Synthetic Data in Healthcare: A Clinician’s Guide

Loading page

Beyond the Bedside: How Synthetic Data is Arming Clinicians for Healthcare’s Next Frontier

Healthcare is drowning in data yet starving for insight. In an industry that generates an estimated 30% of the world's data volume, frontline clinicians—the very individuals trained to synthesize complex information and make life-or-death decisions—are often the last to gain meaningful access to the operational and population-level data generated by their own work. This paradox creates a frustrating chasm between the clinical teams at the bedside and the administrative leaders in the boardroom, hindering innovation, perpetuating inefficiencies, and ultimately impacting patient care. The traditional pathways to data, guarded by overwhelmed analyst teams and months-long regulatory approvals, are no longer sufficient for a system demanding agility and precision.

Bridging this chasm requires more than just new technology; it demands a new philosophy. It requires leaders who can speak the languages of both patient care and data science. Trushar Dungarani, DO, SFHM, Director of Clinical and Data Science at MDClone, embodies this new archetype of clinician leader. With a journey that has taken him from the front lines of hospital medicine at Northwell Health and Johns Hopkins Medicine to the forefront of data innovation, Dr. Dungarani has witnessed firsthand the "beauty and the dysfunction" of modern health systems. His work now focuses on empowering organizations with a transformative tool that promises to democratize data, protect patient privacy, and unlock the insights trapped within the electronic health record (EHR): synthetic data.

The Clinician’s Dilemma: A Culture of Data Distrust

For many healthcare professionals, the relationship with organizational data is fraught with frustration. Clinicians are trained to be masters of data interpretation for a single patient, but when it comes to system-level performance, they are often handed reports and directives with little context or ability to investigate the source. Dr. Dungarani explains that this top-down approach, combined with a lack of access, breeds a natural skepticism.

"These clinicians aren't getting access to data, so they're being told something without that ability to see, see the data," Dungarani notes. "There's always a kind of a defensiveness to it… we have to teach our providers and our clinical team members about data, give them access and, and let them make changes for themselves."

This challenge is rooted in several systemic issues that Dr. Dungarani identifies from his experience in quality, operations, and health information management:

Fragmented Systems: Health systems often consist of patchworks of technology. Data from a previous EHR may not be integrated into the current one, making longitudinal studies nearly impossible without significant manual effort.
Workflow Silos: Data is often trapped within departmental workflows. An emergency department physician may have no visibility into inpatient bed capacity, even though the two are inextricably linked. This makes it difficult for teams to understand and solve cross-functional problems like patient throughput.
The Bottleneck of Access: The traditional process for obtaining data is fundamentally broken. A clinician with a research idea or an operational question must submit a request to a small team of data analysts, who often have a backlog of projects from across a 30,000-person organization. The wait can be weeks or months, extinguishing the spark of curiosity and innovation before it can even ignite.

This environment not only stifles quality improvement but also contributes to clinician burnout. Being held accountable to metrics without the tools to understand or influence them is a recipe for disengagement. Empowering clinicians means giving them the agency to solve the problems they see every day, and that agency begins with access to data.

Forging a New Path: The Intersection of Leadership, Culture, and Technology

Solving the data access problem is not merely a technical challenge; it is a leadership and cultural one. According to Dr. Dungarani, effective change requires clinician leaders who can act as translators, bridging the gap between administrators, data scientists, and frontline caregivers.

Bridging the Great Divide

Administrators, clinicians, and data analysts often look at the same problem through vastly different lenses. The CFO sees cost, the pharmacy team sees medication adherence, and the physician sees a patient's clinical journey. A successful data-driven culture requires a "common language" to align these perspectives.

This alignment is crucial because the narrative told by the data and the one experienced by the clinician can often diverge. "A clinical workflow and a clinical story can be very different from a data journey," Dr. Dungarani explains. "The assumptions and the expectations of how data flows with a patient story need a lot of communication… you need that same type of multidisciplinary team to really communicate how that data is telling that story."

For example, a report might show a delay in administering a specific medication. The data simply shows a timestamp. The clinical story, however, might involve a patient who was temporarily off the floor for a procedure, a pharmacy delay, or a nurse attending to a more critical emergency. Without the ability for clinical teams to easily explore the associated data points themselves, incorrect assumptions are made, and ineffective solutions are implemented. Clinician leaders are essential in facilitating this dialogue, ensuring that data insights are grounded in clinical reality.

The Leadership Mandate

Ultimately, fostering this culture starts at the top. Leadership must transition from a compliance-oriented mindset—"Your throughput times are terrible, fix it"—to an empowerment-oriented one: "Here are the tools to explore the data. Find the bottlenecks and tell us what you need to solve them." When teams are given the tools to identify problems themselves, they become far more invested in creating and sustaining the solution. This shift democratizes problem-solving and builds a true learning health system, where improvement is a continuous, collaborative effort rather than a series of top-down directives.

Demystifying Synthetic Data: A New Frontier for Privacy and Innovation

If the goal is to provide widespread access to data, the primary obstacle has always been patient privacy. The Health Insurance Portability and Accountability Act (HIPAA) and other regulations rightly place strict controls on the use of protected health information (PHI). The traditional method for creating "safe" data for research has been de-identification, a process of stripping 18 specific identifiers like name, address, and Social Security number. However, this method is increasingly seen as inadequate. In a world of big data, re-identifying an individual from a "de-identified" dataset can be surprisingly easy with enough external information. Furthermore, to be safe, de-identification often requires removing or aggregating key variables—like zip codes or specific dates—that are critical for studying health equity or disease progression.

What is Synthetic Data?

Synthetic data offers a far more robust solution. Instead of simply removing identifiers, a synthetic data generation process studies the entire original dataset to learn its statistical patterns, correlations, and distributions. It then generates a brand-new, artificial dataset that mirrors these statistical properties but contains no one-to-one mapping back to any real patient. Imagine it as creating a statistically identical "twin" population. One row of synthetic data does not represent a real person, but an analysis of 10,000 rows will yield the same conclusions—the same rates of diabetes, the same average length of stay, the same correlation between a lab value and an outcome—as an analysis of the original 10,000 patients.

This approach offers the best of both worlds: it provides a high-fidelity dataset for exploration and analysis while completely severing the link to individual identities, thus protecting patient privacy at its core.

Accelerating the Pace of Discovery

The operational impact of this technology is immense. It effectively removes the regulatory and logistical logjams that plague healthcare research and operations.

"The benefits of synthetic data right now in the healthcare system is a, I don't have to jump through a million hoops to look at whether my study is feasible," says Dr. Dungarani. "It may take weeks or months to get a data set to even see if your project is viable. Do I have a magnitude of patients to do this research project?"

With a self-service platform built on synthetic data, a clinician can test that hypothesis in minutes. They can explore the data, refine their research question, and confirm the viability of a study before ever applying for Institutional Review Board (IRB) approval to use the original patient data. This dramatically speeds up the innovation cycle, allowing teams to fail fast on ten ideas to find the one that warrants the time and resources of a full-scale study. It transforms the data analyst from a gatekeeper into a strategic partner, freed from routine data pulls to focus on more complex, high-value projects.

Synthetic Data in Action: From Predictive Models to Health Equity

The applications of this approach are not theoretical. Health systems are already using synthetic data to drive meaningful improvements in patient care and operational efficiency.

Predicting Chronic Disease Progression

Dr. Dungarani shares the example of a health system working to predict which patients with chronic kidney disease (CKD) are most likely to progress to dialysis—an outcome that profoundly impacts a patient's quality of life and represents a significant cost to the system. Using synthetic data, their teams could rapidly build and test machine learning models to identify the key predictors of progression. This allows them to create targeted interventions for high-risk patients, potentially delaying or preventing the need for dialysis and improving long-term outcomes. Similar work is being done to predict which cardiac patients will develop severe heart failure, enabling proactive care management.

Closing Gaps in Healthcare Equity

Perhaps one of the most powerful uses of synthetic data is in the study of healthcare disparities. Analyzing sensitive demographic variables like race, ethnicity, language, and geographic location is essential for identifying and addressing inequities, but it also carries the highest risk for patient re-identification. Synthetic data provides a safe harbor for this critical work.

Dr. Dungarani describes clients using synthetic data to examine geographic "heat maps" for correlations between environmental factors and chronic disease or to analyze whether a patient's primary language or ethnicity impacts the time it takes to get to the operating room for a common procedure like a hip fracture. This type of exploration is vital for ensuring equitable care delivery for all populations. The self-service nature of these tools fosters an iterative, curiosity-driven process that is fundamental to discovery.

"We can ask a question, find an insight, and with anything in healthcare, there are 10 more questions or are 20 more questions," he remarks. "Having that self-service ability to look at synthetic data and ask more questions and really have this dialogue with data is really the power."

Key Takeaways for Healthcare Leaders

For clinician leaders who want to champion a new era of data-driven healthcare in their organizations, Dr. Dungarani's insights distill into several actionable principles:

Champion Data Literacy: Acknowledge that healthcare has one of the lowest rates of data literacy among all industries. Invest in training that creates a common language and understanding of data workflows between clinical, analytical, and administrative teams.
Democratize Data Access (Safely): Move beyond the gatekeeper model. Explore and invest in technologies like synthetic data platforms that can empower your teams with self-service analytics without compromising patient privacy or overwhelming your existing data infrastructure.
Cultivate Curiosity, Not Compliance: Foster a culture where data is a tool for exploration, not just a report card for compliance. Encourage teams to ask questions, test hypotheses, and follow their curiosity to uncover the root causes of systemic challenges.
Embrace the AI Frontier: The rise of AI and machine learning in healthcare is inevitable. Clinicians must be active participants in shaping how these tools are developed and deployed. This requires a willingness to learn, experiment, and engage with new technologies.

Conclusion: Be a Part of the Change

The future of healthcare will be defined by how well organizations can leverage their data to deliver more personalized, efficient, and equitable care. Achieving this vision requires more than just algorithms and dashboards; it requires a profound cultural shift that places the power of data into the hands of those on the front lines. Synthetic data is a key enabling technology for this shift, providing a safe and agile framework for innovation.

As Dr. Dungarani passionately concludes, skepticism in the face of new technology is natural, but inaction is a choice. The path forward demands engagement and a proactive spirit from today's and tomorrow's healthcare leaders.

"Without trying, without learning and, and, and being curious, you know, we aren't gonna move forward," he urges. "So I would say if you doubt it, study it, learn it, and, and figure out where the gaps are and be a part of the change."

without trying, without learning and and and being curious, you know, we aren't going to move forward. So, I would say if you doubt it, study it, learn it, and and figure out where the gaps are and be a part of the change. [Music] Hello, this is Zach McConnell with the American Journal of Healthcare Strategy and you are listening to the clinicians and leadership podcast series where we focus on empowering clinicians from the bedside to the boardroom. Today I have the esteemed honor of being joined by Dr. Trushar Dungarani. Dr. Dungarani, why don't you take a second to introduce yourself? Tell us a little bit about your experience and and your current role. Great. Yeah, thank you for having me. So my name is Trousard Nugarani. I'm the director of clinical and data science at MD clone for our North and South America divisions. And my background is in internal medicine. I'm a physician. I practice hospital medicine. And in my in my the beginning of my career, I was in New York City. I worked at Northwell Lennox Hill. And my my journey started with in an academic setting. And so I taught a patient safety curriculum. I led a hospital medicine group there and then I transitioned to Maryland where I worked in the John's Hopkins Health System and in addition to clinical care I had a big role in our community division from both a quality and operation standpoint and my focus is really to leverage people process and technology. Um my my favorite part of working there was working on high value care delivery for the health system and I got the opportunity to work in machine learning projects um work as a physician adviser and I chaired a health information management um committee. So all that summed up I know the beauty and the dysfunction of a health care system very well. Um, currently at MD Clone, I'm focused on working with healthcare organizations to empower them to use their own data, whether it's original or synthetic, to be able to create change, whether it's um putting out a research paper or driving a quality or operational case and making sure that patient data is safe but still usable. Well, Dr. Dr. Dunarani, we're we're grateful for you joining us today on the Clinicians and Leadership podcast and and particularly I'm I'm excited to get to talk to you and to pick your brain. You've got a unique uh collection of experiences in in a variety of different areas both both serving as a clinician um in in internal medicine teaching. You've mentioned that, but also I'm I'm interested in the interplay between those and uh your your more recent work interacting with data um and and you mentioned synthetic data. We're going to dive into that and what that is a little bit later, but particularly I I love the way how you described the uh your role is is kind of stepping in and empowering organizations to use their data that they they have well. Um and and and so as a result that that necessitates that that organizations like you like you described aren't using their data well. And so what are some of the biggest barriers that that clinicians face today in accessing and utilizing like real time highquality patient care data at the bedside? What what are some of those biggest barriers that you help these organizations navigate? Yeah. Um that's a that's a great question. I was just talking to my colleague. We were wondering why aren't people using more tools? I think it starts with you know our training. whether you're a nurse or a physician, we're not taught to use data in a very um I guess systematic way or we don't have access to data. So, they're not telling us, hey, when you start your job, make sure you get this data set. Make sure you you look at all these variables because it's just not accessible. And so, off the off the bat, we are not in that mindset of asking for things. we're we're given uh reports here and there and we're we're supposed to interpret it and and make decisions with very few data points. Um when when you come into a health system, they may have started on paper charts and then moved to one EMR and then the other. So there's a lot of data fragmentation. And so when you're at your latest and greatest EMR, that EMR doesn't contain legacy data. And so what do what do we do? How do we study things over a long period of time? Um, you know, the other part of it is is workflow. If you are working in the ER, you may not know what's happening on the inpatient side, but anybody who has done hospital operations understands the ER is impacted by the inpatient unit and and vice versa. And so studying it, working in it, and understanding the kinks and the nuances of it is is a barrier for people, especially early on in their career. Um and I think probably lastly is there is an inherent distrust in in data. If you are told you are not doing well and your patient experience scores are low or your patients are getting readmitted, there's always a kind of a defensiveness to it because these clinicians aren't getting access to data. So they're being told something without that ability to see see the data. And so we have to ensure as a health system and as a kind of a broader um healthcare delivery network that we have to teach our teach our providers and our clinical team members about data, give them access and and let them let them make changes for themselves. Well, I I think that's such a key point, Dr. Dr. Dunani that that you brought up and and just the the lack of access that that some of those frontline medical workers can can have to the data particularly when when you're talking you know clinicians and more specifically physicians that they are some of the the best in the world at taking data interpreting it and then planning next steps because that is what not only have they they went through all those years of school and training to do but but that is their job and if they don't do that job well then then patient care suffers and and ultimately they're they're out of a job. And so it's it's not utilizing the the skill sets of your medical staff to the to the fullest to to when you're when you're uh not equipping them to with that data um and and not giving them access to that data. And so kind of coming off of that, we we talked about some of those barriers that that organizations may face. Um but but kind of on the flip side, how how can your clinicians and leadership and other leadership ensure that those frontline providers do have access to that data to that data that they are engaged that they are equipped to to use those tools that you mentioned a little bit uh throughout that last answer to use those effectively? How how can they ensure that they're equipped to do what they need to do? Yeah, that's um that's a that's fascinating thinking about, you know, what does it take, you know, is it is it just access to data? Is it teaching them something? And and I believe it starts with you leadership, right? You have to have a team of people who believe that if we want to be the best in the world, we need data. We need to follow the data and and and make sure we can interpret it in the right way. And so, you need tools. You need a process to say all right am I going to expect um you know my my cardiothoracic surgeon to make a decision about you know why the O is taking longer without giving them any tools. Um no you you have to you have to give them the time you have to give them the access to tools and then encourage them to to find the issues. Um, everything starts with culture and and if you're going to create a change in in a health care system, it's multid-disciplinary. And if you give them the tools to find the problems, they are they are much more willing to create that change versus somebody top down saying, your your O throughput time is terrible. You have to change it, you know, go tackle it. and then you're you're kind of stuck there with um few data points and and you're kind of guessing on what really needs to change. So so so it starts at the top. You have to give them the tools and and then really you know start to learn create that culture of learning and and every every year health systems ask are you are you given the right tools to you know do your job and I think consistently having access to data is one of the biggest challenges in healthcare and that Dr. Dr. Dunarani, I think is is a fascinating area where particularly your clinicians and leadership can step in. Um because because you've you've got administrative teams that that have the data. They may not know exactly how to use it or maybe not know exactly what it means, but but they are the ones that are gathering that data and then you have your your clinical staff that are making the decisions for the patient care. And and both of those teams need to work together. Um, and far too often there's a gap in between them. And that is where I think your your clinicians and leadership can step in. And so before we dive into kind of the rest of our talk, I'm I'm curious to hear your thoughts on how the how clinicians and leadership can build and bridge this gap between administrators, IT, data science teams, and the frontline commission or clinicians, excuse me, to to maximize data usage and then as a result promote high quality patient care. How how can those unique individuals with the clinic clinical and administrative experience bridge that gap effectively? Yeah, I I think it's um a few things. One is a common language of of all right, let's let's make sure apples are apples, oranges are oranges. And a physician may not interpret data the same way that you know a quality department looks at it or you know the pharmacy team and the CFO look at cost at a very different standpoint. So make sure making making sure that we get on the same language and the same um you know set of rules and and though the priorities may be different for different groups there is a common language of hey uh we want to do this specific intervention because a it's going to save us money it's going to improve patient care and it's going to change the way we're delivering care in our health system and that that's something that you know different types of groups can can really attach to Um I I I think you know the other part of it is how do we you know create that that that incentive that that that motivating factor. Some people are motivated by praise. Sometimes it's getting time to do uh put out a research publication. Um you know in other ways people love hearing patient stories. It makes them feel good about why they're in in this field. So we really have to tell a variety of stories so that different types of people in different types of roles can really you know you know build on that on that mission to to to drive forward. Um and I think the you know the other part of it is you know a clinical workflow and a clinical story can be very different from a data journey. Um the you know the assumptions and the expectations of how data flows with a patient story uh you know need a lot of communication and just like changing something on the on the floor requires a multi-disiplinary team you need that same type of team to to really communicate how that data is telling that story because there is a lot of data in that EMR whether it's your current one or legacy one and we have to make sure that we're telling the right One of the things you mentioned earlier in in really I think it was your first your first answer and just telling us a little bit about yourself and your experience was this this concept of synthetic data. And so I'm willing to bet that there's there's a number of people out there that that don't know what you mean when you say synthetic data. And so um I guess my question is what is synthetic data? And and how can clinicians and leadership be be confident and build confidence? So not just confident in themselves but also build confidence in the staff that they're working with regarding the the accuracy of synthetic data. So so what is it and how do we build confidence in it? I think you know the story of synthetic data and we'll just we'll just talk about it in healthcare is why do we have synthetic data? Why is it even a topic? And and I would say the few items that make synthetic data important to think about and and conceptualize is we are trying to protect patient privacy. Right? That's the number one. Let's not expose data. The other part of it is we need data to make change in healthcare. To look at original patient data, you need to go through regulatory steps to use it. And you know, rightfully so. We don't want everyone having access to all of the data because we want to be judicious about that process. Um so so now jumping to what synthetic data is. Uh the current standard for protecting privacy is deidentification. So that means we're going to look at, you know, 18 variables that that are deemed to be high risk, whether it's your name or social security or your address or your zip code, and we're going to pull that out and then give you the data set. Now, does that ensure privacy? Not necessarily. You know, there's a lot of talk about how deidentification doesn't really uh protect it in the most complete way. And so now when we look at synthetic data, it's it's an artificially generated data set, but it is scrambling the data around so that you can't look at one row of data and say this is all the characteristics of one patient. But what you can do is say the statistical properties of one column are are exactly um the same as the original pa patient data set when you're comparing things like age and and um the risk of diabetes or how long did they stay in the hospital and and what zip code are they coming from. So we don't have to necessarily take away data. we just scrambled it in a way so that we can still use important data features uh to make change for our patients. So, I'm I'm curious coming off that you walked us through and I thought that was a great answer just on what synthetic data is and and how it can be used, but um but I'm I'm curious to dive a little bit more into the the role that synthetic data, particularly the the privacy protected synthetic data sets, what role that those sets play in in clinical research. Um and how do they compare to the more traditional patient data sets that that organizations use? um and particularly in terms of reliability and effectiveness. So, so how what role does it play in and how does it compare? Yeah. Yeah. Um so if you think about you know at the end of the day you're trying to get insights and and and protect privacy and so if you protect privacy too much if you a extract too many data points from there you're not going to get insight. On the flip side, if you kind of leave the floodgates open, you're not really protecting patient privacy. So, there is a balance there. Um, and so that right balance is is are we going to, you know, protect our vulnerable populations. Maybe it's a rare disease or something um you know a very small uh amount of community members have or maybe it's in a specific you know ge geographic region. um we have to ensure that we are not exposing that. And so with synthetic data you can you can do that by by scrambling the data around and and there are limits on on what and and how you can study. There are strategies to mitigate small populations but uh you know the the end point is let's let's make sure that we're we're um you know keeping their data safe and private. So the the benefits of synthetic data right now in the health care system is a I don't have to jump through a million hoops to look at whether my study is feasible. Um it may take weeks or months to get a data set to even see if your project is is viable. Do I have a a magnitude of patience to you know do this research project or or create a new service line or um you know make some sort of quality or operational change. um what if we could speed that process up, get insights about our patients and then say, you know what, I I might need original data. So, let let's let's um you know, apply for IRB or REB um to get that data. And now we're not wasting our time on with five or 10 ideas that may take weeks, months, or years to, you know, to get a data set for. Now we can test this ourselves and and and and get the answers so that we can make the most of our time. Um you know the the other part of it is is health institutions have limited resources. um to get data today you need to ask an analyst to to pull a data set using SQL and some health institutions have five people or 10 people for 10 20 30,000 uh employees in there. So you can imagine 95 97% of people aren't asking for data sets because they they can't wait in line. They're not going to get prioritized and it just takes too long to to cycle through that. So if we can empower our our team members with self-service data to answer these questions, then we are making more efficient use of our analysts of our team members and and we're actually creating a learning system and and I think that's such a key component and that just speaks to to I mean what you mentioned at at at the beginning of our conversation is empowering those organizations to use the data that they already have. Um, and you know, just being ex just super clear as well. Synthetic data is not we're not you're not making anything up. You're simply just allowing organizations to use that data to the fullest extent that that that data allows. And and that's that is such a key thing particularly in the field of healthcare which is such a datadriven field. and and and if that can speed up that process for for any small number of things that that those small number of things turn into big number of things and and that's really how change can occur. And so um so there's there seems to be a ton of benefits to to utilizing synthetic data and and utilizing it well. And so I'm I'm curious though, do you have any examples where uh synthetic data has has driven innovation whether that's in you know clinical research or predictive modeling operational improvements just just an example where uh synthetic data has has driven innovation within the medical field. Yeah, absolutely. So I'll say you know with with some of the clients that we're working with um we've used synthetic data in a number of ways. um whether it's look trying to predict what type of chronic kidney disease patients are going to progress to dialysis and what can we do what what are the real predictors of of um those types of patients you know going to dialysis. Dialysis as some people know it's it's it's a huge change in your quality of life. It's a huge cost burden on the health system and wouldn't it be nice to figure out you know why these patients are progressing so quickly how can we stop them how can we create a service to uh to change that and so you know predicting how chronic disease populations are going to you know go to their you know I guess worst outcome or you know the outcome that we're trying to avoid um and and being able to study that seamlessly and add data and remove data um if you have access to it um in a safe way. Um you know another example is is um you know we worked with uh another health system looking at cardiac patients. Can we predict which patients are going to uh lead to more severe heart disease or heart failure? And and we can actually use synthetic data to create machine learning models and it's it's been validated. is in the research and and whether you want to use synthetic data going forward or prove your case, right? Like a lot of a lot of companies out there, their struggle is they don't have data. And if you could use synthetic data to test out your tool, we can increase the speed of innovation, right, before you actually uh get into the original patient data. Well, and Dr. Dr. Jarani, I I love both of those examples because both both diialysis as well as the the cardiac issues. Um those both have dramatic impacts on on patient life on on their on their life and as well as they they have dramatic impacts on health care systems and organizations that are providing care to those patients. And and I mean if if if synthetic data can step in and provide insights and and direction in in providing better care for those individuals and and preventing those worser worser outcomes um then as a result everybody benefits both the health care systems as well as those patients and I and I think that's that's such a that is just such a great example of of why uh synthetic data can play such a keyle role not just in in uh those specific examples but just in healthcare across the board um is because it can make significant differences both for the patient as well as uh for the organization. And so um another area of focus uh that is is becoming more and more uh prevalent and and requires more and more attention are are gaps in in healthcare equity. And so, um, I'm curious, does can synthetic data play a role in in addressing these gaps, particularly when we're when you're looking at under reppresented patient populations? Can synthetic data step in and provide additional insights to allow your healthcare organizations and clinicians to to uh provide better care for for those groups of patients? Yeah, absolutely. Um, you know, you know, we talked a little bit about deidentification and and and that's the current standard of of removing things from a data set. You know, I think synthetic data allows you to to work with, you know, sensitive data and in a in a safe and you know much much more guarded way so that you can get those insights. You know, we worked uh with a a few clients looking at um looking at at at heat maps and and you know, are there geographic concerns around um you know, global warming or or you know, other types of you know, environmental issues and and which type of patients are living in these areas and are we seeing disparities in in chronic disease or or cancer um modeling? Um and so so whether it you know looking at large populations or even looking at your your your own health system and saying all right let's let's look at a common um surgical procedure like a hip fracture. Are all our patients going to the O at the same time? You know maybe we can put in ethnicity or language and figure out are are we slow because of a language barrier? are there some other types of disparities that we need to look into further with um you know with ethnicity or race. You're able to do that with synthetic data and and start off uh with a a great deal of insight and and with our with our platform what I love is we can ask a question, find an insight and with anything in healthcare, there are 10 more questions or 20 more questions. And so um you know having that self-service ability to a look at data, look at synthetic data and and ask more questions and really have this dialogue with data is really the power of our platform and the power of having access to to data. Well, and I yeah, I loved how you just you finished that last one on just the the synthetic data gets the ball rolling. It gets the initial questions asked. So then you can ask more and more questions and and really that's just that is the innovation process. That is the improvement process particularly in healthcare that that you're trying to that you're always trying to do. Um thank you Dr. uh Dingani for for joining us today. Um one last question before I let you go and I promise I will let you go after this question. I know I've peppered you with a lot of questions. Um but but for for individuals particularly your your clinicians and leadership who who want to champion synthetic data who want to champion data analytics and innovation in their organizations. What advice would you would you give to those individuals? I think we talked a little bit about you know what what are the ingredients of of cultivating that culture and you talked a little bit about leadership and access to tools. I would say the biggest thing is is data literacy, right? Like our clinicians know uh clinical workflows and then our IT and analytics team, they know like the data workflow. So developing that kind of common language of data literacy um teaching them and I would say out of all the industries healthcare has the lowest um data literacy by far and so you know why is that why is that so you know our our teams are in front of the EMR I'm sure if you've gone to the doctor's office your your doctor or nurse you know looking at the computer typing in all of these things about what you said and and your weight and your height and and and your medication and all of this data goes in there, right? And and we don't have a great way to to pull it out um unless unless you're working together in a concerted effort to break down silos of data or people and and then you need to have this data dictionary so that people understand what what they're using. Um so I think it starts with data literacy and I think the other part is is that curiosity right we are in the forefront of AI a lot is coming out there and there's a lot of skepticism but without without trying without learning and and and being curious you know we aren't going to move forward so I would say if you doubt it study it learn it and and figure out where the gaps are and be a part of the change uh because it's not perfect now it won't be perfect in the future but We need our health team members to be engaged with with where AI and and machine learning and and data is going. Well, Dr. Dunani, thank you for joining us today on the Clinicians and Leadership podcast. We're we're grateful for your insights and your experience and and just the the ways that you are um advancing and and empowering organizations to to use data to the fullest and and just the change that that is that is going to happen in healthcare because of of your work and and the organizations that you work with. And so um thank you for joining us today um and and we wish you the best. Yeah, thank you for having me and um yeah, we appreciate all that you're doing to talk about healthcare.

Synthetic Data in Healthcare: A Clinician’s Guide

Dr. Trushar Dungarani on Synthetic Data’s Role in Healthcare

Key Takeaways

Related Podcasts

Why Medical Transportation Is Healthcare's Next Strategic Frontier

Healing the Healers: How an AI Co-Pilot is Restoring the Human Touch in Medicine