The 2019 Data Issue

By: Marc Iskowitz

An acquaintance of mine recently found himself in a fairly common plight: His car broke down miles from home, and so he called for a tow.

Tow truck drivers not only ply a noble trade; they also rank among life’s most sagacious observers. During the ride to the garage, this driver told my friend that most of his calls actually involve newer cars whose electrical systems have gone awry, rendering the vehicles inoperable. Given the complexity of today’s autos, the tow man shared, he spends his days bailing out one stranded new car owner after another.

We’ve come to rely on technology in many facets of our lives, from the computers in our cars to the AI-driven home assistants sitting atop our dressers. Doctors have begun to depend on AI algorithms to facilitate diagnostics and treatment. We tend to think that such tools have advanced far more than they really have; they’re actually prone to serious mistakes.

[Click to read the full article]

By: Larry Dobrow

Though you might not know it from the spotty coverage of such crises, pharma and healthcare organizations have endured an abundance of data privacy headaches in recent years.

Yes, Merck was publicly tsk-tsked after bad actors successfully pierced its defenses in June 2017. But how about health system UnityPoint Health, which was hit twice in 2018 alone, to the tune of the potential exposure of 1.4 million patient records? Even the supposedly impregnable Healthcare.gov found its records compromised, with the Centers for Medicare and Medicaid Services acknowledging an estimated 75,000 personal files were surreptitiously accessed this past October.

[Click to read the full article]

By: Alison Kanski

For pharma and healthcare organizations, the benefit of AI lies in its ability to quickly scan behavioral data for unknown patterns, then take action on them. Contrary to its reputation as the slowest of slow adopters, the industry is availing itself of the tech with a great degree of urgency.

There’s no limit to what the industry can do with the data it collects. For instance, pharma can exploit data streams being amassed by product teams. Data on the messages or messaging tactics that trigger a bump in prescriptions can be rejiggered to fuel future campaigns.

In the non-pharma world, the savviest brands and retailers have moved closer to the platonic ideal of 1-to-1 marketing driven by analytics and machine learning. Pharma, industry data execs suggest, is finally ready to use those same tools and tactics to reach doctors, patients, and caregivers more efficiently.

[Click to read the full article]

It's no secret that the healthcare world has awakened to the importance of customer data -- encompassing HCPs and patients. But a Veeva Systems survey of commercial operations execs found the data revolution may be further along than even its biggest advocates have suggested. Customer data has become a legitimate competitive differentiator. That's perhaps the central takeaway from the Veeva survey. 

[Click to view the infographic]

Tell us about your professional history.

I trained as a molecular biologist, but after completing my Ph.D. in molecular genetics at Ohio State, I migrated to computational science as a post-doc at Harvard Medical School – mostly as a way of doing experiments more quickly. I joined Novartis Institutes for BioMedical Research (NIBR) at its inception in a strictly computational capacity, and I have enjoyed growing along with the fast-paced and ever-changing science. I believe a data scientist in drug discovery benefits from having a base of knowledge in biology or chemistry fortified with statistical and computational skills.

Tell us about the “eureka” moment in your career, if there was one?

Around 2005, there was a powerful confluence of machine learning in cheminformatics and the availability of drug bioactivity databases with well-structured annotations. I realized we could apply machine learning to large-scale chemogenomics data to train models that predict targets solely from chemical structure. We published one of the first papers that addressed a critical need in phenotypic drug discovery: computational target prediction for compounds.

In the broadest possible sense, what do you do?

I manage a team of data scientists who characterize models of disease biology, identify therapeutic targets, and find pre-clinical drug candidates, all while growing our company’s knowledge base through dedicated data stewardship.

What is a typical day for you?

Coffee and a commute into Cambridge, Massachusetts. Then I meet with scientific teams and coordinate the application of data science to our drug discovery problems. This works out to be a mix of project team meetings, emails, more coffee chats to grease the wheels for my team, and — if I’m lucky — reading a scientific paper and a bit of coding, data wrangling, and analysis.­­ We have the mantra that we are the “brokers between data and the next experiment.”

How far along is pharma in its so-called “data revolution”?

Pharma is flush with data and always has been. The difference today is the expectation that experimental and clinical data should outlive those experiments or trials, along with critical metadata for mining and predictive modeling. Every pharma company is building out their AI capabilities and shifting toward the philosophy that data is an asset and that every experiment builds a deeper foundation of institutional knowledge. We have seen breakthroughs in specific drug discovery tasks when data and metadata are standardized and amenable to machine learning.

What are your hopes (realistic or otherwise) for AI’s potential in healthcare?

The healthcare field is a vast composite of activities, from preclinical drug discovery to clinical trials to manufacturing to commercial aspects. Every step of the process generates data and therefore is susceptible to disruption by advanced predictive machine learning methods. Wherever decision-making is boosted by predictive modeling, we will see techniques such as deep learning improving our accuracy and efficiency. However, our grand challenge in pharma is less about implementing AI and more about complex data and metadata management.

What’s something about the role people don’t fully appreciate?

Agility is paramount to scientific data management. Due to constant innovation and disruption, the lifespan for technologies may ultimately be shorter than the time it takes to design robust databases or software. The IT role in pharma requires rather unique flexibility to roll with the changes and a data engineering, rather than software engineering, bias.

How will the job and its responsibilities be different a few years from now?

There will be an awakening among all research companies as to the importance of data for long-term sustenance and survival. While singular discoveries from experiments or trials may not depend on data stewardship, sustained innovation absolutely requires reframing research as the continuum of all experiments, where the data generated is part of a greater ecosystem of knowledge. Data architects that have a wider lens and understand how any one type of data fits into the overall picture will be the most impactful.

What are your must-have work items?

Laptop, power, Wi-Fi, data, and a scientific problem. Curiosity and purpose I will bring with me.

Who is your biggest inspiration, professionally or otherwise?

Charles Darwin, Isaac Newton, Carl Sagan – any scientist that has shown the courage to shine a light on the dark corners of human ignorance in spite of cultural resistance to change.

[Click to read profiles on two other data architects]

Tell us about your professional history.

I trained in New York as a medical oncologist. After completing my fellowship, I moved to the University of Wisconsin Cancer Center as an assistant professor seeing patients, conducting clinical research, and teaching. It was then I realized my true passion was helping more patients than the ones I was seeing in my clinic, and that I could better do that in the field of clinical research and drug development. I joined AbbVie in 2009 and I have held numerous roles, including medical director, project lead, and most recently head of the development design center.

Tell us about the “eureka” moment in your career, if there was one?

There have been a few. One was during my time at the University of Wisconsin. There, I recognized that instead of working to help one patient, there was more I could do to bring new treatments to whole groups of patients by moving to the pharmaceutical industry. It takes a village to develop a drug, including everyone from statisticians to marketers, from discovery scientists to toxicologists, from safety physicians to regulatory colleagues. All play unique and important roles.  

In the broadest possible sense, what do you do?

I work with AbbVie clinical teams all over the world to help them design innovative and efficient clinical trials in order to get new and better treatments to patients as soon as possible.

What is a typical day for you?

I work with our clinical teams to add more innovation and health technologies, and to use novel data tools to drive toward the most efficient and successful clinical trials possible.

How far along is pharma in its so-called “data revolution”?

I believe the best is yet to come. My group is passionate about positioning AbbVie as a leader in digital health in all aspects of the patient journey, including disease prevention, disease awareness, diagnosis, and, of course, treatment.

What are your hopes (realistic or otherwise) for AI’s potential in healthcare?

There is so much potential for AI in our field. It seems like every week I think of another way we can use AI and, in particular, machine learning. From predicting efficacy or safety parameters to picking the best countries or sites to minimizing protocol amendments, almost everything we do can be improved by the use of advanced data analytics. It is super-exciting, but we must continue to practice a fair amount of scientific skepticism to ensure we understand the limitations of the data sources and that we are using the right data for the right purpose.

What’s something about the data architect role that people don’t know?

The data source is as important as the data itself. Back-tracing data to the source is the only way we can be assured we can trust the results.

How will the job and its responsibilities be different a few years from now?

As data sources get larger and larger, we need to make sure integration is less manual and as seamless as possible. We also need to make sure we have systems that can handle large data transfers to multiple individuals who are working on projects around the world. And we need to make sure we are capturing everything we do in a digital format that can be processed and analyzed to answer critical questions. If you plan in advance for the possibilities that come with machine learning/AI, then you can structure your process to collect relevant data points.

What are your must-have work items?

Data access, data accuracy, extensive amounts of data, and a passionate team dedicated to making a difference.

Who is your biggest inspiration, professionally or otherwise?

Early on in my career, I worked with a medical oncologist who had recently been appointed to lead our cancer center. She had this amazing ability to be all things at all times – a widely recognized expert in the field, an outstanding and dedicated educator, an astute administrator, an inquisitive clinical researcher, and, most importantly, a compassionate physician to her patients. Seeing her in action gave me hope that if I could just be one of these things, then I could bring value to the world and be a small part of creating better treatments for people living with cancer.

[Click to read profiles on two other data architects]

Tell us about your professional history.

After graduating from Marquette University, my first few jobs gave me unique visibility into different areas of the medical and pharmaceutical field. Positions included labs in the hospital setting, medical sales, and eventually in technical consulting mainly around business intelligence and data warehousing across various industries including finance, insurance, pharmaceuticals, and utility.

For the past two decades, I’ve dedicated myself to the pharmaceutical space, concentrating on implementing and engineering business intelligence, analytics, and enterprise solutions, utilizing the best practices and key insights I’ve gathered from the industry along the way.

Tell us about the “eureka” moment in your career, if there was one?

I’ve been fortunate to have several moments of major insights and inspiration throughout my career. However, one insight I learned early on – and confirmed throughout my career in various ways – is the realization that success in my roles is highly dependent on ability to communicate and educate, not just technology expertise.

I was once brought in to help navigate a project that was behind schedule with plenty of resources and experts. Aside from the technical perspective and experience I provided, the most impactful role I played to get the project moving in the right direction was the ability to communicate various highly technical aspects of the project in a way that all parties could comprehend. This required listening, understanding, facilitating, and communicating back in a relevant manner.

In the broadest possible sense, what do you do?

I am responsible for supporting and enabling our advanced analytics department across all of our services and solutions. My team and I work closely with the data scientists to understand their needs, both now and future. From there, we work on enabling and implementing our core solutions, while prototyping and investigating new methods and technologies. I also lead and support global initiatives to transform our company by having our data and analytics be a competitive advantage in the industry.

What is a typical day for you?

My days are far from typical. As we all know, data and analytics is a popular topic today, so there are more discussions around how to architect and engineer to meet current needs while considering the future growth of data and a variety of analytic workloads and demands. Therefore, aside from daily operations and architecture, many of my days are taken up by fostering strong relationships with customers and building strategies to advance capabilities that drive growth.

How far along is pharma in its so-called “data revolution”?

Pharma has been a laggard behind industries such as banking and retail from an adoption and capabilities standpoint, but I see that gap shrinking. As companies continue to realize the direct impacts of data on patients, they will be more willing to take strategic opportunities, such as leveraging real-world data to identify new indications and patient populations for drugs and extending product life cycles. I do not see pharma leading the charge in “data revolution” due to the nature of the industry, but I believe the realizations and impact that data can have are more quickly being realized.

What are your hopes (realistic or otherwise) for AI’s potential in healthcare?

The ultimate goal is for AI to have a positive effect on patients – making the possibilities endless. AI can enable better discovery of drugs and streamline drug development, as well as more effective patient experience programs. Time and money are always key factors in these areas, so the aspiration of leveraging AI technologies is to be able to drive costs down while facilitating quality drugs to production in the safest manner possible.

What’s something about the data architect role that people don’t know?

The glamour and glitz are in the insights that people see from the analytics, but data engineering and architecting is a key component to that. There is often the expectation that all data is consistent, high-quality, and trusted while optimally organized for any analytics. This requires architecting the user requirements, which is a major design consideration, but not the only one. The role requires the architect to have a balanced understanding of data, technologies, and methodologies to be used and how these methods and solutions work, and to anticipate additional use cases.

What’s something about the role people don’t fully appreciate?

As laws and regulations regarding the use and protection of data change, balancing the optimization for analytics while ensuring the data security and controls is – and will remain – an ongoing challenge that is both exciting and demanding.

How will the job and its responsibilities be different a few years from now?

The more technologies and companies mature in data and analytics, the more emphasis we’ll begin to put on developing balanced expertise in technology and analytics. I believe my role will have more demands and expectations around data management, security, data regulations, and understanding new analytic methodologies and the ability to educate others.

What are your must-have work items?

Internet access is an obvious must. But I still work regularly with a whiteboard and notebook. The whiteboard is more for visualizing and brainstorming, which I feel is even more important now than ever in my role. As much of my day is spent in meetings and building relationships, the notebook allows me to disconnect and focus on really understanding the person and their needs, while still being able to jot down thoughts and ideas in the moment.

Who is your biggest inspiration, professionally or otherwise?

Immediately my mind goes to icons such as Elon Musk, Richard Branson, and Albert Einstein. And although I have and still get many inspirations from these folks, I would say one of my biggest inspirations was an elementary school teacher who taught me that meeting expectations based on another person’s definition is not success. I had to define what success was for me and use that stand.

[Click to read profiles on two other data architects]