Not many people can claim to be a chief data officer and a practicing neurosurgeon.
Dr Nicholas Marko heads the data team at Geisinger Health System and is director of neurological oncology at Geisinger Medical Center. He reveals to Communication Director how his dual roles complement each other and how the development of data science helps him in his work – and could help yours, as well.
Interview By Jan Wisniewski
How do you combine your roles as chief data officer and director of neurological oncology? Which came first?
My formal training is in clinical neurosurgery, but all through that time I had been doing data-related research. It started with molecular biology and then genomics – analysing large data sets is a part of that. That grew in scale to analysing whatever kind of data sets I could get my hands on and trying to integrate patient and clinical data with all the other molecular stuff. Therefore, as I was evolving as a neurosurgeon and going through my training, I was also going through my evolution as a data scientist in my research.
When I got to Geisinger, I was hired to do clinical neurosurgery and computational research in our research and development work. The latter half of that grew into a chief data officer role for our organisation. We have always been a data-rich organisation but like many organisations within recent history, we have started to realise that having a specific leadership role for data was necessary.
What are the parallels between your two areas of expertise?
There are operational parallels and there are conceptual parallels. I think operationally, clinical medicine and data science are converging. As doctors we are more ready to look to the information and analytics and at the same time the analytics are becoming more robust and capable of providing answers. It is is great to have a background in both because it really gives me a good perspective as to where the real value propositions are for the organisation and for the doctors.
Now conceptually, there are things about neurosurgery and the way that we look at the world that are helpful for me as a data scientist. With neurosurgery we are very used to dealing with risk under conditions of incomplete information. It is always challenging when you are managing a patient with a brain tumour as often there isn’t clearly one answer. There are a couple of different ways you could approach it and you have to make a complex decision after thinking about all the technical parts and all that the individual patient values and expects.
In data and in analytics, there is so much complexity that there isn’t always one right answer either. There is always risk that you might get certain things right and you might get certain things wrong. Conceptually being in a clinical field where we think about everything as lying on a distribution of risk has helped me in the data world because every decision we make about what to do with our information lies along a distribution of efficacy and risk as well. Just like I have to make decisions about what I have to do clinically, somebody has to make decisions about what we do on the data side or we will never get anything done.
You spoke about how these two disciplines are in some ways converging. How do you see the use of data science affecting health and treatment services?
That is the key question. I think the future is actually very promising as long as health care systems continue to adopt data-driven cultures. That is clearly the way many systems are going right now – a culture where you look to data for answers. You will see doctors using computer-based data modelling tools to help make decisions about what we should do for patients.
If you have got a patient in your clinic and you have six, eight or 10 chemotherapeutic agents that you can pick from, how do you do that? Well, you can do it based on your knowledge and experience of clinical trials. But imagine if you could sit down at a computer, input information about that patient and it could help give you some data-driven answers about what may or may not work for that individual. That is at the level of the individual patient. I think that will scale all the way up to the level of the health care system where you will start to see the same sort of data-driven approaches taken to help dictate how we do things operationally.
For example, we may see schedules in the operating room relying more on computer-generated scheduling modules than on a series of people sitting around with papers spread out on a desk trying to figure out which operations go in which room at which time.
Could you provide an overview of what data strategy is and what it means?
Data strategy is the process by which an organisation thinks about the value of its data and the application of the data and information across all operational arms of the business unit. It’s about treating data like any other limited resource that you have in your organisation. People are beginning to recognise the information that organisations collect through their daily operations has inherent value and that information can be leveraged to help anything from optimising operations to accessing new markets, to communicating better with clients and with employees.
It is about figuring out how you are going to: leverage your data, making secondary use of that data to continue to add business value and how to drive your operations in a more objective, data-driven fashion.
Would you say that all kinds of organisations need a data strategy?
Any organisation needs one that finds itself in a position where they are a) an organisation that generates a lot of information in the course of business or b) they are an organisation already saving their data but aren’t quite sure what to do with it yet. Small organisations with a tiny footprint or a highly-focused mission may not need to worry about a data strategy because the secondary use of data is not an important part of what they do. But anybody who is thinking about collecting their information and trying to get something out of it, or trying to leverage the information that they are gathering for something more than just its primary transactional purpose, will benefit from having a data strategy.
What should the relationship between a chief data officer and a chief communications officer be like?
I think it is one of the most exciting, evolving relationships in an organisation’s C-Suite. For a long time communications has collected much of its own data, has looked at data collected by other pieces of the system, and has largely been perceived as a different world than some of the other operational pieces of the organisation.
I think the opportunity to have a chief data offer and a chief communications officer working closely together will enable the enterprise to tap into the full spectrum of data information that is being collected across the system and strategically use it for their communications. At the same time, the data arm of the organisation can tap into the resources provided by communications, both internally and externally, to help get messages out about what the organisation is doing with its data. There is also an emerging field of customer relationship management, customer satisfaction, which is a much more data-rich field than some communications operations have dealt with in the past.
There is a lot of interesting and novel data-types that are coming into play.
So you foresee a mutually beneficial relationship for communciations and data people?
Communications teams can really derive benefit from tapping into their colleagues on the data side and helping to figure out how to strategically and technically use that data. It it may not be an area where communications people are incredibly familiar with but they are very clear about what they want to do with the data, and the data folks can help with getting your arms around the information and linking it with other pieces of information in the organisation that can help answer the operational questions of the communications people.
Traditionally, the relationship between data and communications has been inherently tangential. But we are getting into an era where communications are becoming much more data driven and at the same time the data people are realising the importantance of communications and change management. Two departments that have been somewhat independent of each other are really starting to realise the great opportunity to overlap and work together. It will be one of the most exciting points of convergence in the coming few years.
Nicholas Marko’s four steps to move to a data-centric culture
- High-level executive endorsement: get data on the radar of senior executives and make them understand why it’s important.
- Ensure internal messaging is consistent. Moving to a more objective and more computational, way of thinking takes a lot of consistent internal messaging.
- Ongoing change management is vital. Data management leaders are not always necessarily good change messengers. It is critically important because as the data fields change and update so quickly, change management concerning data is an ongoing process that never stops. Change leadership is the responsibility of chief data officers and chief communications officers alike.
- Decision makers have to get in the habit of asking for data. Whenever strategies are laid out, the people leading those efforts should make a habit of saying, “What information is this based on? What analyses did you do to back this up?” so that the people implementing the day-to-day operations of an organisation get in the habit of asking for data.