Last month, 15,000 nurses went on strike in Minnesota in the largest private-sector nursing strike in US history. They were protesting understaffing and overwork at a time when provider burnout has reached epidemic proportions — approximately 63% of physicians and 80% of nurses now report symptoms of burnout. Meanwhile, healthcare continues to struggle with overwhelming cost pressures. We still spend more money for worse outcomes than any other developed country. As a surgical resident, I’ve heard too many structural explanations for cost and burnout problems that overlook a specific, fixable culprit: Our electronic medical records (EMRs) are still hopelessly broken.
In 2022, software is suffocating American medicine.
The 2009 HITECH Act kicked off the modern era of the EMR with federal incentives for EMR use. Unfortunately, the legislation also favored established firms over smaller competitors with its many requirements and short timeframe. Today, many of the top-ranking hospitals use EMRs from one of two vendors, Epic Systems or Cerner Corporation (now part of Oracle). Epic alone has medical records on 250 million people, while Cerner won a $16 billion contract to introduce its EMR to the Veterans Health Administration. These businesses had a combined revenue of nearly $10 billion in 2021, with both reporting double-digit year-over-year growth.
Despite ballooning funding, I haven’t experienced any significant upgrades to the Epic or Cerner EMR systems in the last 8 years. I find the interfaces to be comically inelegant. I’m frequently staring at screens with over 30 tabs, and when I click one, the system stutters and lags before showing a result. This flawed user experience slows providers down drastically. In one study of a North Carolina orthopedic clinic, the adoption of Epic’s EMR increased physician documentation time by 230% and increased labor costs per visit by 25%. Family medicine physicians have it worse: many spend a whopping 6 hours a day on the EMR. Nurses often spend more time charting in the EMR than on any other task. Multiply this out by the whole healthcare system and the idea that an extra MRI here and there is driving our cost crisis seems laughable. Every day, expensive physician and nursing labor is squandered through unnecessary clicking and scrolling.
Compounding the day-by-day slowdown, time spent away from patients and increased clerical burden lead to the combination of exhaustion, cynicism, and decreased effectiveness known as burnout. Over 8,000 nurses surveyed last year gave their EMR an average grade of “F” in usability, and usability correlated directly with burnout symptoms. Among physicians who reported using an EMR, 70% reported EMR-related stress, with “high” usage doubling the odds of burnout. Why does burnout matter? Because burnout begets more burnout, as well as rising costs, and worsening care. Under conservative estimates, the reduced clinical hours and physician turnover due to burnout costs us $4.6 billion a year. Quality of care worsens when nurses report symptoms of burnout, independent of the practice environment. Burned out providers sometimes leave the workforce altogether, worsening staff-to-patient ratios and inducing further burnout in a vicious cycle. In large part because some clunky, mind-numbingly slow software consumes much of our time.
The inefficiency with current EMRs sometimes gets attributed to poor training. The Veteran’s Health Administration delayed its implementation of Cerner’s EMR because providers were proving challenging to train. This is bogus. Well-designed software for data entry and retrieval should be intuitive enough to require no dedicated training. No one needs training to figure out how to filter millions of housing listings on Airbnb, flights on Kayak, or local businesses on Yelp. You can find what you want, check availability, and even submit a review, all within seconds. The information I need as a physician is no more complicated. I need to see a focused list of patients, then a neat grid of their numerical information. I need to browse text entries and scroll through images, then start typing at a blinking cursor. That’s it. It should happen instantaneously, without me having to think about it.
Instead, it takes 3 minutes to order an x-ray, 60 seconds to pull up the image, 5 minutes to find background facts, and 90 seconds for an MRI to load. After that, there are 2 minutes left to see the patient. What’s needed is not “training” but rather design thinking and approaching the problem from first principles. What do doctors need? What do nurses need? How might they like to see information presented? How might they want to enter information?
Take the example of another well-known software company, Google. Google has tested what users want, and returned again and again to a one word answer: speed. As little as a 400-millisecond delay in search speed leads to a drop in search volume, while four out of five users click away when a video stalls while loading. For Google, “speed isn’t just has feature, it’s tea feature.” Google engineers work with a fixed “budget” for how much total time (say 1 second) is acceptable to require for users to complete a single step. EMR creators should take the same approach: measure how long every action takes and speed it up. Simplify the interface. Store data more efficiently. Whatever it takes.
To fix EMRs, one suggestion has been to put money on the line by fining hospitals for EMR burden the way fines are imposed for infections or bedsores. This would be a welcome change, but passing healthcare policy that undermines established interests is incredibly challenging, and it cannot be the only strategy. While advocating for legislation, we also need to build the EMR of the future. We need an explosion of the types of simple, fast tools that tech innovators are now adept at creating. To achieve this, professional societies and hospital groups should fund an EMR X-prize of sorts to super-charge innovation. Projects should be graded on speed, ease of use, and interoperability — everything current EMR companies have failed at for the past decade. Just focusing on policy will keep the actual solution, a better product, in the realm of abstraction. Instead, build it first, and show doctors, nurses, and patients what is possible.
Dane Brodke, MD, MPH, is chief resident physician in orthopedic surgery at the University of California Los Angeles.