Big Data in Healthcare: Its Potential to Save Lives
What is big data in healthcare and why is it important? In healthcare, “big data” describes the use of large amounts of information to help guide operational and clinical decision-making. It uses methods like machine learning and predictive analytics to identify trends and patterns that would be difficult or impossible for healthcare teams to discern using traditional methods.
Just how big is the data? It depends on the use case. In some cases, organizations want to study nationwide or even global trends, while in others, they want to analyze trends within their own system or communities. These types of healthcare data can come from sources ranging from electronic health records (EHRs) to connected medical devices, medical imaging, genomics and beyond.1 Regardless of the scale and the source, big data can be a powerful tool to help healthcare teams anticipate potential needs and risks2 — and its potential is growing every day.
Learn how Baxter’s connected care solutions can support your digital health strategy >
Big Data vs. Small Data: What’s the Difference?
Big data is often classified based on the “three Vs:”3
Volume: Big data deals with large amounts of data — often measured in terabytes or even petabytes.
Velocity: Big data is generated in real or near-real time, and its applications often require equally quick analysis.
Variety: Big data is heterogenous, meaning it is drawn from a variety of sources, both structured (e.g., data that can be organized in spreadsheets and databases) and unstructured (e.g., data gleaned from images, audio files, etc., which can be more difficult to organize).
Real-World Use Cases of Big Data in Healthcare
Healthcare organizations are only beginning to tap into the value of big data, but they are already using it to help improve patient care and operational efficiency in many ways. For example:
|
Early Detection of Potential Patient Deterioration |
Patient deterioration can be difficult to identify, with symptoms that may be subtle and difficult to distinguish from other potential causes. Machine learning algorithms and predictive analytics have the power to synthesize data points like a patient’s vital signs, lab results, comorbidities, medications, demographics, clinical documentation notes and much more. The algorithm can then act as a sort of safety net — analyzing a patient’s data throughout their stay and alerting clinicians to potential signs of deterioration.4
|
Operational Efficiency |
Big data can also be leveraged to help improve operational efficiency. For example, healthcare systems can analyze large data sets to help forecast patient demand and allocate staffing appropriately. Big data can also help address wait times, bed management, equipment monitoring/usage and more.
|
Personalized Treatments |
Predictive analytics also have exciting possibilities for supporting personalized care. The concept of a digital twin, for example, is a virtual representation of a patient, pulling in the patient’s data from sources like wearable devices, lab results, imaging and more. These “patients-in-silico” can be used to help predict biological responses, disease progression and even treatment outcomes.5
Why Big Data Matters: Benefits for Patients and Providers
The benefits of big data can be felt throughout the healthcare journey — from the patient room to the surrounding community. As early as 2021, the World Health Organization highlighted many potential benefits of big data in healthcare, including:6
|
• Improved clinical decision-making through real-time analytics |
Whether healthcare analytics are used to help providers harness data on a global scale or gain a deeper understanding of trends within their own organizations, the potential is significant — and it’s growing every day.
Expert Insights: How Baxter Leaders See Big Data Shaping the Future of Healthcare
To help us explore big data’s potential in healthcare, we spoke with our own Ryan Hochworter, Head of Digital Solutions, and Chris Gutmann, Vice President, Enterprise Technology & Value Generation, who shared how healthcare organizations can harness data at a macro level to help support more personalized, data-driven care for each individual patient.
Q: What comes to mind when you think of “big data” in healthcare?
Hochworter: “Big data is different to everyone. How you look at it and what you can do with it depend on a number of factors, including the organization’s size and technological maturity. Some of our customers have large data science teams, while others may not have a formalized IT department.
One large IDN I’ve been working with for nine years has captured three billion sets of vital signs over that time; that’s a large data pool to analyze. A smaller regional hospital may be only in the hundreds of thousands over that same time period — but they can still find important insights if they know what they’re looking for and have a plan in place.”
Gutmann: “For a large health system talking about ‘big data,’ they are often thinking at a community level. For example, ‘in this suburb of a major city, there are a lot of retirees with orthopedic issues, so let’s look at acquiring more orthopedic practices here.’ That’s one type of big data healthcare organizations care about.
Another type is the data that is generated within the walls of the organization. That data comes in three forms: the data researchers see (which is the largest and noisiest data set), the data clinicians see (where they have the ability to layer in their expertise and judgement), and the data patients see (which tends to be the cleanest data, having gone through that clinical review). Like Ryan said, there is value to be found in all of these feeds — but your ability to tap into it may depend on the data science resources you have available.”
Q: How are you seeing hospitals harness big/macro-level data today?
Hochworter: “They’re using it to help solve for all sorts of things. Identifying signs of potential patient deterioration is big. Throughput is important. Equipment monitoring, too — will we have enough beds or infusion pumps? And with concepts like digital twins, the possibilities just keep growing.”
Gutmann: “At the patient level, organizations are using data to support better decisions that aren’t guided by each individual clinician’s habits or preferences. They’re identifying trends that show us what analysis we can take off clinicians’ hands, and introduce updated workflows that support new standards of care and best practices.”
Q: What advice would you offer healthcare teams to help make “big data” less intimidating and more empowering?
Hochworter: “If you have a data science team available to you, tell them clearly which data points you are looking for to support your research or initiative. Approaching them with a broad request like, ‘I need to solve for falls’ may be too vague. But if you can tell them which data points you need in order to make progress on the issue, then you can be off and running much easier.”
Gutmann: “Focus on the problem you’re trying to solve. Many organizations want all the data available to them from any given device, even if they don’t have a plan for what to do with it. You can certainly do that, but if you go into an implementation knowing what data you want to glean from the new technology and how you’re going to use it, that’s when you can harness the data points you need to help create a ‘big data’ pool that supports your goals.”
Take the Next Step in Your Digital Transformation
Big data is uncovering new possibilities in healthcare every day. Discover how we are powering the future of connected care, then reach out to your Baxter representative to learn more.
Featured Contributors
Ryan Hochworter, Head of Digital Solutions | Baxter
Ryan is a healthcare leader with over 20 years of experience driving digital transformation. At Baxter, he leads teams that develop technology-enabled strategies to address clinical complexity, healthcare value, and accessibility. Ryan has worked with...Read Full Bio
Chris Gutmann, Vice President, Integration Solutions of Connected Care Group | Baxter
With a background in engineering, Chris Gutmann has been in pursuit of improving processes and solving challenges. In 2008, Chris changed his career path to healthcare, working to see all the delivery pathways and understand where technology… Read Full Bio
Keep Exploring
Frequently Asked Questions
+ What is big data in healthcare and why is it important? |
| In healthcare, the concept of “big data” refers to the use of large data sets to help guide operational and clinical decision-making. Its importance lies in its ability to help identify trends and patterns that would be difficult or impossible for a human being to discern using traditional methods. |
+ How can predictive analytics help improve patient care? |
| Predictive analytics can help healthcare providers predict a patient’s likely health outcomes — for example, how they might respond to a certain treatment. These advanced analytics can also be used in operational capacities (e.g., helping forecast needs for supplies, equipment or staff). |
+ What types of data are used in healthcare analytics? |
| Big data healthcare analytics use both structured and unstructured data sources. Structured data is information that can be organized in spreadsheets and databases, while unstructured data is gleaned from images, audio files, etc., and can be more difficult to organize. |
+ What challenges or risks are associated with big data in healthcare? |
|
• Cybersecurity: When an organization collects large amounts of data, they must also ensure the data is protected. Data breaches and security vulnerabilities are chief concerns of big data management.2,7 • Interoperability between data sets: Big data comes from a variety of sources, ranging from EHRs to connected devices to wearable sensors. Discrepancies between data sets can make them difficult to analyze consistently. The use of common frameworks like Fast Healthcare Interoperability Resources (FHIR) can help.2 • Bias and discrimination: Big data algorithms, like the data sets feeding them, may exhibit bias and discrimination based on race, gender, socioeconomic status, etc. All data sources should be closely examined for potential biases, which can have significant and lasting negative impacts.7 |