Authority Magazine
October 6, 2024
Invest in training and upskilling — Ensuring teams are equipped with the necessary skillset to work alongside AI is critical for success and responsible use of the technology. This includes comprehensive training on the foundation of AI, implementing AI, and special use cases of AI in healthcare.
Artificial Intelligence is no longer the future; it is the present. It’s reshaping landscapes, altering industries, and transforming the way we live and work. With its rapid advancement, AI is causing disruption — for better or worse — in every field imaginable. While it promises efficiency and growth, it also brings challenges and uncertainties that professionals and businesses must navigate. What can one do to pivot if AI is disrupting their industry? As part of this series, we had the pleasure of interviewing Joseph Zabinski.
Joseph Zabinski, PhD, MEM is VP, Head of Commercial Strategy & AI at OM1. He oversees design, implementation, and evolution of OM1’s commercial strategy, including prioritization, go-to-market planning, and alignment with emergent market needs. Dr. Zabinski also serves as subject matter expert and thought leader for OM1’s AI work, including applications of the PhenOM digital phenotyping platform with life science and healthcare provider partners. Prior to joining OM1, Dr. Zabinski was a consultant in the Pharmaceutical and Medical Products Practice at McKinsey & Co. He specialized in advising life science and healthcare clients on using AI and advanced analytics, including identification of unmet medical needs, portfolio optimization, and AI strategy. His academic research focused on applications of Bayesian network modeling methods to predicting and stratifying environmental human health risk. Dr. Zabinski holds a master’s degree in engineering management from Dartmouth College, and a doctorate with a focus in health analytics from the School of Public Health at the University of North Carolina, Chapel Hill.
Thank you so much for joining us in this interview series. Before we dive into our discussion our readers would love to “get to know you” a bit better. Can you share with us the backstory about what brought you to your specific career path?
Mypath to my current role at OM1 was shaped by a unique blend of experiences in healthcare, technology, and data analytics, combined with my goal of bridging the gap between emerging AI technology and real-world clinical applications. I got started in this world as a methods person — I studied mathematical modeling, originally in applications in the energy industry — and fell in love with healthcare research during grad school. To me, the ability to use data and AI to generate better answers to real questions around diagnosis, risk, and treatment remains something like magic, and I love working to make that magic real and sustainable.
What do you think makes your company stand out? Can you share a story?
In healthcare, outcomes are critically important and can often be a matter of life or death. Accessing and making full use of the comprehensive data needed to understand a patient’s story has been challenging, and often where OM1 has been able to help the most.
We stand out because of our data, which includes more than 350 million patients, as well as our unique AI-powered digital phenotyping technology that we use to identify patient populations and hidden patterns to predict health outcomes. Beyond the tools we offer, we take pride in the quality and insight we provide to our clients and partners. We’ve built a technology company with deep roots in the clinical space, and I’ve found our ability to translate what we do to physicians and patients to be distinctive and a key advantage we have.
For example, we partner with clinical associations like the American Academy of Dermatology (AAD) and the American Academy of Otolaryngology–Head and Neck Surgery (AAO-HNS) to create rich, linked datasets that we can use to understand patient journeys as accurately as possible. Throughout the past year I’ve led our work with the AAD in using digital phenotyping technology to improve diagnosis in generalized pustular psoriasis (GPP), a rare but serious skin disease. These patients can go for years without an appropriate diagnosis and consequent inability to access specialty care and treatment. With digital phenotyping, we’ve been able to highlight patients at much greater risk for GPP, giving clinical teams a place to start when trying to reduce underdiagnosis. As AI continues to mature, it is essential for clinicians to understand how they can be involved in data generation and use so that we can better calibrate our models to personalize assessments around diagnoses, risk, and treatment response.
Our goal is to empower stakeholders to cost-effectively access, analyze, and use outcomes data at scale in more robust, clinically meaningful, and precise ways — both individually and at the population level — ultimately driving innovation and the frontier of clinical care.
You are a successful business leader. Which three character traits do you think were most instrumental to your success? Can you please share a story or example for each?
- I think maintaining a simultaneous focus on long-term strategy and short-term tactics has been critical, especially in an emerging field like AI. Understanding the potential of this technology in healthcare can be so mind-bending that it’s easy to get distracted from the day-to-day task of making incremental progress, but you also need vision to keep motivated and focused on the path ahead.
- I do well in fluid, dynamic environments with a ‘north star’ goal and a lot of flexibility in how to get there. I’ve been lucky to be able to work in environments where this skill is useful, and it’s helped me solve some challenging problems when the most direct path to the solution gets blocked. As my industry matures and my own career advances with it, I’ve found that a flexible approach to achieving goals is key to connecting emerging AI technology to real-world applications in the generally conservative field of healthcare.
- Finally, I try to cultivate patience — with clients, colleagues, technology, and myself! I am a naturally impatient person and have found that tempering that tendency has let me keep a motivating energy and drive towards results, while improving my interactions with others and the core business and technology problems I work on.
Let’s now move to the main point of our discussion about AI. Can you explain how AI is disrupting your industry? Is this disruption hurting or helping your bottom line?
Getting a diagnosis can often feel like a long and difficult journey for patients. And even after reaching a diagnosis, developing a personalized treatment plan can be challenging. That’s where automation and AI can really help. By applying advanced algorithms to real-world healthcare data, clinicians can better understand disease-signaling patterns in undiagnosed patients, which speeds up the diagnostic process and leads to more effective treatment plans. With specific insights into a patient’s likelihood to respond to a therapy, clinicians can tailor treatment recommendations, significantly improving patient outcomes and access to care while decreasing ineffective use of treatments.
AI tools are also expediting research by facilitating the identification of potential clinical trial participants that meet the criteria needed for a trial, but are overlooked due to the constraints of traditional enrollment methods. This kind of application can accelerate trial recruitment, and help prevent trial failure through insufficient enrollment.
Overall, the AI tools we use make research and clinical interventions more precise and personalized, answering questions we couldn’t without these tools and improving efficiency. This leads to a positive effect on the bottom line.
Which specific AI technology has had the most significant impact on your industry?
Machine learning and deep learning algorithms have had a significant impact on healthcare in recent years by providing clinicians with a synthesized view of patterns across patients with similar conditions to understand earlier who will end up with those clinical manifestations. By bridging population data to the personal level, advanced algorithms can also determine which treatments will have the greatest efficacy for patients and if they are likely to be adherent or experience adverse events.
ChatGPT and tools like it are also changing the industry. First, these tools can function as really powerful summarizers and synthesizers of information. For patients, this could mean greatly expanded access to information that’s hard to find or absorb now. For example, a ChatGPT-type tool could describe what to expect when newly diagnosed with a condition, or starting a new medication, in language that’s clear and meets the patient where they are. Physicians could benefit from this kind of summarization too, particularly once these tools are able to access real-time data and compile the ‘latest and greatest’ in emerging knowledge and treatments.
Powerful generative AI tools are also helping with ‘intelligent documentation’ — assembling patient histories and new findings into a clinical note, for example. Down the road, these tools may be used to more quickly and reliably extract information from patients’ medical records to help generate personalized assessments and treatment plans.
Can you share a pivotal moment when you recognized the profound impact AI would have on your sector?
Several years ago, I was about to speak at a conference on the theoretical impact of AI when I got a message from our team. We were running a pilot implementation study of our technology for finding undiagnosed patients, and the first results of confirmatory diagnostic screening had just come in: patients we had flagged had completed screening, and tested positive. These patients were otherwise ‘off the radar,’ and because of AI, they now had received an accurate diagnosis (and access to potentially life-saving treatment).
How are you preparing your workforce for the integration of AI, and what skills do you believe will be most valuable in an AI-enhanced future?
As generative AI tools improve and proliferate, understanding how to use these tools to enhance daily workflow will be critical. We’re moving away from an era of ‘science experiment AI’ to a future where these tools will be fairly mature and widely available. Getting familiar with how they interact with existing tasks and goals, as opposed to being totally new technologies unrelated to the day-to-day, will be critical.
In addition to preparing our workforce with core competencies, we are also focused on the industry-wide need for equity in medicine. That is why we’re working towards criteria to better understand AI bias, to ensure alignment between clinically-oriented AI solutions and medical facts. We have made a large investment in building AI explainability technology that allows our in-house as well as our external users to ‘open the box’ and observe which data are being used by AI models to reach outputs. This might sound easy, but AI models are complex and optimized for prediction or estimation rather than transparency. Making them transparent enough to ensure usability in the real world, while maintaining analytic performance, has required creativity and a lot of hard work on the part of our team.
What are the biggest challenges in upskilling your workforce for an AI-centric future?
Some obstacles organizations may face are lack of foundational knowledge, ability to seamlessly integrate AI technologies into current practices, change resistance, access to training, and ethical considerations, to name a few. I think teams can reduce these challenges upfront by establishing a solid foundation of understanding before implementation, building a robust change management strategy, allotting investment dollars for continuous education, establishing definitions and criteria of AI bias, and most importantly, continuing to have humans in the loop to ensure explainability of the data when leveraging AI in practice.
What ethical considerations does AI introduce into your industry, and how are you tackling these concerns?
AI tools generate outputs based on the data on which they’re trained. So, if these data is not fully representative of the world that the AI system operates in, or misrepresents that world, outputs can be biased. No dataset is perfectly representative, and no AI model is without some biases.
To combat this, we are tackling ethical concerns around bias through our investments in AI explainability technology. This involves interdisciplinary collaboration, and the integration of health records with diverse data sources to create large, enriched datasets that are representative of various patient sub-populations. This approach fosters more equitable research, helping to ensure medical advancements benefit all demographics, and the future of medicine is more inclusive and personalized to the needs of every patient.
What are your “Five Things You Need To Do, If AI Is Disrupting Your Industry”? If you can, please share a story or an example for each.
1 . Address barriers to access and acceptance — If we do this, the gap between what AI can do in theory, and what it actually does in the real world can begin to close. AI technologies are reshaping healthcare at a speed never seen before. What are the risks and opportunities for innovation in clinical environments and beyond? The industry needs to be specific about the value AI creates; efficient and seamless in its delivery for providers and patients; and resolute in the use of insights to help make decisions — only then will we see acceptance and the next step of adoption.
2 . Compare what AI can do to the world as it is — For example, we must compare AI to real patient journeys, real diagnostic procedures and (lack of) uptake, and real treatment decision-making under tremendous uncertainty. AI will always disappoint if we expect it to be perfect, but it can be incredibly helpful if we use patients’ lived experiences as our baseline to create well-supported solutions that address real problems. You want to be mindful of leveraging data effectively (think collecting and analyzing) to inform point-of-care and treatment decisions as well as improve the AI solutions in clinical environments.
3 . Recognize that data is a first level of defense against AI bias — As mentioned before, AI relies on data to learn, so feeding the training of algorithms with the richest, deepest, most holistic, most diverse and broadly representative datasets possible is critical to greatly increasing the objectivity of AI algorithms and reduces bias. Organizations seeking to integrate AI into their operations should develop criteria through collaboration with AI researchers, domain experts, and business stakeholders. The latter would set up the bias criteria and the AI expert would create the technical solutions to monitor it, constantly improving their AI solutions to minimize biases.
4 . Invest in training and upskilling — Ensuring teams are equipped with the necessary skillset to work alongside AI is critical for success and responsible use of the technology. This includes comprehensive training on the foundation of AI, implementing AI, and special use cases of AI in healthcare.
5 . Learn where AI doesn’t make sense — Sometimes the ‘disruption’ we see from AI is actually the result of attempts to use it where other tools would be more appropriate. Recognizing these situations early — ideally, before they happen — can save a lot of trouble and friction, and preserve team members’ belief in AI to be useful.
What are the most common misconceptions about AI within your industry, and how do you address them?
Often, stakeholders who are hesitant to adopt AI in healthcare may feel the advanced technology is replacing humans in clinical settings. However, this is not the case or the goal of tech-enthusiasts in the industry. AI should be seen as a great add-on to clinician workflows — increasing efficiency and reducing administrative burden — versus a replacement of doctors when patients seek in-person care.
‘Disruption’ can be valuable in advancing the state of clinical medicine, but we can’t disrupt ongoing care in ways that cause friction or harm for patients or increase the burden on clinicians.
Can you please give us your favorite “Life Lesson Quote”? Do you have a story about how that was relevant in your life?
There are quite a few, but here are two simple ones, both from my dad. The first is something he used to say to me when I wanted to rush through my homework or chores to move on to something more fun: ‘if you don’t have time to do it right, how will you have time to do it twice?’ There’s a fine line between efficiency and cutting corners, and that phrase helps me remember to stay on the right side of it.
The other comes as part of the territory with a last name that starts with Z: whenever people get sorted by name, I end up at the back. In circumstances like that my dad’s philosophy was ‘somebody’s got to be last’. I keep that in mind to remember we can’t control everything — we should keep focused on what we can achieve and avoid getting distracted by what’s beyond our control.
Off-topic, but I’m curious. As someone steering the ship, what thoughts or concerns often keep you awake at night? How do those thoughts influence your daily decision-making process?
I think a lot about health data and technology being used well, and not carelessly. This sounds simple, but the impact of what we do when our work helps inform medical decisions can be enormous, and there are many ways for it to go wrong. It’s a combination of worrying about the underlying data, the technology, and the people — it drives me to work hard on these different fronts to make sure what we’re producing is really high-quality.
You are a person of great influence. If you could start a movement that would bring the most amount of good to the most amount of people, what would that be? You never know what your idea can trigger. 🙂
I am a huge fan of the concept of disease eradication. I love the idea of using technology and data to help eliminate conditions that cause suffering. We’re not going to get there anytime soon with diseases like diabetes or cancer, but especially with infectious disease I think there are real opportunities to do a tremendous amount of good and make life better, permanently! For an inspiring view into this world, check out Jimmy Carter’s work on eradicating Guinea worm disease.
How can our readers further follow you online?
You can follow me on LinkedIn and check out the latest with OM1 by visiting its website or social profiles on LinkedIn and X.
Thank you for the time you spent sharing these fantastic insights. We wish you only continued success in your great work!