Innovation

3 Health-Tech Tips From the Team Behind VA’s COVID-19 Clinical Decision Tool

Presidential Innovation Fellow Kaeli Yuen shares tips for developing clinician- and patient-centered healthcare tools.
Kaeli stands outside in Washington, DC. She wears a floral shirt and a white mask to protect herself from COVID-19. It is a sunny, blue-sky day, and behind her you can see city buildings and the DC skyline, including the Washington Monument.

Kaeli Yuen.

Innovating in healthcare can be intimidating. Stakes are high (people’s health and lives!). The U.S. healthcare system, health data, and clinicians’ workflows are complex. However, the field is ripe for change. As a Presidential Innovation Fellow (PIF) working with the U.S. Department of Veterans Affairs (VA) Office of the Chief Technology Officer (CTO), I saw the pressure that COVID-19 placed on VA clinicians first-hand. Working with teammates across the VA Office of the CTO, I developed clinical decision support tools to help VA clinicians care for COVID-19 patients.

Below, I share three tips that digital health professionals can apply when developing provider-facing healthcare tools:

  1. Work closely with end users at every stage.
  2. Work closely with the broader medical community.
  3. Try small improvements; they can have profound impact.

First, some context: Clinical decision support uses automated methods to assist healthcare professionals with evidence-based decision making. It’s ready for innovation because medical knowledge (what we know about great patient care based on the latest evidence) is rapidly expanding. This puts significant pressure on clinicians. They need to keep up with evidence-based best practices to provide the highest standard of care.

The COVID-19 pandemic exacerbates this issue because:

  • COVID-19 is a novel disease,
  • Clinical guidelines are rapidly evolving, and
  • These changes demand constant clinician attention and action.

When I spoke with front-line VA clinicians, they shared a recurring challenge: determining a patient’s disposition, or the location where they should receive treatment.

To address this challenge, we developed the COVID-19 Patient Manager, a clinical decision support tool for emergency medicine physicians. This tool presents patient-specific recommendations from the American College of Emergency Physicians (ACEP) in a digital format. We began piloting the COVID-19 Patient Manager this summer, and early user feedback has been very positive.

We identified three best practices throughout product development:

1. Work closely with end users at every stage

People often develop health information technologies without sufficient input from clinician end-users. This presents problems like:

  • Poor clinician workflow integration,
  • Clinically irrelevant features that decrease clinician trust, and
  • Poor usability, contributing to clinician burnout.

To avoid these potential pitfalls, we recruited a group of highly-engaged VA emergency medicine physicians to serve as project clinical leads. We met weekly with this group to deeply understand the problems they experienced while caring for COVID-19 patients, understand their clinical workflows, get input on our tool’s clinical content and presentation, and more. One of many COVID-19 Patient Manager features inspired by this engagement is a “copy to clipboard” function, which enables clinicians to easily transfer tool recommendations to their electronic health record notes.

These clinical leads also served as fantastic champions for the project within the VA. Their project involvement helped us obtain essential buy-in from agency leadership. By working closely with our end users at every stage of product development, we guaranteed useful features, valuable feedback, and long-term advocates.

2. Work closely with the broader medical community, too

Clinician trust indicates success for many clinical decision support tools, including ours. Clinicians often look to professional medical organizations for trustworthy, up-to-date, and evidence-based guidelines. Our development team built trust by:

  • Forming strong collaborations with professional medical organizations, and
  • Clearly communicating how our tool relates to these organizations’ evidence-based guidelines.

Our development team partnered with the American College of Emergency Physicians (ACEP), an organization that our clinical leads vetted and trusted, to implement their COVID-19 management guidelines. Our development team relied often on ACEP to confirm that we were implementing their recommendations as they had intended.

As our clinical leads practiced these COVID-19 guidelines, they shared feedback with ACEP. This feedback improved the next iterations. For example, feedback from our clinicians about how often arterial blood gases are sampled and recorded in their practices led ACEP to reconsider how that test factored into their recommendations.

A tight collaboration between the COVID-19 Patient Manager team and ACEP benefitted both groups. This ultimately benefits veterans and patients as well, since physicians gain better access to tools that support their clinical decision-making.

3. Start small and iterate

A screenshot of a PDF. In thick black text, a title reads, Emergency Department COVID-19 Severity Classification. Small black text guides clinicians through appropriate evaluation and disposition for adult patients with suspected or confirmed COVID-19. There are 7 columns, each a different color, with many text boxes and directions. The columns ascend in severity, from mild-low risk to critical.

Original PDF of ACEP COVID-19 patient management guidelines.

Even a small improvement can have a big impact. COVID-19 Patient Manager faced bureaucratic processes that impacted our launch timeline. This included many procedures for accessing real-time patient data from electronic health records, which would be used to compute ACEP’s recommendations.

Given COVID-19’s urgency, we wanted to get something in clinicians’ hands as soon as possible. We decided to “soft launch” a version of our tool that required clinicians to manually enter patient data into a digital tool, rather than have the tool auto-populate with patient data from the electronic health record. This was an incremental improvement over the original version of ACEP’s guidelines, which comprised an eight-page PDF that clinicians had to print and manually complete and calculate.

A screenshot of the COVID-19 Patient Manager application. Compared to the PDF screenshot, this image is much simpler. There are clear white text box areas to input patient info. An orange in-app message guides clinicians and reads, All fields are optional, but completing the empty fields will improve the quality of recommendations. Below this message guide, a summary of recommendations appears, along with guidance on whether to admit the patient. All of this information is included neatly in one screen, with a clean and clear grey, white, and black color scheme, with key info highlighted in orange, and calls to action in blue.

A screenshot of the VA COVID-19 Patient Manager, with imaginary patient data.

Despite this seemingly incremental improvement, we received highly positive feedback from our clinical leads and other emergency medicine physicians. We heard that virtually no physicians had used ACEP’s PDF version of their guidelines in practice due to printing and manual computing burdens.

With our soft-launch tool, physicians integrated ACEP’s guidelines into practice, and said that it improved their decisions regarding COVID-19 care. For example, a physician reported changing their decision to discharge a patient home after inputting the patient’s data to COVID-19 Patient Manager and reviewing its recommendation to admit the patient.

Our COVID-19 Patient Manager approach is inspiring many VA colleagues about the potential of clinical decision support to strengthen veteran healthcare, during and beyond the pandemic. I hope it’s useful to you, too!


Are you passionate about technology as a tool to improve lives? Interested in using your expertise for public good? That’s what Presidential Innovation Fellows (PIF) is all about. Learn more about PIF at our website, and follow us on Twitter and LinkedIn to find the latest work from our Fellows.