By Joe Stanganelli 

July 26, 2016 | CAMBRIDGE, Mass.—Want to succeed in the healthcare or life-sciences sector as a top-tier, influential executive in the world of big data and analytics?  Top experts convening at the Massachusetts Institute of Technology this month advised that, among other things, the key lies in fostering a collaborative corporate culture and working well with others.

At this year's 10^th annual MIT Chief Data Officer and Information Quality Symposium (MIT CDOIQ), the heart of almost all discussion and the vast majority of the panels and other sessions focused on the past, current, and evolving role of the CDO.  Organizations are increasingly viewing the CDO—once a vanity C-suite title that didn't really fit in with other C-suiters in terms of power or prestige—as a vital connector facilitating the flow of insight among all members of the organization, and thereby empowering the role.  What's more, if some of the best-attended sessions over the course of the three-day Symposium are of any indication, nowhere is the CDO role having more of an impact today than in the industries of healthcare and the life sciences.

"[Life sciences need] a Chief Data Officer [in order] to start using data as a strategic asset," Mark Ramsey, CDO of GlaxoSmithKline, told an overflowing lecture hall full of Symposium attendees at the outset of a panel discussion titled "Dynamics between the CDO and the CIO: The Case of a Major Pharmaceutical Company" (GlaxoSmithKline being that "major pharmaceutical company" in the case of this session).  "Within a pharmaceutical company it's a little different; [whereas] marketing organizations … are usually the ones who are passionate about data, within a pharmaceutical company, R&D is the largest part of the organization.  That's really the heart of how to bring new products to bear, and it's … not talking about 'When was the last time someone purchased something?'"

Instead, posited Ramsey, data—and the CDO's role in facilitating data in a pharmaceutical company or other life-science organization—"is about trying to support the scientist to make better decisions."

Helping Those Who Don't Want to Be Helped

Ramsey's notions on the support of CDOs and other data scientists in medical research apply equally to doctors and other healthcare professionals, noted other MIT CDOIQ speakers in agreement.  Unfortunately, the healthcare and medical research fields have only recently begun to yield in their long-storied resistance to data officers' insights and suggestions inspired by analytics and algorithms.

"I go back to the early '90s and the [nursing] work that I did with [surgeon] William Shoemaker at USC [Medical Center], where we were using non-invasive biomedicine … and with that [resulting] data we were applying predictive algorithms against it," answered Charles Boicey, Chief Innovation Officer of healthcare analytics firm Clearsense, when asked in a CDOIQ panel about making healthcare's technological and cultural organizational move from merely descriptive analytics to more powerful predictive and prescriptive analytics.  "We actually learned [about] predicting outcomes for trauma patients[.]  In 1992, 93, that was not well looked at, and we made the very big mistake … of calling that prescriptive analytics, so basically we were going to tell our [colleagues] what they should be doing.  It went over, as you can imagine, very poorly."

The panel in which Boicey participated on the first day of the Symposium—appropriately titled, "The Evolving Role of the CDO in Healthcare—Challenges and Opportunities"—continued to poke and prod at this idea of people—and especially doctors—not liking being told what to do.  Panel moderator Nicholas Marko, a neurosurgeon with CDO experience who also serves on Clearsense's advisory board, poked fun at his practitioner colleagues' sense of self-importance by relating the old joke about the difference between God and surgeons (punchline: "God doesn't think he's a surgeon").

Now, with nearly 25 years of experience of giving data-driven yet unsolicited advice under his belt (including work at UC Irvine Health and other healthcare centers), Boicey believes that he has worked out the best way to present predictive and persuasive analytical information to healthcare providers for optimal effect.

"20 years later, I find myself in a very similar… position, where we're now going from—again—descriptive into something predictive.  So what we did at UCI … is more pushing out of information and also saying, 'Look, you've got a patient that in the last 30 minutes the heart rate has increased, the… blood pressure has dropped… and here's some information—and, oh, by the way, X number of patients who have presented similar to this, this is what their outcome was,'" Boicey related to his audience.  "So we're not even suggesting anything at this point; we're just putting information out—and we have found [healthcare providers] very receptive on the clinical side of things because [of] how fast things are moving… By providing value in that manner… we are, I think, showing value and not assuming a position of 'We're here to tell you what to do.'"

The Role of Control

MIT CDOIQ Symposium speakers agreed that the issue of control is a sensitive subject when it comes to data—especially when it comes to healthcare and life science.  Compliance and geocentrically informed data-privacy issues aside, a big part of the problem is the perpetuation of traditional constructs of data ownership.

"I think historically people were very protective of their data in healthcare," Boicey's fellow panelist James Noga, Vice President and CIO of Partners HealthCare, told the panel participants and audience.  "I don't know about your system, but I can remember having discussions with a cardiac surgeon, 'I'm not gonna give you my data because that data is of value to me.'"

Still, Noga is optimistic about current trends in the data culture in healthcare and life science.

"[Today,] I think those barriers have broken down and people are more willing within an organization to share data.  They've really provided that basis for being able to move into what we call predictive analysis," said Noga.  "We're seeing new datasets with patient-reported outcomes[, and] I think the diagnostic realm is where people are really pushing for advanced [predictive] analytics."

Ramsey, speaking at his own panel session the next day, demonstrated a similar yet very pragmatic outlook regarding how the CDO relates to these notions of data ownership, data control, and data responsibility.

"The business owns the data, and we [as CDOs] are custodians as it passes through; that ownership comes with making sure the [data] quality is at a certain level," said Ramsey when he was asked by an audience member about the division of data-related labor between the CDO, the CIO, and others in the organization.  Ramsey went on to emphasize that CDOs fill the roles of conduits, analysts, and facilitators of data (and all of the functionalities that stem from data).

"We're the catalysts [as CDOs]," argued Ramsey.  "It's the subject-matter experts who are the ones who are harmonizing [data]."

As such, urged Ramsey, CDOs must put "a lot of pain back on the business owners," i.e., the CEO, CIO, and CFO who are more directly involved in general business functions, when it comes to ensuring data quality "because they've gotten away with doing some things in the past that aren't going to be acceptable as we [move forward]."

"It's going back to the folks who created the data.  It's, 'Wait a minute; if I just do it right, I don't get all these questions, I don't get the overhead,'" Ramsey went on to elaborate.  Only then, he said, can he then in his role as GlaxoSmithKline's CDO get into what he calls his "key focus"—leveraging data for "exploratory uses, driving out value, driving out complete transformations of R&D, [and] ramping up [a] new information platform[.]"

Meanwhile, Ramsey's co-panelist and company counterpart, GlaxoSmithKline CIO Daniel LeBeau, voiced a more hedged—albeit seemingly contrary—opinion.

"I still have a difficulty with the concept of data owner[ship] after 20 years," said LeBeau.  "Do you buy data?  Do you sell data?"

Perhaps, then, LeBeau's CIO-ish attitude buttresses Ramsey's arguments, providing the reason CDOs like Ramsey exist.  If no one "owns" or otherwise controls data in an organization, how can its quality or its value be assured?

On the other hand, LeBeau's fluffier, socialistic viewpoint of data ownership seems to better support the breaking down of silos than Ramsey's stricter, less forgiving, and more practical viewpoint.  Indeed, LeBeau's anti-ownership ideas of data have become more popular than ever in the healthcare and life-sciences realm, as genomicists and other medical researchers wring their hands over data privacy—accusing data protectionists of stifling scientific progress and even harming children.

How, then, can practical and effective CDO-ing foster collaboration and combat data silos?

CDOs Bulldoze Silos

Data hoarding and information silos have also long been status quo—cultural walls that healthcare and life-science CDOs have but barely started to tear down.

"[I]n siloing… your whole data enterprise is decentralized over everywhere, and nobody really knows what you're doing with [data] except everybody knows that primarily it's theirs," said Marko.  "That makes it very difficult to get any kind of multidisciplinary work done [because] the data[-related] work… particularly [in life sciences and] healthcare is really a constantly multidisciplinary effort, and… running and managing a multidisciplinary team is not the same thing as hosting a meeting for 43 people."

For Marko, therefore, the solution for the healthcare or life-science CDO is to create and maintain a collaborative culture that shares and leverages organizational data so as to foster efforts that are truly multidisciplinary.  Referring to these concepts in concert as "data strategy," Marko emphasized to attendees that this data strategy is one of the core jobs specific to the healthcare or life-science CDO.

And one of the core tenets of Marko's data strategy is that technology is a red herring.

"So many places say we want to do good stuff with our information [because] we know there's a lot of potential value in there, so what we're going to do is—centrally—start throwing modern technology and training people at it and let them tell us how to find the value," said Marko.  "[It's] not that some of that isn't sometimes workable, but if you don't have a vision around it and you don't have some sort of general strategy to how you're going to put these resources in these places and which parts you're going to build up and which parts you're not, what you end up with is a series of one-off solutions to specific problems that work for the one thing they work for, that don't really fit together very well, and don't make it any easier to build the next one once you've mastered the last one, and—often times—a lot of random people with random skillsets scattered throughout the whole place doing bits and pieces of everything but not really owning or working in concert with anybody very well."

Ramsey, for his part, echoed Marko's sentiments.

"[You] keep people involved [with the data] because you always have to have people, but what you don't want is a bunch of people sitting in a room with Excel spreadsheets arguing over data attributes," said Ramsey.  "When you're curating data and rationalizing it, you can spend the rest of your life trying to do that."

Instead, according to Marko and his panel, the data strategy of a healthcare or life sciences CDO is a culture strategy.

"I think that's what data strategy does," said Marko.  "It synchs your organization's vision [and] what you're doing with your data large scale, [and] it helps to shape your culture around what it means to have a multidisciplinary team working on things."

Indeed, while Marko arguably stole the show during his panel session, doing by far more talking than the rest of his panelists combined, it was his third panelist—Eugene Kolker, Vice President and CDO of Global Technology Services for IBM—who perhaps summed up the CDO challenge best and most succinctly.

"Data is not a technology issue and it's not a statistics issue," offered Kolker.  "It's a people issue."