Secondary Logo

Journal Logo

Nursing Informatics

Big Data Fuels Unstoppable Change

Simpson, Roy L. DNP, RN, DPNAP, FAAN

Author Information
doi: 10.1097/NAQ.0000000000000373
  • Free

THE ADVENT of big data has the potential to be the biggest disrupter nursing has ever experienced. Leveraging data science, nursing can, for the first time, demonstrate and describe the profession's value, a requirement in a capitalistic society. However, if nurses themselves do not step up to determine the profession's value with measurable data, nonclinicians will make their own determinations about our value to care across the continuum.

How big is big data? It is huge; terabytes of information are involved. How big is a terabyte? To understand the immense size of big data, consider these traditional metrics: 1 terabyte equals 2 million books, 100 hours of high-resolution video or 1000 copies of the Encyclopedia Britannica.1

Never before has the computing power needed to reveal big data's hidden insights been cheaper or more readily available. Aggregating individual data sets and applying big data algorithms identifies nuances in subpopulations so rare that they are not likely to be apparent in small samples.

Chief nursing information officers (CNIOs) need to lead the significant use of big data, and they will need to leverage their doctoral education to harvest and analyze the data accurately. As fast as big data is moving, even a doctorate may not be enough to unlock the insights hidden in big data's massive datasets and relational databases.

To determine and portray nursing's value, CNIOs will be challenged to change the way they use big data today. For the most part, nursing explores descriptive data, which explains what is happening now. Descriptive data is the least valuable information type on the big data landscape. Instead of concentrating their efforts on low-value descriptive data, CNIOs need to refocus on high-value prescriptive data. Prescriptive data is defined as using analytics to find a solution for a particular issue.2 In the prescriptive data model, researchers use a laser focus to answer a specific question. Not only does prescriptive data balance humans' bias to believe their odds of success are better than they actually are, it directs researchers to next steps.

To pull the hidden information and insights from big data, CNIOs will need scientific researching skills. Being a successful data miner in the 21st century does not require a need to know it all—just the tenacity to know where to look to find it. One of the most crucial skills for nursing's big data scientists will be the ability to frame questions specifically in order to derive not just a high-value answer but the insights that go with it. Nursing's challenge is akin to the conundrum experienced in clinical drug development. Half of all clinical drug development trials fail. That means that no one died from the drugs, but no one got better either, and no one can explain why.

One of the least obvious and most important aspects of big data is the amount of data normalization that must take place to prepare it for the analysis phase. Most nurses are unaware of the time, expense, and expertise needed to populate a relational database with the information needed for big data analysis. At the Nell Hodgson Woodruff School of Nursing, Emory University, the data scientists spent 3 years smoothing and normalizing data from 20 000 patient records to build the database needed for big data research. This labor-intensive process, which is called “data cleansing,” is required to transcend the different data formats and computing platforms involved in a single electronic health record.

Big data lacks the packaged analytics toolsets solutions to “slice and dice” the data. This means nurse researchers must take the time needed to frame questions with specificity. If you don't ask the right question, you won't get a valuable answer. The lack of big data-strength toolsets creates opportunities for data scientists and nurse informaticians. These professionals are perfectly positioned to be the inventors and developers of the first toolsets designed by nursing for use by clinical data scientists.

Once data has been cleansed and the relational database populated, nurse scientists move to the first of 4 steps—selecting the demographic group and corresponding diagnosis that is of interest. The second step is to create a graphical representation of the relative frequencies of all diagnoses in the target patient group. The study group is then further refined by choosing a full profile of the medications, lengths of stay, orders, and procedures associated with the diagnosis—the third step. Selecting one of these factors and drilling down into that one factor in the group of interest is the last step before deciding which area of data warrants further pursuit.

If all of these steps seem too abstract, the value of the output is certainly clear. Consider the long-held belief that a patient with congestive heart failure (CHF) who could walk the unit was ready for discharge. At Emory, patients with CHF needed to complete an uninterrupted 6-minute walk to be discharged. However, big data proved that discharge readiness did not require the traditional 6-minute walk—a 60-ft walk was just as effective.3 Refinements in the standards by which medical milestones are measured have the potential to revolutionize the way care is delivered by replacing “one size fits all” thinking with a more person-specific approach. In essence, it is big data that has the most potential to take health care consumers to the promised land of personalized health care.

It is not just CNIOs who are being called to big data. Nurses at the bedside and those leading hospital units have a contribution to make to big data as well. For too long, nurses on the frontlines of patient care have depended on rote memory to practice to the standard of care. However, rote memory discounts emerging best practices, breakthrough research, and new clinical techniques. Today, nurses at the bedside are forced to wait 17 years for the latest research to reach them.

All the data nursing needs to speed research to the bedside is available now, which proves it is not the data that is the problem, it is the nurses. Instead of waiting for the research to make its way onto their radar, nurses who deliver patient care need to see that keeping current with the latest information, clinical strategies, and care techniques are part of their professional responsibility.

What can CNIOs and their counterparts who lead patient care in America do now to make nursing more cost-effective, more efficient, and safer for patients? I believe nurses at all levels need to jump in with both feet to the pool that is called big data. First, CNIOs need to be creating the toolsets we need to manipulate big data—who else is better qualified? Second, nurse executives need to advocate for nursing to be included in STEM (science, technology, engineering, and math) to access more funding to accelerate big data's exploration. Third, nurses leading patient care must step up and lead by example to compress the time it takes for research to be implemented at the bedside.

After all, 17 years is too long to wait!


1. Simpson R. Big Data: Nursing's Future to Practice. Viana McCown Lectureship. Paper presented at: the University of South Carolina, 2018.
2. DeZyre. Types of analytics: descriptive, predictive, prescriptive analytics. Accessed July 10, 2019.
3. Turner ON, Dean A, Simpson RL, Cranmer J, Smith A, Reilly CM. Implementing the Sixty Foot Walk Test: Knowledge, Beliefs and Attitudes of the Healthcare Team. Paper presented at: Emory University Nell Hodgson School of Nursing, DNP Project, April 2018.
© 2019 Wolters Kluwer Health, Inc. All rights reserved.