Desperately Seeking Data

The Quest to Fill Data Scientist Jobs in Healthcare

Data Scientist Jobs Decoding Big Data in Healthcare

Big data is everywhere. In fact, we’re all generating it every day. So much of what we do creates a digital trail including the goods we purchase, the mobile calls we make, the social media posts we write, the hotels we check into, and on and on. Every digital process is captured in a database somewhere, and the volume, variety and velocity (principles of big data coined by Gartner analyst Doug Laney in 2001) is ever increasing.

Big data is changing the way people and companies work and interact. It can provide valuable insights to inform decision-making processes and improve operational efficiency and effectiveness. But as its name suggests, big data can be a big mess.

"big data is changing the way people and companies work and interact"

In contrast to relational databases where there is a place for everything and everything in its place, big data is largely unstructured and stored in its raw form. This means, while it’s easier and cheaper to collect and store data, finding the data you need is analogous to finding a needle in a haystack. And so, querying and making sense of all this big data requires sophisticated tools and perhaps even more sophisticated data analysts.

Today, these data analysts are largely known as data scientists, part of a discipline born from statistics and raised alongside computer science. Members of this new breed of number crunchers have been steadily scooped up by companies like Google, LinkedIn and Facebook, making the data scientist what some have called the sexiest job of the 21st century. And while data scientists and related positions in business intelligence, analytics and informatics are now common in social media and retail businesses, their presence in healthcare is still somewhat nascent.

Big Data in Healthcare

Until recently, healthcare data was primarily collected for the purpose of clinical research. But the mandated transition to digital health records has caused an exponential increase in the amount of data being collected. Now that health systems are storing the details of every patient diagnosis, treatment and outcome, there are infinitely more possibilities for how to leverage the data. The industry has been slow to act on it though, because, let’s face it, how many Ph.D level data scientists like those found at data giants like Google or ivy league universities are kicking around your average hospital?

So why the increasing focus on big data? As mentioned above, healthcare organizations are now producing more data than ever before. The largest public health information exchange (HIE) in the US, Healthix, which serves New York City and Long Island, currently manages over 100 terabytes of data. Add this to other HIEs, large provider systems, insurers, and the scores of health data managed by government agencies such as CMS, and the volume becomes overwhelming. Not to mention healthcare’s growing contribution to the Internet of Things, a topic for another day.

Second, while early data scientists developed their own custom big data platforms and tools, market demand and technical advances have brought new tools to the table, making it easier to store and analyze these oceans of information, even for industries known as late adopters like healthcare. This means companies don’t need an army of doctoral level data scientists and can hire people with more varied educational and professional backgrounds.

Big Data in Healthcare

Third and perhaps most impactful is healthcare’s move from volume to value. Providers have been increasingly moved from fee-for-service that pays based largely on volume of treatments and procedures, to a pay-for-value system. The Medicare Access and CHIP Reauthorization Act of 2015 (MACRA) provisions will sunset fee-for-service under Medicare and instead tie payment to quality measures and outcomes data. Many private payers are following suit. For example, the nation’s largest health insurer, UnitedHealth Group, says it expects to increase provider payments linked to value-based arrangements to $65 billion by year-end 2018.

This means providers need tools to help them use all the data they collect to better manage population health. And so because of market demand from its participating provider organizations, Healthix recently launched an analytics service. “With a database with records of over 16 million patients, it just makes sense to leverage that asset to help our participants manage risk and succeed in shifting payment models,” explained Todd Rogow, Healthix Senior VP and Chief Information Officer. Other HIE’s such as Maine’s HealthInfoNet, and many large health systems and payers have taken similar steps.

The Market for Data Scientist Jobs

So what does this mean for those seeking big data roles in healthcare and for the employers recruiting for those positions? Why do we need big data, analytics, or informatics in healthcare, what does the job market look like, and what health IT skills are required to break in? We asked a few people both hiring and training individuals looking to take on big data roles in healthcare.

Adam Moody is Director of Data Operations at RowdMap, which provides a platform to assist providers and payers to ready themselves for value-based contracting. He said that as data is used more heavily to drive decisions around quality and patient outcomes in healthcare, roles like data scientists and analysts are critical to the work of his organization and others like it. “Our customers need help with the move from fee-for-service to pay-for-value,” he said. He explains that with data as the basis for most decisions and strategies related to this shift, “it’s no longer just about writing code, we need people that can understand the conversations and ask the right questions.”

“it’s no longer just about writing code, we need people that can understand the conversations and ask the right questions”

Eric Just, VP, Technology at Health Catalyst, a data warehousing, analytics and outcomes-improvement company, emphasized the need for a healthcare background as well. “We need people that understand clinical practice so they can tease out normal variation from trends,” he said. He added that some of his company’s clients are new to big data and analytics and that his staff must articulate to the clinician why they should use analytics and help build their competencies in this area. Because of this, he explained, the big data related jobs he’s hiring for almost always require healthcare experience. “Understanding the nuances driving our analytics is critical. You can’t just give them the numbers; you have to also help clinicians act on the information.” He said he believes it’s harder to acquire a depth of healthcare knowledge than to learn to build and deploy one of his company’s analytic models.

Moody agreed. “We look for folks that are passionate about healthcare and informatics and hire toward that passion. We can teach people to write lines of code, but we can’t teach true passion for healthcare improvement. Moody admitted that for some entry level jobs it’s possible to hire folks with no healthcare experience and teach them. But he felt that they need to come in with a strong interest and aptitude for learning about it.

Megan Landry, is the Program Manager for Health Informatics Programs at the University of New England (UNE) in Southern Maine. She recently helped launch a Master of Science Degree in Informatics at the university. The program focuses on both technical skills like data visualization, analytics, and database design as well as soft skills like project management and leadership training. She said all of the current students have some kind of healthcare background.

The program began taking students this past January and already the employers are calling. “They’re pleading,” she said, “for employees or interns with experience in healthcare data analytics or some variation.” She’s been contacted by health IT vendors, government, providers and insurers alike, all looking for the same thing. “Folks that can crunch data and use programs that also have critical thinking and communication skills,” she stated.

Data Scientist Jobs in Healthcare

Moody back at RowdMap, has an office in Maine and is one of the employers Landry spoke to. “These are difficult positions to fill,” he said. “We tend to build toward openings by understanding what the talent pool will look like in the future partly through relationships with local universities like UNE. We’re always recruiting.” He said that with offices in smaller markets such as Louisville, Kentucky and Portland, Maine, RowdMap has to be more active than companies in larger markets. “But we don’t see this as a disadvantage,” he said. “We end up with better candidates in the long run. In a small company they can be part of a project start to finish. We don’t have a lot of turnover.”

While Health Catalyst has over 400 employees and is located in Salt Lake City, Just reports that his company is also challenged in hiring. The health analytics market is getting crowded with new companies starting up all the time. In response, his company has turned internally to build organization-wide competencies and cross train employees. “Our data scientists are teaching our data architects and clinical team members the science behind the algorithms.” In this way he explained, his clinical teams can help often skeptical physicians understand what is behind the information their analytics tools are giving them.

Defining Big Data Roles and Skills

While the jobs are out there, Landry admitted that she struggles articulating to her students what the jobs are. “These jobs don’t tie up into a simple one-liner job title, and all companies use a different word to describe the same job. I think it’s confusing job seekers,” she said. She explained that while one company might advertise for a Data Analyst, another may be recruiting for an Associate in Informatics, both with the same job responsibilities.

And on the flip side, while she is seeing a lot of interest in big data related training and degree programs, she’s having the same issues communicating what her degree program is about as the employers are with jobs. “This field of study is so new that it’s difficult to put your finger on the pulse,” she said. “It’s not like nursing programs with the all same accreditation and curriculum regardless of where you go. A Master of Informatics degree is going to look different at each institution.”

And that might be OK, according to Moody, because the skills he’s looking for are broad and varied. “We are a smaller organization, so we don’t have overly specialized jobs. Our training is tailored to the individual, but relies heavily on hands-on and increasingly difficult assignments.” He said he finds that people with a breadth of skills do a lot better in big data type roles than those with a contained career path. “I find it less interesting when someone has a computer science background and moved from entry level to management as a java developer for example.” Rather, Moody is looking for someone with varied experience. “While we used to be tied to certain tools, we are now more platform and tool agnostic.”

Just points out that because the field is moving so quickly, formal technical training is less important than real world experience. “We need people that can take the analytics, determine what is clinically relevant, and walk through that with the clinical teams. “In a job interview, I’m not going to ask about training, I’m going to ask for stories about deploying and using analytics to change business operations and clinical practice.”

“we need people that can take the analytics, determine what is clinically relevant, and walk through that with the clinical teams“

Both Moody and Just acknowledge that technical experience is needed, but both argue that soft skills such as critical thinking and presentation style are equally, if not more important.

A former hospital director of informatics herself, Landry agrees. “While people may be great technological minds, they can lack the ability to communicate effectively.” So that is where she explained her program focuses most heavily. “You can teach someone to crunch numbers on the job. But if they don’t have the self-motivation and the social skills, they aren’t going to be very successful in this field.”

Breaking Into Big Data

So what do Landry, Moody and Just suggest someone do if they want to get into the big data field?

First off, all three suggest the person get acquainted with several database and visualization tools, and agree that SQL is the big daddy of big data. “If you can learn SQL, you can learn R, SAS, or Tableau, but it’s harder in the reverse,” advocated Moody. “When I look at a resume, I’m also looking for someone who has picked up and mastered multiple tools and platforms so that I know they can pick up the next tool that comes along.”

And along with a baseline technical background, all suggest those interested in the field get at least some healthcare exposure. “I had a student who worked in office management for a large electronic health record company,” described Landry. “He wanted to move over into the clinical side and so he got some training and worked weekends as an ER tech.”

Breaking Into Big Data

Just suggested students seek an internship at a healthcare system where they can work on a real project. “The book knowledge isn’t going to be terribly helpful. What will be is working in the field as a change agent.” He added that students should get involved with a real world task or project, even as an intern or volunteer.

Moody also said internships working on real projects can be very helpful and can give the student a sense of whether the field is for them. To help applicants decide if they should apply at RowdMap, they offer a link on their website where potential applicants can complete a few exercises to see if the kind of work they’d do at the company is something that interests them or they have an aptitude for.

Reading the Tea Leaves

Last we asked Landry, Moody and Just to tell us what was in store for big data in healthcare and what that would mean for future health information technology jobs.

Landry thinks interest in education in the field will continue to grow, and sees more informatics degree, certificate and continuing education programs on the horizon for her university and others. That’s good news for Just as he projects big data and analytics will become increasingly competitive. “We’ll need to hire the best and brightest to maintain our edge,” he admitted. But he also said he’s optimistic and feels that while there will be more employers competing for the same job candidates, there will be a corresponding increase in the number of people interested in and qualified for the field. “I see this as a rising tide to lift all boats,” he said. “We’ll be a better company for it, and we’ll have better candidates for it too.”

Moody said the future of big data will be getting better and better at answering the ‘so what’ question. “It’s not data for data’s sake. We can’t be the rear-view mirror evaluating strategies after the fact, we must look ahead to inform future discussions and strategies.”

“we’ll need to hire the best and brightest to maintain our edge“

So as healthcare data continues to accumulate, spurred by better technology, a larger and more varied data science workforce, and incentives linked to pay-for-value programs, big data will continue to be a big deal in healthcare. Big data related roles should be plentiful for those with some solid technical skills, a curious mind, and a passion for healthcare improvement.