that we can store double the amount of information in the same amount of physical space.
But as humans, evolution isn’t working on an 18 month scale. So while the technical side of working with data is making lots of progress, we haven’t made nearly as much useful progress in understanding all of it.
I wrote another piece about what I called the “Four Cs of Data + Design,” which was a way of framing some of where I think the focus should be in this kind of work. And last week, I spoke about those examples, and then more specifically about how they apply to healthcare, at the Mayo Clinic’s Transform Conference. (Video is here if you prefer listening to reading.)
The transformation of the last twenty years…
Our current moment is a decision point (another theme of the conference) that requires a shift in how we’ve worked with data over the last 20 years.
When I started doing lectures about this kind of work in the late 90s, I had to spend a lot of time setting up the idea that this deluge of data was on its way, and that design (by way of information visualization) was an important part of the potential solution. The data is coming, the data is coming! (A feeble attempt at playing a kind of Paul Revere of data and design.)
Five years later, Google had launched, and people were all too familiar with information overload, but were impatient for answers.
Ten years ago, big data and analytics were going to save us. Your organization is just a data scientist (or two) away from having all the answers! Everyone was wrapping up their dashboard projects. We’re leaving the data warehouses and jumping into data lakes! With enough analytics, the story went, we can make sense of the black box of what’s in all that data.
Now in the last couple years, these organizations, having stocked up on data crunching and analytics capacity, still find themselves stuck.
Where’s the insight?
We know this because this is usually the point at which a potential client will get in touch with us: they were sold on something like Tableau and have just figured out it’s not giving them the answers.
And in the meantime, the hype cycle has shifted its attention to machine learning and artificial intelligence. Maybe the machines can figure it out for us! These are really powerful tools, but the problem is that now you don’t have just one black box you’re trying to figure out, you’ve got two! Why did the AI make that recommendation? It may work great for certain kinds of first order problems, but you still can’t generate insight from algorithms. Never could. People create insight, and the focus should be about giving them tools that help them develop these insights.
Healthcare and the next twenty…
Thinking about where this puts us, let’s now think about the next twenty years, and apply the four Cs to priorities around understanding data in healthcare. Those aren’t the only four categories to consider, but I think they’re helpful for focusing on a few threads that are really important.
It’s about people communicating. The primary focus of EMRs should be about communication. How do patient and provider communicate? How does the provider convey results to a specialist? How do multiple specialists work together on a problem, as represented by a single patient’s data in that EMR?
We’re social animals, and we need better ways to socialize a set of information amongst all the people who need it. This shouldn’t be a tacked on feature like an email inbox embedded in an EMR application, it should be the central focus of how people work.
You look outside healthcare and you have tools like Slack, which is “a real-time collaboration app and platform.” But that’s just a 15 or 20 billion dollar way of saying “glorified chat client with document sharing.” There’s a reason they’ve been really successful: simplifying communication matters.
This is about data availability and new capabilities. So for instance, a promising recent development is what Apple is doing with its the Health Records API, where medical records are cracked open ever so slightly, on an individual basis. With it, users can access parts of their health record directly on their personal mobile phone.
But the availability of an API like this means a small studio like ours can start pulling data from providers and building apps based on that data. We don’t have to do that within an App Orchard the way that EPIC wants us to, we don’t even have to partner with a larger institution who has access to patient data, but we can start digging into this data and thinking up our own patient-centric ideas about what to do with it. (I’ll save the ideas for another post.)
That’s a significant shift of power, but for me personally, I get most excited about how changes like this can unlock a lot more of the creativity that’s out there.
This is maybe the most straightforward. Everybody’s dealing with too much data. Drowning in it. I have never, ever, ever met someone who doesn’t feel overwhelmed by the amount of information they deal with every day. Too many emails. Too many journal articles. A very busy internet. Too much data jammed into patient records.
So the ability to condense all that, to bring hierarchy to it, only grows in importance. But here’s the thing… We’ll never have less data. There’s no going backwards. Sorry to stress everybody out. For me it’s great because hopefully that suggests some type of job security. I think it’s possible to do drastically better than we are today, it’s just that most people are focused on solving what seem to them like interesting technical problems, rather than putting their focus on making all that data accessible, useful, and actionable for the end users.
We need better tools, and the best solutions are probably much smaller than the EMR. Nobody is going to unseat EPIC or Cerner in the next five or ten years, but there are a lot of interesting things that you might be able to do at the edges to help deliver better patient care.
We worked with a group at Mayo Clinic for whom an Excel spreadsheet was a central part of their understanding of various symptoms and outcomes. We were able to build something far more useful, but if as blunt an instrument as Excel can deliver useful results, I’m absolutely confident that there are far better approaches and ways we can improve.
I like to use these more pragmatic improvements as a way to understand a problem well enough to where we can begin to conceive ways to really rethink problems in their entirety. The problem isn’t the spreadsheet, it’s about how the contents of that spreadsheet change over time, how that information is shared around a department — amongst researchers, doctors, nursing staff on the floor — and how to use that data in a way that better supports patient care.
The title for the panel where this was presented was “The Answer is Right in Front of Us,” which couldn’t be more apt. It’s not a lack of data, or computational power, or hardware solutions, it’s a matter of giving people better tools to make sense of it all. It’s also a bit like the quote from William Gibson: “The future is already here — it’s just not evenly distributed.”