Most major Electronic Health Record (EHR) systems today are nothing more than basic, fragmented measurement reporting tools with user interfaces that make it difficult to manipulate enormous sets of medical data. To their credit, they’ve become a fundamental component of today’s healthcare system, a physician’s digital consultant, yet unlike human colleagues who communicate meaningful information, these systems burden their users to find the meaning themselves. Simply put, an EHR shouldn’t just report raw medical data; it should explain it. And it can do that through good design.
Last week I was given a consult to evaluate a patient suspected to have acute kidney injury. The patient’s serum creatinine level, an indicator of kidney function, had elevated significantly over the past two days – not a good sign. Searching throughout the EHR for the usual culprits – imaging contrast, nephrotoxic medications, sepsis – left me clueless. Later I learned from the nurse that a few days ago the patient had indeed received several doses of a nephrotoxic drug, yet those events were virtually invisible as the EHR’s default medication screen contained only currently active orders. After adjusting the date range to display two weeks’ worth of data (to be safe), I now had to sift through about a hundred struck-through discontinued medication order entries to find what I’d been looking for. But why wasn’t the EHR already doing the legwork, displaying these potentially pertinent correlations and data points, painting a more descriptive clinical picture, instead of leaving me to wade through mounds of information? The answer to that question came down to poor interface design, which, not surprisingly, is a big reason physicians are reporting decreased satisfaction and productivity with their EHRs.
Two possible solutions to the above problem are based on an understanding that’s often overlooked by today’s EHRs: that medical data is highly contextual and time-dependent. A contextual solution would have been to display a list of nephrotoxic interventions the patient had recently received alongside creatinine levels. Vancomycin, the drug the patient was given, is widely accepted as nephrotoxic, and anyone, doctor and patient, can readily find that information with a simple web search. And most of today’s major EHRs have access to entire databases of drug interactions and side-effects, they’re just not connecting the right dots. Solutions of this type do not require sophisticated A.I. algorithms, but rather that designers clearly and explicitly display basic associations and correlations that are staples of clinical medicine. That aspirin can cause abnormal liver function tests (LFTs) is an almost indisputable fact, and so it seems painfully obvious that recent occurrences of aspirin ingestion be grouped contextually with LFT results. And beta-blockers with blood pressure values, antibiotics with white blood cell count, etc.
With the Rise of AI, What IP Disputes in Healthcare Are Likely to Emerge?
Munck Wilson Mandala Partner Greg Howison shared his perspective on some of the legal ramifications around AI, IP, connected devices and the data they generate, in response to emailed questions.
A more time-dependent solution would include providing the capability to easily juxtapose various lab values, vital signs, and order history alongside one another, allowing for more effective relationship discovery. Trending creatinine levels in parallel with medication history would have made it incredibly easy to spot the correlation I labored to find earlier. Unfortunately, today’s EHR systems categorize and segregate tightly-associated data sets across multiple screens, resulting in users having to memorize the timestamps and details of various values and interventions as they flip tab-to-tab in order to construct their own personal timeline of events – an immense amount of unnecessary cognitive strain. By presenting categories of information as parallel timelines, you more accurately depict the natural chronological interdependencies of medical data. Progress notes that contain written changes in management can be more easily confirmed by comparing timeline data, adjustments in medications and their dosages can be more finely tracked and correlated to a new onset of symptoms or to lab abnormalities. Designing health information in this manner would afford users the freedom to more effectively focus on and complete important tasks, which is good design at its core. Others seem to share similar sentiments and are exploring this new direction by utilizing existing social media timelines.
The ultimate goal of these solutions is to design the EHR to serve as a physician’s true digital partner, communicating data clearly and meaningfully. With some clinicians now spending more than 50% of their time interacting with medical software, it’s imperative that this time spent isn’t wasted on wrestling with interfaces and straining our cognitive capacities. Rather, it should be spent on attaining a deeper understanding of patients, arriving at more accurate diagnoses, and devising more effective treatment plans. And although design doesn’t seem to be a priority among most major healthcare software vendors, recent efforts by institutes such as NIST and HIMSS to establish healthcare usability frameworks only support the claim that good design in medical software is crucial. This is the type of design that can literally save lives.
This post originally appeared on Kyro Beshay’s blog.
A Personalized Approach to Medication Nonadherence
At the Abarca Forward conference earlier this year, George Van Antwerp, managing director at Deloitte, discussed how social determinants of health and a personalized member experience can improve medication adherence and health outcomes.
