What does a photographer and a clinical informatics researcher have in common?
There may be more than you’d think at first glance.
When James Mitchell was pursuing his undergraduate degree in the United Kingdom, he studied photography—working across both digital and film mediums.
Today, James Mitchell, PhD, is an assistant professor of biomedical informatics at the University of Colorado Anschutz (CU Anschutz). His research focuses on human-computer interaction, clinical decision support, and artificial intelligence (AI)—designing systems that improve how information is delivered at the point of care.
“I’ve got a weird, eclectic story,” said Mitchell.
While studying photography during his undergraduate career, Mitchell got a job in the retail stores at Apple. “Back then, Apple Retail was a small company,” recalled Mitchell. “There was only like three stores in the UK. There was only like six stores in the US, I think at the time.”
Mitchell started off as what’s known as an Apple ‘Specialist.’ At the time, the way that Steve Jobs and Ron Johnson had created the retails stores was that each part of the store focused on a field or creative domain. For example, you would have a section for photography, movie making, music making, office applications and more. The stores would then hire people who were experienced in their respective domains to support with those sections of the store.
After that, Mitchell completed his certifications and became an Apple ‘Genius.’
“[As a genius], you knew how these things were made, how everything inside worked, why things didn’t work, how to fix it. We used to have to use thermal paste on the processes and do heat checks and all those types of things,” Mitchell recalled.
Mitchell continued his career with Apple, working as a Lead Genius. He later transitioned into user acceptance testing, where he focused on the retail systems.
In 2015, Mitchell’s son was born, and his wife went back to university. This got him thinking about the future. His wife and him sat down to make a five year game plan, and part of that was defining the things he did at Apple and turning them into a degree. After these discussions, Mitchell completed a Master’s degree in Advanced Computer Science.
After his Master’s, Mitchell went on to complete his PhD and a postdoctoral fellowship where he focused on clinical information delivery at the bedside.
In 2021, the Department of Biomedical Informatics at CU Anschutz had posted a job, and Michell came across it. “Basically, they were looking for somebody that did human-computer interaction, natural language processing, and clinical informatics—and I had experience in all of that. So, I applied... then started here at the end of August 2022.”
Mitchell and his lab, the Mitchell Lab at CU Anschutz, focus their research on information delivery in technology, as well as other challenges that occur at the bedside.
"In a nutshell,” said Mitchell, “my research focuses on human-computer interaction, clinical decision making, human-AI interaction and how we solve real world problems through innovative apps and solutions that provide people with support in the decisions that they make. That could be from an academic perspective, a clinical perspective or just for general people. "
Much of their work is around clinical problems, working with mobile applications, Electronic Health Records (EHRs) and other forms of information operating in clinical settings. Some of the questions driving the Mitchell Lab’s work are:
One of the key pillars of their work is looking at how information is delivered, accessed and received—focusing on how humans interact with computers and digital devices.
“The problem,” said Mitchell, “is trying to figure out how you deliver what type of information, when—and especially in what domain.”
Yet clinical settings pose unique situational and environmental challenges when looking at information delivery human-computer interaction and AI-computer interaction. “Because, you know,” emphasized Mitchell, “for you on your iPhone, for something like Google Maps or Apple Maps, that's different to a high-stakes environment like a clinical environment where a clinician is trying to make a decision.”
Mitchell has learned strategies and best practices that can improve the delivery of information in clinical environments. For example, one strategy is to show simple information at the top and give the user the option to uncover more complex information. “You see this in websites all the time,” said Mitchell. These techniques improve the accessibility and usability of information for users.
“You have to look at things like information architecture—so how you lay things out,” he explained. “There's a hierarchy to it as well. So you have simple information, and there’s a gradient to complex information.”
“So a lot of my work is looking at how we deliver that information on mobile devices or in the EHR in a simple way that allows [clinicians] to make a decision, and learn to trust the system, and all those kinds of things.”
A large part of this work is also ensuring that these systems are actually effective. The Mitchell Lab thus has to gather user feedback to make sure patients and clinicians can interact with and benefit from the architecture, design, and workflow of the system.
Yet, clinicians are often extraordinarily busy, and healthcare is a challenging place to collect this data. It may not always be feasible to, say, ask healthcare workers to attend an hour-long focus group about a system or app.
To address this challenge, the Mitchell Lab also devotes a lot of energy and effort to create iterative feedback mechanisms that are built into their innovations, allowing them to receive vital feedback. This allows them to make revisions to improve the efficacy of the system while adapting to and respecting the time constraints of the environment.
For example, the Mitchell Lab has developed ClarifAI, an AI-powered tool that facilitates higher quality feedback received from surveys and Asynchronous Online Focus Groups (AOFGs). It accomplishes this by determining the quality and detail of an answer and, if applicable, asking follow up questions to better understand the feedback.
These efforts further their mission of improving human-computer interaction and human-AI interaction at the point of care.
Working at the intersection of cutting-edge technology and high-stakes clinical care—the Mitchell Lab’s research is innovative.
Yet the Mitchell Lab is also innovative in their approach to research.
Research can begin in many ways. It often begins with an identified gap, question or problem that warrants further exploration. Researchers may develop a hypothesis or an inquiry framework, then gather and analyze data to deepen understanding, test assumptions or identify possible solutions.
Mitchell’s approach to research is a bit different.
"I have a sort of flipped approach,” he explained. “I see a problem. I try and solve it with maybe an app, or software, or a pipeline or something like that—and then I look at ways at how we can extract science out of it.”
Although he no longer works at Apple, Mitchell’s engineering mind is as unwavering as always. "I'm a builder—I enjoy building things,” said Mitchell. “I enjoy solving problems. You know... I don't enjoy writing them out as much as I enjoy actually fixing the problem. Then if I'm lucky enough that it has some research output, then great; if not, maybe there's some commercial opportunity there and I can file patents, like we've done for certain things... so I do a multifaceted approach.”
So, what does a photographer and a clinical informatics researcher have in common?
A photographer captures data in the form of light and translates it into a photograph—seeking to create a composition that will depict a scene, capture a moment and resonate with their audience.
A clinical informatics researcher takes data in the form of conversations, health records and biomedical data and finds ways to represent it—aiming to support clinicians, help patients and improve information at the point of care.
Both are designers, innovators, and engineers—working to capture, represent and share information in a useful, meaningful way.
And both are part of James Mitchell’s story.
“I'm just lucky enough that I've got a platform, which is my lab, to be able to do the work that I want to do, and solve the problems that I want to solve.”
— James Mitchell, PhD, MS