Healthcare Needs Software That Handles a Messy Reality

EHRs took off in the 2000s, and there have been calls for a better user experience ever since. Despite significant advances, healthcare personnel still spend hours a day toggling between screens, reentering data the system already has, waiting for clearance from insurance companies, and attempting to satisfy payer requirements. This is no small undertaking.

Research shows that clinicians devote between one-third and one-half of their work hours to EHR systems. This results in inefficiency that drains more than $140 billion from care capacity each year.

I would argue that the problem is not that companies have failed to make easy-to-use applications. Rather, it’s that there have been limitations to what software can realistically do in a modern healthcare environment. Healthcare is messy. As such, the systems and processes that support it tend to feel messy too.

This does not mean that clinicians will be stuck dealing with clunky platforms for the rest of their days.

The drive for simplicity

One of the oldest ideas in science comes from William of Ockham, about 700 years ago. He posited that when multiple explanations fit the facts, the simplest explanation is best. This principle, which has guided centuries of innovation, is known as Occam’s razor, because metaphorically, a razor cuts away unnecessary assumptions.

Applying this principle to product design, the best products are those that reduce friction, shorten learning curves, and allow users to focus on tasks over tools. They don’t expose their underlying sophistication.

This speaks to the desire for a more intuitive experience. But there is another principle pulling on the opposite end of the spectrum that comes into play.

That which cannot be simplified

Much more recently (the early 2000s), Stephen Wolfram introduced the idea of computational irreducibility, which argues that some systems, even those governed by simple rules, generate behavior that is just too complex to shortcut. The future state or end result only emerges by working through each interaction; there is no jumping ahead. Let the system run, step-by-step. This applies to everything from weather to biology to a patient encounter.

Think of the typical sequence of a patient’s trip to their physician’s office. A single visit involves scheduling, charting, coding, billing and more. Then the clinician’s judgment is applied to a set of incomplete data points, shaped by patient behavior, constrained by payer rules, and passed through the practice’s operational bottlenecks. There is so much happening that it cannot be reduced into a static workflow, no matter how clean the interface sitting on top looks.

This is why software has yet to offer clinicians the relief they seek. The technology industry keeps sanding down the interface while the underlying complexity associated with patient care still requires providers’ input.

A new lever delivers meaningful change

To this point, healthcare software has been dedicated to documenting care as a system of record. AI is now capable of taking on a substantial amount of the requisite reasoning load that makes the system run, alleviating the burden from clinicians.

AI can coordinate workflows, interpret context, surface key information at the right moment, and automate administrative work. A human in the loop is still necessary, but we’re talking about hours of work disappearing each day. Imagine the impact that can have on patient care and clinician burnout.

Now consider the irony that to make software seamless at the surface, it must become far more complex underneath. Magic happens as the orchestration layer grows deeper, as AI models continuously evaluate and adapt to changing context, and as multiple systems communicate with one another to exchange data dynamically.

The mess underneath doesn’t disappear. It is absorbed.

Assessing what matters

Entering this new territory, the questions software buyers have always asked — about feature lists, integrations, implementation timelines, and pricing — still matter, but new factors need to be considered as well. These include:

• Does the solution reduce cognitive load? Features and operational relief are different animals. Buyers need to focus on the output. Does this truly make a clinician’s or administrator’s life easier?

• Is the AI woven into the workflow itself? Practices don’t need another disconnected system. If a user has to click outside of their existing workflow, adoption — and therefore impact — will be limited.

• Can the system handle ambiguity? If software only works in perfect scenarios, it will fail user expectations every time. The real world inevitably deviates from the ideal daily. Systems must adapt accordingly.

• How much hidden operational burden remains after implementation? Onboarding, retraining and workflow fragmentation can sometimes cost more than the software subscription itself. A new platform should minimize hidden costs rather than add to them.

The best is yet to come

We’ve been talking about the promise of healthcare technology for more than two decades. Now we sit at the cusp of reaching its potential. There is a tremendous opportunity, thanks to AI, for us to finally build, deploy and maximize software that accepts, even welcomes, all of the complexity and mess while dramatically simplifying the user experience.

When that happens, technology disappears into the background and allows clinicians to focus on patients instead of platforms. This will become the defining characteristic of the next great healthcare software offerings, and we will be better for it.

Photo: pixelliebe, Getty Images

This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.