Working at Top of License and Improving Care Delivery: A Q&A With Debra Patt, MD, PhD, MBA, on AI's Practical Promise in Oncology
As Debra Patt, MD, PhD, MBA, explains, the oncology patient population is growing, and professionals across the field must keep pace despite a workforce shortage and rising clinician burnout. AI and other digital tools, however, offer pragmatic solutions to oncology’s most pressing operational challenges, helping to ease the burden on clinical teams.
Dr. Patt sits at the intersection of large-scale practice operations, professional society policy, and national advocacy, giving her a uniquely comprehensive view of cancer care and AI in healthcare, both in terms of where they stand today and where they are headed. As Executive Vice President of Texas Oncology, President of the Community Oncology Alliance, and Chair of ASCO’s AI Task Force, she has led efforts to expand access to digital solutions and integrate AI into clinical and administrative workflows, with an emphasis on scalability, workforce sustainability, and patient-centered care.
In this interview with ASCO AI in Oncology, Dr. Patt discusses AI’s impact on oncology practice, how organizations can build trust in new tools, the role of professional societies like ASCO in accelerating responsible adoption, and what is needed to fully realize AI’s potential in oncology care in the coming years.
What led you to focus on AI in cancer care?
My goal is to make sure that patients with cancer can get what they need. We are lucky to live in the golden age of modern cancer therapy, where the goal is clear for patients not to succumb to their cancer and for their livelihood not to succumb to their cancer therapy. That means we do everything we can to deliver the greatest and latest scientific innovations close to their homes and their communities. So not only can they have effective treatment of their cancer, but they get to be at their dinner table, pick up their kids from soccer practice, sleep in bed next to their spouse, and work their jobs—they get through their lives without being compromised. I think protecting patients’ livelihood makes such a difference.
I've always been interested in health services research, and that quickly evolved into using digital tools for quality improvement and then AI for care delivery. AI has been a perfect marriage of where this can work effectively to help practices deliver on this dream of modern cancer therapy for the patients we serve.
What are some of the key ways AI can be used in clinical practice? How are AI tools being used at Texas Oncology?
Some of the easiest applications for AI include applying it to administrative tasks to decrease staffing burden and improve patient satisfaction. Like anything, you have to test it first in small populations, evaluate your results, and optimize the process for a broader roll out.
At Texas Oncology, we have an agentic pilot answering phones in a limited area. This allows patient calls to be answered faster, which reduces anxiety and diminishes the amount of time they have to spend on their healthcare. Additionally, it allows our transformation team to evaluate whether this solution requires modification and whether it is appropriate for broader implementation.
Another place where AI can be very helpful is in reducing administrative burden to allow clinical staff to operate at the top of their license. One way we use this today is by creating triggers to pull up triage management pathways. This allows our very talented nurses to have documentation and clinical guidance at their fingertips to respond to patient symptoms; it also eases decision fatigue and the burden of pushing it out to patients. By implementing clinical decision support in this way, we enable our nurses to work at the top of their license and improve their efficiency. This allows the decisions about care to remain with clinicians, and decision support reduces administrative burden and staff turnover. We are in a nursing shortage crisis, so these types of investments can be transformational.
We see similar efficiencies with medical scribe services where we are using agentic AI scribes to help clinicians with the documentation burden of clinical visits. This reduces documentation time in a practice and allows our clinicians to work at the top of their license and maintain a better work–life balance. This helps improve esprit de corps and mitigate burnout. I call it the “home for dinner” campaign.
What evidence do you look for from AI tools before you trust them enough to implement in your practice or workflow?
“What is the use case?” is the first question, because we have some pretty strict boundaries around use cases today. That's really important, because as we think about AI in health care, nobody wants it to be making clinical decisions—not at the nurse level, not at the doctor level. That is not the desire patients have, and it's not the desire organizations have. That is not where we are today.
Where we are today is that we want administrative burdens to be relieved, and we want our people to be able to work at the top of their license. We want to reduce decision fatigue and support their work–life balance. I recognize that some of these tools, especially in administrative use cases, can easily fall within boundaries we're comfortable with. So if they do have challenges—like bias, errors, and hallucinations, which we know are present with AI—those are generally manageable and don’t pose morbidity and mortality risks to patients.
I think the harder lift is clinical decision support, but that’s also some of the most exciting areas, too. One example is OpenEvidence. It is a wonderful tool for clinicians and can help generalist oncologists expand their subspecialty knowledge and expertise. That said, it is limited by the prompt you give it, so you have to ask the right questions and provide the right data to get the results that you want.
All of these tools, like any ancillary instrument, come with challenges in implementation and identifying optimal use cases. I believe we all need to be trained in how to use these tools—just as a carpenter learns to use their tools more efficiently and effectively—to ensure we’re using them appropriately to achieve our desired result.
As we think about AI in health care, nobody wants it to be making clinical decisions—not at the nurse level, not at the doctor level. That is not the desire patients have, and it's not the desire organizations have. That is not where we are today.
Where we are today is that we want administrative burdens to be relieved, and we want our people to be able to work at the top of their license.
- Debra Patt, MD, PhD, MBA
Where have you seen the greatest benefits and challenges with AI adoption in your practice, and how have your colleagues at Texas Oncology responded to the adoption of new AI tools—with more reticence or excitement?
There is both reticence and excitement around the implementation of AI tools. The biggest challenge has been change management. Digital tools are easiest to implement when they meet us where we are and when they include some degree of vendor-managed implementation. That said, sometimes we do need to change. We need to alter our workflow, but that is very hard to do.
We have hired change management experts at Texas Oncology and are working proactively with vendor solutions to do this better. When I started down this road, I thought “the road to Nirvana,” or the journey to a perfect practice, would be fraught with difficult vendors and integration challenges. While those are real issues, they are much easier to sort out than it is to manage change internally within organizations. Change management is the heavy lift here.
As we have managed these escalating needs in my own organization, we have hired experts to optimize change management and communication. They are critical assets to this transition. Looking at how we’ve implemented different vendor solutions in the past, we’ve become accustomed to bootstrapping it, doing everything ourselves. But that’s not the right answer because it strains our teams too much. It is not sustainable. Instead, we’ve learned that we need to partner and collaborate more closely with vendors to facilitate implementation. That is a better approach for us because it doesn't tax our teams as much and makes the change management process more effective.
What do you tell oncologists who are more concerned about using AI tools in clinical practice, and which areas tend to generate the most concern?
There are concerns, and they are real. Most of them center on quality and safety. This is why important guardrails must be in place. ASCO has crafted a wonderful policy on the responsible use of AI that can help guide this work.
Do you believe concerns about AI—and change more broadly—reflect a generational divide among practitioners?
I think about generational divides a lot in the context of our organization—when recruiting young doctors—and even when thinking about my professional society, ASCO, and how people interact with it. It occurs to me that younger generations interact with their profession differently—they want different things from their jobs and think differently about what they want in their life
I was born in 1974, so I'm Generation X. When I was growing up, I felt like I would forge the road ahead myself. Purposeful work was really important to me, and I wanted to work hard in my job.
It's been my experience now though, when working with younger doctors, that they have realistic desires around work–life balance and being part of socially progressive organizations. We need to adapt and change, and so I want to understand these differences better so we can pivot as needed, to ensure that we're continually evolving.
How can generalist oncologists use AI to keep pace with specialist-level knowledge and expertise?
There are many tools, such as OpenEvidence, ASCO’s Guidelines Assistant, and clinical decision support, that put care delivery solutions at our fingertips and allow generalist community oncologists to benefit from subspecialty expertise. I feel these tools make it easier to keep up with the latest advances in therapies, more so than it had been in the past.
As chair of ASCO’s AI Task Force, what are the top priorities for practicing oncologists right now?
The three priorities right now are:
Managing workforce shortages—we need all of our staff to operate at the top of their license
Managing burnout for a workforce with growing demand and administrative burden
Clinical trial identification so we can continue to innovate
To achieve this, ASCO is focused right now on how we can strengthen collaborations and how we can heighten awareness of available tools. ASCO has developed a partnership with Google Cloud and is using that partnership to create a more meaningful way to search guidelines, to extend their reach and put guidelines at clinicians’ fingertips. It is ongoing work, and ASCO is constantly trying to understand to best leverage that collaboration to serve its members and the patients we serve together.
At the Annual Meeting we can expect to see some foundational talks on AI along some very specific use cases—again with the goal of heightening awareness. Because what people don't understand is that they will face challenges navigating AI in oncology, but if we can heighten awareness as to what’s happening, that can help clinicians and their administrative partners better navigating this evolving landscape.
Do you believe greater AI literacy or training could improve the uptake of AI tools in practice?
I think there's a lot of work to be done in AI literacy training. Heightening awareness is the first step in giving people guidance on tools that have been effective and helping them adapt. Thankfully, oncologists are accustomed to learning new things. ASCO has taken a big role in that through the AI Task Force, this new platform ASCO AI in Oncology, and its educational programs. We are also now planning AI topics for discussion at the ASCO Annual Meeting. These are critical ways ASCO can help facilitate change.
I'm also the President of the Community Oncology Alliance, and innovation is the theme of our annual meeting this year. We want to put some of these agentic AI solutions in front of practices because they may not be meeting with vendors regularly, and may need a little help with their thinking as they begin to change and adopt some of these solutions. I think that will be transformative. As professional societies, we have an obligation to serve our communities by helping them envision what is possible and guiding them along this journey.
I was Editor-in-Chief of JCO Clinical Cancer Informatics for 6 years before serving on the ASCO board. That journal also plays an important role in raising awareness of digital tools and helping people understand what can be done. It presents research on what has been achieved, as well as editorials on how clinicians are thinking about digital tools, which is incredibly meaningful.
What operational safeguards should be mandatory before an AI tool is embedded into oncology workflow at scale? What regulatory frameworks are needed to facilitate that, and what barriers stand in the way?
There are many policies and they are necessary, though my bigger concern is that state and federal policies vary. I do think there will be a path to resolution. Also, there is no question that important care delivery decisions need to be made between doctors and patients. But AI, when used appropriately, doesn’t threaten that, it can enhance it—because doctors and nurses can have more meaningful interactions with patients when they are less burdened by administrative tasks.
There are many policies emerging around AI and digital tools used in clinical practice, and it can be a bit confusing, honestly. We work with many vendors that have a national presence, but state-level policies around disclosure and consent vary widely. That poses a real challenge in terms of compliance.
There's a real goal to see federal guidance emerge to serve as something universal so there’s no state-to-state variability. As we look to partner with vendors on implementation, we all want to ensure compliance solutions, and that inter-state variability can pose a real challenge. I would like to see a similar set of standards or parsimony in the standards and the compliance and regulatory frameworks.
I think that the administration has a real desire to do that—from the Office of the National Coordinator for Health IT (ONC), from the Centers for Medicare & Medicaid Services (CMS), from policies from the Executive Branch. I think that is emerging, but not as quickly as AI is evolving, and that gap presents a challenge for where we are today.
AI, when used appropriately, doesn't threaten [the doctor-patient relationship], it can enhance it—because doctors and nurses can have more meaningful interactions with patients when they are less burdened by administrative tasks.
- Debra Patt, MD, PhD, MBA
Over the next 3 to 5 years, where do you expect AI to have the greatest and most meaningful impact on cancer care in the community setting, and what is needed big picture to get us there?
I think managing communication and the workforce will be the biggest opportunity—and challenge. We are in the midst of a staffing crisis, and it impacts practices in a myriad of ways. We need to make this better, and we will.
As we look at the current landscape, we are facing a workforce shortage alongside a growing number of patients driven by an aging population. More patients with cancer are surviving because of the wonderful investments and innovation in cancer care. As a result, we need to sustain a workforce capable of treating this growing patient population. I think that AI and digital transformation will be essential to meeting those needs, to helping us keep up. We must adapt by using our staff more efficiently and extending our reach to better serve our patients and meet their needs.
Our staffing solutions will need to look quite different. There will need to be whole new departments focused on transformation and informatics. I don’t expect that we will employ fewer people; rather, we will have the same number of people in different roles, with many focused on the optimal use of technology and organizational change management, because it’s a heavy lift.
ASCO AI in Oncology is published by Conexiant under a license arrangement with the American Society of Clinical Oncology, Inc. (ASCO®). The ideas and opinions expressed in ASCO AI in Oncology do not necessarily reflect those of Conexiant or ASCO. For more information, see Policies.