James Chen founded DrFirst and is the CEO and chairman. He is a seasoned entrepreneur with experience starting and leading tech businesses.
From the Jetsons to the Terminator, we’ve enjoyed a cinematic view of the future where artificial intelligence (AI) achieves human-like consciousness, taking on everything from cleaning our homes to doing all kinds of jobs we’d rather not. With over-the-top expectations like these, is it any wonder healthcare leaders are becoming increasingly disillusioned by artificial intelligence?
Reclaiming The Narrative
As architects and advocates of AI technology, it’s time we reclaim the narrative and rebuild lost trust. Business and technology leaders can’t rewrite TV and movie scripts, but we can—and do—tell the story of AI’s potential every day, whether we realize it or not. And people listen in ways we may not expect.
Investors read between the lines of quarterly financials for evidence that AI’s impact has loftier goals beyond saving time and money. Customers tune out from tired cliches about AI and machine learning and want proof of results that matter and ROI.
If we choose our words carefully, we can strongly and positively influence conversations about AI and decisions about its use.
In healthcare technology, some of us have decided to talk less about the fantastic things machine learning is supposed to do in the future and talk more about the practical jobs AI performs every day. We’re working to be more intentional with our language, eliminating buzzwords and tech jargon in favor of plain talk that defines our digital solutions by the improvements in patient outcomes they achieve.
Previous overpromises have many executives wondering if AI will ever prove that it can transform healthcare when, in fact, it already has. Millions of small tasks that once weighed down healthcare providers are now done safely, reliably and accurately—words we don’t use lightly but can say with confidence because we have clinical outcomes to back them up.
To have meaningful conversations about artificial intelligence that move beyond buzzwords and cliches, such as “AI-powered,” we need to focus on these types of practical use cases. I’m talking about examples where AI and machine learning are proving they can transfer data from disparate systems, even filling in the blanks, so that the data can be used to trigger safety alerts—not just stored in a digital file cabinet.
Real Talk About AI And The Free-Text Mess
Prescription instructions are an example of how what seems like a small thing can have a tremendous impact across the healthcare system. Medications are ubiquitous, with 69% of physician office visits and 80% of emergency department visits involving drug therapy. That means that every month, millions of electronic prescription instructions are shared among disparate systems using different terminologies. Most people assume that these data transfers happen entirely electronically, without human input. But that’s not the case.
For every e-prescription written, there are multiple drug databases that play a part in how the prescription instructions are written and electronically transmitted. This means that a single, common prescription instruction—take one tablet by mouth once daily—has 832 commonly used permutations, including “1QD” and “tk 1 t po qd.”
Disparate systems and different vocabulary mean that data typically gets dumped into lumps of “free-text” instead of parsed into actionable fields. As a result, clinicians are reading and manually entering data, over and over, which is tedious, mind-numbing and can introduce keyboard errors. This happens more frequently than you may think. A recent study revealed that pharmacy staff has to manually edit 83.8% of e-prescription instructions.
Here, too, it’s important that we reframe the perception that digitization and artificial intelligence created inefficiency when they were meant to solve for it. For AI to deliver on its promise of patient safety and provider efficiency, healthcare systems must have a clearly defined implementation strategy that makes shared data truly interoperable in a practical sense.
Reducing Tedium To Alleviate Clinician Burnout And Staffing Shortages
Burnout is rampant today among healthcare providers facing the strain of after-hours work, excessive administrative tasks and chaotic workplaces. The pandemic only exacerbated this pattern, now affecting about 30% of all clinicians: physicians, pharmacists, nurses and others.
This is where AI can (and does) make a difference.
Data collection and processing, including prescriptions, are among the repetitive work activities highly feasible for AI-driven automation, according to a seminal study from McKinsey. In pioneering artificial intelligence to improve healthcare workflows, my own company is finding that certain data entry, data sharing and data synchronization tasks are ripe for optimization.
When implemented, that optimization means doctors, pharmacists and other healthcare professionals don’t waste precious time typing and transcribing data—time they should be spending with patients. And we are talking about a lot of time; healthcare professionals spend hundreds of thousands of hours on data entry instead of interacting with patients. With workforce shortages and clinician burnout continuing to jeopardize healthcare systems, AI can make an extraordinary difference in everyday work.
The Way Forward
We can achieve macro results with AI in healthcare by focusing on seemingly “small” issues like prescription data entry that, when taken together with other forms of data management, have proven to save lives, time and money. Let’s lead the AI in healthcare conversation to these areas, where it can finally live up to its promises. It might not be the next big blockbuster at the theater, but it is a realistic narrative we can all get behind.
Published by Forbes – September 9, 2022