A Digital Transformation at the City of Hope

December 19, 2019 | As Chief Digital Officer (CDO) of the City of Hope National Medical Center, Mark Hulse is responsible for determining the organization's digital transformation and strategy. While the role has many of the same components as the more traditional CIO role, Hulse says it really takes on cultural change aspects of the organization as well, interacting with patients and leveraging data and systems to create a more seamless digital experience.

The role is one in constant evolution, fueled by the healthcare industry as it moves more into the realm of data science, as well as undergoing a digital transformation.

"I really see my role as an amalgamation, in many ways, of the technology, the systems, the data," says Hulse. "But overall, this is really thinking about how we continue to improve the experience and connectedness of our clinicians, researchers and staff, but even more importantly for our patients?"

Thinking about how do to improve that experience by leveraging critical data and system assets is key to personalizing cancer care, Hulse argues.

On behalf of Bio-IT World, Mana Chandhok spoke with Hulse about his role at City of Hope, the changing point-of-care landscape, and future digital transformations for the industry.

Editor's Note: Chandhok, a conference producer at Cambridge Healthtech Institute, spoke with Hulse in preparation for the upcoming Bio-IT World West, part of the Molecular Medicine Tri-Conference, March 1-4 in San Francisco. Hulse will be speaking as part of the Digitization of Pharma R&D track. Their conversation has been edited for length and clarity.

Bio-IT World: Are there any other CDO positions we should be keeping an eye on as Pharma undergoes a digital transformation?

Mark Hulse: It may not be officially named yet, but leadership around data science is making its way into these roles. As cancer care in particular becomes both more personalized and complex, there is a rapid expansion in both the complexity and size of data. This makes the application of more traditional analytic methods more challenging. However, machine learning and AI can be better suited to develop predictive models and identify patterns within these complex data sets. One approach would be to use supervised learning methods to predict the outcome in a certain type of cancer, and therefore help guide the patient through treatment decisions. Another approach might use unsupervised learning methods to identify patterns in large, complex data sets. For example, the link between certain gene mutations and specific types of cancers has been well established. However, we are just beginning to scratch the surface of this from a research perspective. The application of AI could reveal patterns in this data that would focus and accelerate future research efforts. As you consider the growth of patient-generated data from wearable devices, in-home monitoring equipment, social determinants of health, etc.; the complexity and scale of data just continues to grow. So an emerging role might be Chief Data Science Officer or something along those lines, but really focused on taking it to the next level in terms of the development of machine learning and applied AI.

City of Hope was recently named one of the best cancer hospitals in the West by US News & World Reports 2019-2020. Do you think that this type of role is essential to success in healthcare?

This was a huge achievement for City of Hope and we're very proud of that. While I wouldn’t take any direct credit, I think overall what we are developing here will continue to expand City of Hope's recognition as a highly innovative cancer center, and one that's incredibly focused on accelerating cancer research. We think of this as delivering tomorrow's discoveries to the patients who need them today. As new data is generated at each step in the patient journey we can apply some of the methods we've talked about in terms data science to learn from this data. Then we can quickly develop the insights to help both clinicians and patients make the best decisions at the right time.

Can you go into a little bit more detail about what informatics tools or data strategies that you guys are using at the City of Hope to bring these kinds of actionable results to the patients?

A big focus that we have here is around precision medicine. In the world of cancer care it's often thought of as integrating clinical and genomic data to guide treatment decisions that are based on an individual patient's phenotypic and genotypic profile. And that's certainly true for City of Hope, but we're really looking to go beyond this in incorporating patient experience data. Often we think of this as patient's self reported outcomes, which are today more in the format of survey tools. But increasingly we are looking to incorporate data from streaming devices, wearable devices, in-home monitoring equipment, etc. It's really how do we continue to increase the level of connectedness that we have with our patients? And then again integrating and analyzing this data create what's been often called within healthcare now as real world evidence.

And then there is the final step of real world action. For clinicians there can be uncertainty about when to order a genomic test, and then how to interpret that test. There can be a lot of confusion when you have several different types of variations or mutations in the genes, and the evidence connected with those mutations is confusing. So we are looking to reduce some of that cognitive overload that our clinicians experience by leveraging the system to assist them in that level of decision making.

Another example we see in patients who are immuno-suppressed is sepsis, which is a huge issue. Now, there are a lot of academic centers that are developing predictive models for sepsis. But what we've found is that, in our bone marrow transplant patients who are severely immuno-suppressed, those predictive models really don't apply to these patient populations. So we've analyzed several hundred data elements and have come up with a predictive model that is now highly predictive of patients who might experience a septic event. So when you have that advance information, you can now begin to surface that automatically into the clinical workflow.

For example, if a patient is at low risk for sepsis, perhaps you just want to put some kind of a notification in the electronic medical record to alert the clinician and have them do some teaching with the patient around signs of infection. If it's a moderate risk, maybe you want to send an alert to the clinician to get the patient started on antibiotics. And if the patient's at high risk, the system, without even having to interact with the primary physician, can just go ahead and initiate an urgent infectious disease consult so those specialists can go see the patient right away. This is just an example, again, of taking real-time data and leveraging a well-validated predictive model and the system to take actions that can improve the patient’s outcome and simultaneously reduce a clinician’s cognitive overload.

You had mentioned just connecting more with the patient. Do you consider at-home monitoring, the wearables, etc., in the point-of-care category, or is that something that's separate?

I really see it as both from a clinical point-of-care and research standpoint. So if you're monitoring the patient, let's say they've got an in-home scale and they're weighing themselves every day, and all of a sudden you notice a sudden increase in the patient's weight. That's probably more fluid related, or it could be a part of perhaps the chemotherapy that they're taking or something along those lines. An algorithm can send an automated alert to clinical staff who could reach out to the patient and see if perhaps they need some adjustment to their medication. So that's a direct benefit from a clinical standpoint, but I also see it as being a huge benefit on the research side. As you're collecting and integrating all of this data, getting back to this concept around machine learning, you can use these AI methods to help identify patterns in the data that might predict the problem ahead of time.

Are there any digital transformations or tools that you're most excited about bringing to City of Hope in the future?

I think a lot of it really has to do with aggregating and analyzing data and information in order to learn and improve both the level of connection we have with our patients and to individualize and personalize their care. We think of this as every single patient being on their own clinical trial. If you're continuously monitoring this data and comparing it to patients who most closely match that patient, you can better predict a patient’s outcome and offer them choices that then meet their individualized needs. I think we are looking at three to five years in the future to mature the platforms that will provide us with the capabilities to rapidly accelerate discoveries, reduce clinical cognitive overload and deeply personalize each patient’s care.