Google Health UX lead on designing EHRs that work for clinicians

A user experience manager at Google discusses technology fatigue, machine learning's role in electronic health record design, workflow optimization and more.
By Bill Siwicki
10:42 AM

Melissa Strader, UX manager at Google Health

Photo: Google Health

Given that existing health IT tools – especially electronic health records – notoriously are not very user-friendly, many healthcare information technology vendors are spending time and resources designing IT that makes clinicians happy.

One such vendor, Google Health, recently introduced Care Studio, software designed to streamline healthcare information for clinicians. It brings together information from different EHRs, and lets doctors and nurses browse and search in one streamlined interface.

Google Health has a wide array of IT design expertise. Healthcare IT News has tapped that expertise, interviewing Melissa Strader, UX manager at the company. She discusses keys for designing EHRs for clinicians, getting past tech fatigue, understanding clinician behavior and workflows, and machine learning's role in EHR design.

Q. What are some keys for designing EHRs for clinicians, given current tools are often not very user-friendly?

A. In the last few decades, health IT solutions have done a great job at digitizing healthcare information and processes. Today's EHRs are powerful tools and are built to address many needs, including supporting hospital administrators, IT teams, billing departments and more.

This can make it challenging for clinicians to quickly find the information they need at the point of care. Doctors and nurses often have to navigate through multiple, disconnected systems to get all the information they need to care for a patient.

My biggest takeaway when designing technology is to start with who's using it – in this case, clinicians – and build to address their needs. Clinicians can be overwhelmed with too much information, so bringing meaning to that information is the key to making clinicians feel like their tools are truly designed for their needs.

What we know about clinicians is that their daily activities and environment are complex and fluid, and therefore any tool needs to be sensitive to those variable workflows and constant task-switching and fit into this way of working.

When applying this approach to Care Studio, Google's solution for bringing health records together, we designed the product around a few key principles that put clinicians first.

First, we wanted to make the time spent between doctors and their patients more valuable, productive and human. Second, we wanted to avoid the temptation to build something that was only incrementally better than was available since we understood that clinicians not only had information overload, they had an overwhelming number of tools as well. And third, we wanted to imagine, with clinicians, how to present information in a meaningful way.

Q. Clinicians are used to hearing big promises about EHRs and are still often disappointed. How do you get past this tech fatigue?

A. It's true that clinicians are fatigued when it comes to technology. They've heard promises of technology streamlining their workflows time and time again. They often have to navigate through multiple different EHR systems to read a patient's history or switch between their computers and fax machines.

Even more, learning a new tech tool can often mean taking days away from caring for patients. Technology is meant to be a tool, but it's easy to see how it becomes a burden.

Given this, you could assume clinicians would automatically see the value of a simple tool. But doctors are familiar with their tools and rely on muscle memory to complete workflows and enter in billing codes, so making a change, even to a more intuitive tool, comes with barriers.

We knew we'd need to approach changes with humility and not expect a new tool to be embraced immediately. Part of developing a product for clinicians was knowing that we needed to listen and learn, and be open to the fact that some of our assumptions may turn out to be wrong.

One way we hope to address tech fatigue and skepticism has been to build a tool that could demonstrate value immediately. Drawing on familiarity, setting expectations on workflow changes and solving clinically important challenges can go a long way in providing value quickly.

Pretty much every clinician is familiar with Google Search, and they know it just works. By designing our product interface around the familiar Google Search bar, we could tell clinicians that this is something they already know how to use and it will just work and help you find what you're looking for.

As a clinician looks for health information, we have many nods to Google Search. For example, once a clinician has found the patient record, the search bar remains at the top so they can continue their journey. As clinicians begin to use the product, they appreciate that it is medically tuned and intuitive to support them in delivering care.

Q. You've said that understanding clinician behavior and workflows should be the focal point in EHR development. How so?

A. For clinicians to adopt a new technology, it must fit into existing behavior and workflows. If a technology isn't designed in this way, it will only contribute to the tech burden that so many doctors and nurses are experiencing. Everyone deserves the respect of feeling like their needs are honored in the design of a product that they use, and clinicians are no different.

Many of our designers and researchers have experience working in healthcare, and we have a team of incredible clinicians at Google. That gave us a strong foundation for understanding, from the beginning, how and why healthcare professionals are constantly pulled in different directions, spending up to half their day on a computer. This means that new technology is fighting for attention from a person who is too overwhelmed to be enthusiastic about using it.

Healthcare is a service, and a lot more goes on in a doctor's day than visits with patients. A technology product is one of the many touch points a doctor has throughout the day.

It needs to perform in its intended way and at the right time along the journey to be successful. Keeping this in mind helps ensure your product can have the biggest impact on a clinician's day, as they care for their patients.

For example, we've learned from our clinical team that typically when a patient is admitted to the hospital, they receive intravenous meds/fluids. In order to be discharged, the patient would need to either be on oral meds, or the IV meds would have to be stopped.

This is pretty laborious for the doctor to check. To make it easier, we created a visual display that shows and color codes the routes. So when the doctor is preparing for discharge, they have multiple visual signals to help spot any issues.

Q. You've also said there is huge potential in using machine learning to make day-to-day EHR work easier for clinicians. For instance, if a doctor is looking for a specific medication or procedure, wouldn't it be helpful if other information related to this was displayed as well? How else can AI help?

A. Working more easily for clinicians is the key here. We all have products and services in our lives that just seem to work without needing to learn how to use them. While EHRs have come a long way in transforming how healthcare data is stored and presented, we've heard from clinicians that this magic of finding exactly what you're looking for is missing.

Although AI and machine learning in healthcare often are associated with predictive analytics, this technology can also be useful to helping clinicians search for information across a medical record. This is difficult to do with health information, since different systems use different codes and units of measurement.

We've used machine learning to connect medically related concepts. Machine learning can draw connections between things like medications and conditions so that clinicians can more easily see all the information needed to understand a patient. Another example is using machine learning to provide "one-line" summaries of patients, with critical information extracted from the notes section, so that providers can quickly get up to speed on new patients.

These examples are rooted in designing for the doctor. While there are endless ways AI may be able to transform healthcare, in design for clinical tools it's important to think first about how it can make a clinician's day to day easier.

Twitter: @SiwickiHealthIT
Email the writer: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.

Want to get more stories like this one? Get daily news updates from Healthcare IT News.
Your subscription has been saved.
Something went wrong. Please try again.