By Lisa A. Eramo, MA
For The Record
Vol. 29 No. 3 P. 22
Data scientists are charged with making sense of large amounts of data. Are HIM professionals candidates to fill this role?
Dubbed by Harvard Business Review as the "sexiest job of the 21st century," data science has only begun to make a dent in the health care industry. Organizations want—and need—highly skilled individuals who can take a deep dive into data and provide valuable insights to improve patient care and lower costs.
Just how much data are we talking about?
By 2020, the health care industry is expected to generate 25,000 petabytes of data (equivalent to 500 billion four-drawer filing cabinets), according to IDC Health Insights.
This digitization of health care data continues to revolutionize the field of data science, says Alvin Rajkomar, MD, an assistant professor in the division of hospital medicine at the University of California, San Francisco. Rajkomar recently wrote about health care data science for quality improvement and patient safety in an article published by the Agency for Healthcare Research and Quality.
"Health systems are generating incredible amounts of data about patients and health care operations. Previously, these data would sit in paper charts or storerooms and out of analysts' eager hands," Rajkomar says. "Now that the data are digital, they will not be used just to perform basic analysis—they will drive insights into the nature of health care delivery, open areas of inquiry to improve health, and power artificial intelligence algorithms."
Oddly enough, over the next decade the majority of data won't be derived from EHRs or originate in a traditional clinical setting. Instead, thanks to the Internet of Things, data will flow from wearable devices, biometric sensors, and other types of Bluetooth and wireless devices, says Susan H. Fenton, PhD, RHIA, FAHIMA, associate dean for academic affairs at the University of Texas School of Biomedical Informatics.
"What are you going to do with all of this data?" Fenton asks. "You have to figure out how to manage it, extract information and knowledge from it, and then put that back somehow into your delivery of care or processes."
Not surprisingly, experts say that HIM professionals may be among those who are most qualified to help health care organizations answer these and many other questions about the data they produce and collect.
What Is a Data Scientist?
Data science is an interdisciplinary field that draws on topics such as statistics, computer programming, and mathematics. However, successful data scientists must also possess a vast knowledge of the domain in which they work. In health care, this requires an understanding of not only how and by whom health information is generated but also whether that information is accessible and can be trusted in terms of accuracy.
Domain knowledge is literally one-half of the equation—and it's something that HIM professionals already possess, says Justin Starren, MD, PhD, FACMI, director of the Center for Data Science and Informatics at the Northwestern University Feinberg School of Medicine in Chicago. "Taking data and converting it into information and then into knowledge and insight is part of what HIM professionals do," he says. "Way back when, it was about chasing down pieces of paper and making sure they were organized so someone could look at them and gain insight. Now it's about bits and bytes."
HIM professionals already understand the context and lingo, says Mayank Gandhi, senior director of product management, information management, analytics at NTT Data, which provides information management and big data analytics services for health care organizations. "That gives them an advantage when transitioning into a data scientist role."
Consider Felicia Owens, MS, RHIA, a terminology mapping specialist at Intelligent Medical Objects (IMO). Owens, who has a bachelor's degree in HIM and a master's degree in health informatics, says it was the combination of her education and work experience that helped her move into data science. "A lot of times, employers will go after a physician or nurse for these positions," she says.
Prior to joining IMO, Owens worked as a clinical data analyst for a contract company specializing in HIT. Although she didn't have the formal title of data scientist, she performed many of the same job duties such as standardizing clinical data from various sources, writing SQL queries to validate and identify missing data, providing quality assurance of automated mapping, and more.
"Part of what most data scientists spend their time on is taking data that's dirty, ugly, and comes from a variety of systems and finding out how to clean, organize, and restructure it so you can actually do some analysis on it," Starren says.
These analyses usually yield big results, he notes. "The gain is usually in millions of dollars, whether it's through reducing readmissions, improving patient flow, reducing adverse events, or optimizing staff," Starren says.
Data scientists continue to be in high demand in a variety of industries, including health care. A recent job search for data scientist on LinkedIn.com yielded more than 8,000 jobs nationwide, with nearly 12% of these jobs in the health care industry.
Similarly, a recent search for health care data scientist on Indeed.com generated listings for more than 3,500 jobs nationwide, many of which require previous health care experience and knowledge of patient data.
The good news is that demand for data scientists continues to outweigh supply—at least in the health care arena. That's because few individuals possess the unique combination of computational expertise and health care domain knowledge, Starren says.
"You're almost always dealing with people who understand health care and need to beef up the computation or people who understand the computation and need to beef up their understanding of health care and where the data comes from," he says.
Those who possess both the technical and domain knowledge are highly coveted by health care organizations, insurers, pharmaceutical companies, the government, and others. "It's very much a seller's market," Starren says, adding that the salary for a health care data scientist often starts at or quickly climbs into the six-figure range. According to Salary.com, the median annual salary for a data scientist is approximately $92,000. Data scientists earn an average annual salary of $113,436, according to Glassdoor.com.
Are They Worth It?
This begs the question of whether health care organizations can actually afford to hire data scientists.
"In many cases, folks are hiring those who are trained in the advanced computation and trying to bring them up to speed on the health care, or they're trying to bring up the skills of their reporting staff to do more advanced analytics," Starren says. "Just going out and saying, 'I want to hire a bunch of biomedical data scientists will break the budget."
Adam Nelson, vice president of business consulting and offering development at NTT DATA, agrees. "When you look at midsized or smaller hospitals, they just don't have the budget," he says. Instead, these organizations often outsource their data analytics projects to companies that can hire data scientists, modelers, and developers internally. Large health plans and health systems (as well as systems that include an integrated health plan) may be more equipped financially to hire data scientists, he adds.
However, this may change over time as organizations of all sizes realize the power and insights inherent in health care data, Gandhi says. "You have partnerships like accountable care organizations and patient-centered medical homes where data are the key to achieving the common goal of increasing quality of care while reducing the spend or making the spend more efficient," he says. "More organizations are identifying that they need [data scientists] so they can make decisions in a more involved and methodical manner."
In the future, health care organizations may even integrate data scientists more directly at the point of care, enabling physicians to access a centralized database of updated and insightful information prior to each patient's visit. However, experts agree that centralizing patient data in this manner is probably not something that the industry will see in the short term, though it is something to anticipate going forward.
Gaining Data Science Experience
For now, there are many other ways in which HIM professionals can join the data scientist ranks. For example, with proper education and training, some HIM professionals who transition to health care data scientist roles work for vendors or outside parties that collect and analyze data for hospitals, Gandhi says.
Although there is no single formula for success in terms of becoming a data scientist, HIM professionals may find that obtaining a master's degree and/or landing a position as an analyst, data integrity specialist, or clinical informaticist is a good first step.
That's what Kiaja Earlywine, MS, RHIA, a clinical informaticist at Community Memorial Hospital in Cloquet, Minnesota, is doing. Earlywine, who obtained her master's degree in health informatics from the College of St. Scholastica, hopes to eventually obtain AHIMA's Certified Health Data Analyst (CHDA) credential to further her fondness of working with data. "I enjoy turning data into information that can be used to make changes and improvements within health care," she says.
A colleague of Earlywine's, Rachel Hendrickson, RHIA—who will graduate in May from the College of St. Scholastica's master's program in health informatics—plans to apply her current experience working as a clinical informaticist to more advanced analytics roles within health care. "I believe anywhere data are captured, there is a potential for data analytics to provide benefits, and it would be exciting to see data analytics more widely utilized," she says.
After obtaining her master's degree in HIM from the College of St. Scholastica, Marketa M. Bumpus, MS, landed a position as a quality data integrity specialist at Reid Health in Richmond, Indiana, where she currently serves as subject matter expert for all data collection and abstracting. Like Earlywine, Bumpus plans to obtain the CHDA credential to be better prepared for an anticipated promotion to quality data coordinator at Reid Health.
Enhancing Your Data Science Knowledge
One of the challenges HIM professionals face when moving into data scientist roles is that they're often competing with those who have degrees in computer science, information science, biology, or even nursing.
Higher education and/or additional certification may be necessary to stand out in the crowd and justify one's ability to work as a true data scientist, Fenton says. This may require going back to school to obtain a master's degree in data science, health informatics, business analytics, or biomedical informatics, all of which tend to focus on the skills that HIM professionals lack (ie, data mining, data visualization, data structures, and developing algorithms), she adds.
Online programs may be particularly helpful for those who are already working in health care because they're often more flexible in terms of scheduling and may even offer a self-paced or competency-based curriculum.
For example, Northwestern University offers online programs to obtain a master of science degree in either health informatics or predictive analytics. The University of Texas Health Science Center at the Houston School of Biomedical Informatics also offers various online options, including a graduate certificate in biomedical informatics or public health informatics as well as master of science and doctor of philosophy degrees in biomedical informatics.
Fenton says the students enrolled in informatics programs at the University of Texas are equally divided between those with a health care background (eg, clinical or HIM) and those with a background in computer science or information systems. The average student age is 38, and most already possess at least one higher degree and a significant amount of work experience.
Gandhi advises prospective data scientists take a few free or low-cost courses online before enrolling in a formal education program. This will help determine whether data science is a legitimate interest and provide an idea of skill level, he adds.
Next, identify a data science educational program that caters specifically to health care. This may be easier said than done because many programs are domain agnostic, meaning they're designed to educate professionals working in a variety of fields, Rajkomar says.
Consider the following questions when evaluating master's degree programs to determine whether they are suited to launch a career specifically in health care data science:
• Does the program description reference health or biomedical data?
• Does the program include electives that focus on health care data?
• Do any of the professors have a health care background?
• Does the program include a capstone project or practicum in a health care setting?
• Where do graduates typically end up working? Is it in the health care industry?
"You need to be looking around and observing what's happening in our world with the data and plan appropriately," Fenton says, adding that even if HIM professionals don't intend to become data scientists, they must upgrade their skills to stay relevant and competitive. "Take a good hard look at your skills … and try to figure out your own path going forward."
— Lisa A. Eramo, MA, is a freelance writer and editor in Cranston, Rhode Island, who specializes in HIM, medical coding, and health care regulatory topics.