Stacy O’Connor, MD, is the new Associate Vice Chair of Informatics and Quality as well as Associate Professor of Radiology in Abdominal Imaging with UNC Radiology.
Dr. O’Connor received her bachelor’s degree from MIT in 2000 before earning her Medical Degree with distinction in research from Mount Sinai School of Medicine in New York in 2006. She completed her intern year at Leigh Valley Hospital in Pennsylvania before moving on to a Radiology Residency at The University of Wisconsin. She continued her education with an Abdominal Imaging and Intervention and Informatics fellowship, followed by a Biomedical Informatics Research and Training fellowship at Brigham and Women’s Hospital, which she completed in 2015. During that time, she also received her Master’s in Public Health from the T. H . Chan Harvard School of Public Health and a Master of Medical Science with a dual concentration in clinical informatics and imaging informatics from Harvard Medical School.
During her second fellowship, Dr. O’Connor started her career as a Radiologist in the Community Division of Brigham and Women’s Hospital. She then accepted a position as an Assistant Professor of Radiology in the Abdominal/Body Imaging Section at Medical College of Wisconsin/Froedtert Hospital, where she spent the past seven years of her career before relocating to Chapel Hill. During her tenure at Wisconsin, Dr. O’Connor took on additional roles in the radiology department, including Medical Director of IT Operations, Patient Safety and Quality Officer, and Medical Director of the Quantitative Imaging Lab.
Dr. O’Connor has multiple publications and presentations across a spectrum of clinical, informatics, and quality improvement topics. Her recent research includes characterization, disparities in follow-up, and outcomes of incidental findings, using deep learning to estimate organ-specific radiation dose decreases when optimizing CT protocols, and the use of artificial intelligence in imaging.
What attracted you to the Department of Radiology at UNC?
There were a couple of different things that piqued my interest in UNC. I first heard about the position through a friend, Dr. Dorothy Sippo, the new Vice Chair of Informatics. We did Fellowship together in informatics, and she asked me to consider coming down as her Associate Vice Chair. I investigated the position and had a chance to talk with Maureen Kohi, the Chair, and Dr. Mossa-Basha, the Vice Chair of Quality, whom I am acquainted with from a national group. I was able to meet multiple members of the Abdominal section before and during my interviews, including a video call with Dr. McGinty, who was in Malawi! Between the people, I already knew and the people I met while looking into the position, it seemed like a great group of people to get involved with.
What inspired you to pursue medicine and become a doctor?
I wanted to be a doctor for as long as I can remember. As a little kid, I remember going to the doctors and thinking I could see myself doing what they did. I love puzzles, problem-solving, and figuring out how things work. The human body is the most complex machine in the world. It’s exciting to see how it all works together. And when things go awry, to determine why and what you can do about it. I went to MIT, and while I graduated with a science degree, my thinking process is very much like an engineer—wanting to know the how and why behind everything. Between my love of problem-solving and the ability to continuously learn in an ever-changing field, medicine was the right path for me to pursue.
What made you decide to pursue Radiology?
I fell in love with Radiology because of its diagnostic aspect and the ability to incorporate my interest in the hard sciences of physics, chemistry, biology, and computer science. I did an Abdominal Imaging and Intervention fellowship, where I also discovered a passion for procedures done by diagnostic radiologists. Not only was I assessing a patient’s problem on imaging but also using those images to figure out in 3D space how to be able to place a drain or do a biopsy. I find it fascinating.
How did you decide to pursue your current specialty?
All of what I said above, and in terms of general Diagnostic Radiology, I love the variety and the chance to work with people across Departments and specialties. The collaboration is something I really enjoy about my specialty and radiology in general.
One of the things I love is when the referring providers come down to the reading room to talk about the patient. For example, someone came down to talk about a challenging GI case. I started asking questions about medications because I noticed something on the scan, and they were surprised I could see that on the screen. It then gave me more information because if the patient felt the need to take that medication, then with what I’m seeing, I can discern that it’s probably X versus Y. By working with the referring provider, I have a better idea of what’s going on with the patient, thus allowing me to give them a better read.
Why abdominal, in particular?
I have found that abdominal/body imaging is an area where I can do the most good. There are multiple organs we deal with in the abdominal division, with processes that can be confined to the abdomen/pelvis or are part of a bigger picture. If a patient complains of belly pain, I look at the scans to evaluate what might be going on. Is it their gut, their liver, or their kidneys? There are many different things that it could be, and trying to understand how that all interplays speaks to my passion for puzzle solving.
What about informatics drew you, and how do you find that it integrates with Radiology?
In college, I was trying to decide between computers and medicine. I decided to go with medicine because I thought it was a different type of problem-solving, and ultimately, I would have the chance to help people. Although I understand the attraction, I wasn’t as interested in building a nice piece of software.
When I was in residency, I went to a leadership program, and Dr. Nabeel Safdar came by to talk about informatics. He showed how you could use your computer skills to make medicine even better, improving care for patients and the whole healthcare team. For me, it was the perfect way to marry two of my passions and be able to use both to help others.
How do informatics and quality come together in your role?
Computers and informatics tools can keep track of many things, never need sleep, and can be deployed to help many people at once. For example, primary care providers (PCPs) see many patients. They put in orders for so many labs that it’s hard to remember what hasn’t been done. When people get imaging done, they often follow up with their provider at the next visit because that’s how medicine works nowadays. Everything is episodic. These are old numbers, but a primary care provider can get at least 60 Path or Radiology reports a day, and they are being hit left, right, and center with all this information. One of the things informatics can do is to offload from our providers. Some tools can be used to flag essential reports, “Read this one Now!” There are only so many hours in the day, and we’re trying to reduce what many people call “pajama time,” where doctors work on their charts after they get home.
Informatics tools can also improve quality and patient safety. One of the things I have been working on is developing tools and workflows to make sure patients get follow-up recommendations in their radiology reports or make an informed decision not to. So, a classic case of when this doesn’t happen is a patient with a lung nodule that is incidentally noted. For example, they get a shoulder X-ray where part of the lung is seen, and the radiologist notes a lung nodule. The patient and provider are concerned with the shoulder, so the lung nodule fell through the cracks. Three years later, the patient returns with a cough, and they have metastatic lung cancer that could have been avoided.
So, the informatics tools I’m working on to improve patient safety and quality track that—making sure the follow-up chest CT or x-ray is done. For example, an orthopedist ordered a scan of the shoulder. They don’t handle the lungs, so they may ask or expect the primary care provider to get the follow-up imaging. But, this communication can fail. Even if it doesn’t, things can still go wrong. Say the PCP knows about the nodule, they ordered the CT chest and did everything they were supposed to do, but the patient never showed up for the scan. They got busy with their life, and they forgot. One of the challenges for the PCP is to realize that the patient never got the scan they ordered. So again, it’s a chance for computers to catch that because computers don’t forget. By catching things like this, we hope to improve quality by making sure these patients get what they need.
There are also other quality things that we do. There is software that can keep track of all the latest recommendations from various medical societies and published literature. For example, if a patient has an 8 mm left upper lobe pulmonary nodule, the report can say,“ according to the latest recommendations, you should recommend a follow-up CT in x months.” Computers can keep track of all the latest recommendations and can provide appropriate language and citations for your report. This is important because if you look at the studies, it takes about 15 years between when something is proven in the literature or becomes part of a guideline to when it’s common practice. So, some of the quality interventions are to allow people to perform at their best with the world’s knowledge at their fingertips. There is no way physicians can be expected to read all the literature because it’s just not humanly possible. This is an instance where it’s something appropriate for a computer to do. The computer can grab the recommendations and present them to the provider. Maybe the provider chooses not to take the recommendation because there’s something else going on with the patient, but the software gives them the opportunity to make a better-informed decision using the most up-to-date guidance.
What medical advances would you like to see in the next five years in your specialty?
There’s exciting work being done now that can have a big impact on our profession as a whole in the next five to ten years with machine learning and AI. For example, these tools can help to make better images from the existing data in our scans. They will be able to take the same level of radiation and create a beautiful CT scan; not only can it show me the anatomy but also, through dual-energy or spectral imaging, give me even more information about the tissues themselves. Same with MRI. I can’t wait until using AI, and machine learning can produce a scan that takes the same amount of time but gives you beautiful pictures even if the patient can’t hold their breath, which is one of the challenges for abdominal imaging, especially MRI.
Another advance that I would love to see is to allow for closer collaboration amongst all people caring for a patient, including referring providers and radiologists. In the future, a computer could distill down the information found throughout the medical record and provide each person with a complete picture of the relevant subset of information about the patient, including medical history, labs, other comorbidities, diagnoses, and genetics. All the details can provide more comprehensive information about the patient that can help with a more accurate diagnosis without the need to scan through reams of irrelevant records. This whole picture will depend heavily on the primary care provider, the relationship that they’ve built with the patient, and the ability to quickly and easily enter all observations and discussions in the record. The only information radiologists currently have for the patient is limited to what is shared or can be found through diving into the medical record as much as time and patience allow. The paradigm could be shifted if more was provided in the medical record and a tool could extract, summarize, and present the relevant information at the time of image interpretation or other interaction.
Why did you decide to go into academic medicine?
Academics provide another level of a challenge since our institution gets complex cases, pushing us to think deeply, pull from all our resources, and work together. Not only do I enjoy the challenge but also the shared dedication to working out a solution.
The other part is that I love teaching. I enjoy going through cases with our fellows, residents, and med students. And you get to see the trainees grow. Radiology residency is four years of radiology-specific training after one year of internship, allowing us to see the resident’s progression. When we first see them, it’s much more basic explanations, but by the time they graduate, their understanding of scans and making connections for a diagnosis is so much more mature, and this is really rewarding.
What is one thing you wish that your co-workers knew about you before they met you?
I am always doing something, but don’t mind being interrupted at all. I love it when people come over and want to talk, especially if I can help them with something. Even if it’s some random project that they’re just thinking about doing, I’m never too busy to talk to colleagues, trainees, or referring providers.
What’s one piece of advice you would give somebody who’s considering going into Radiology?
Don’t believe all the hype about AI. I remember when Dr. Ezekiel Emanuel spoke at the American College of Radiology conference in 2016 and told us not to train any more residents because AI was going to take all our jobs. Surprisingly, you still hear that from many people. If you’ve ever trained in AI, you’d know that is not going to happen. I do think AI and machine learning are going to become a big part of radiology, but I think it’s going to be the radiologist using it as a tool, not as a replacement.
How would you describe yourself in one word?
Multitasker.
If you could have one superpower, what would it be and why?
I think being able to read people’s minds, to get the gist of what they’re thinking and how they think rather than the specific words. I consider myself at least a little neurodivergent, and I’ve worked hard to connect with individuals and understand their perspectives. Sometimes I just don’t understand people, specifically how they reach certain conclusions or react to certain situations. To be able to understand how people think would be helpful.