In health care, a focus on small data rather than big data can provide a better way to analyze potential causes of underperforming KPIs.
With data, quality is more important than quantity when it comes to delivering the best in patient care and business results.
Big data gets a lot of press nowadays and is often used as a proxy for data science technologies such as machine learning, artificial intelligence (AI), and natural language processing (NLP). Computers’ ever-increasing storage and processing speeds mean huge data sets can be crunched to reveal associations and trends that might have previously gone unnoticed. In many businesses, big data has been invaluable in predicting customer behavior and interests, but in health care, providing just the right data at the right time in the right context is what clinicians need to reduce costs and improve clinical outcomes.
And, medicine isn’t just any business.
Some health care professionals believe big data is one of the more overrated trends in health care. Vitaly Herasevich, M.D., Ph.D., associate professor of anesthesiology medicine at the Mayo Clinic in Rochester, Minnesota, shared his reasons in a recent article in Becker’s Healthcare:
Two principles are usually ignored. The first is that association is not causation. What works in retail and other industries using ‘big data’ is not working in medicine. Second is the idea of ‘pre-test probability.’ Data exists in EMR because clinicians suspect something and order tests and labs. Predictions on top of existing suspicions do not produce breakthrough results.
The most impactful analysis could well come from applications that are more specialized and focused on “small data.” These applications are more effective at showing potential causes of underperforming KPIs, identifying how process adjustments can impact profitability and patient-flow patterns, and for determining specific actions that would provide the greatest impact.
More Data Doesn’t Mean Better Data
Part of the health care industry’s recent efforts to shift to value-based care is digitizing the system, including health records. There’s evidence, however, that electronic health records (EHRs) fail to be as useful as intended.
“Big data isn’t always smart data,” wrote Jennifer Bresnick, editorial director of HealthITAnalytics.com:
Actionable insights are the key to using big data analytics effectively, yet they are as rare and elusive as a patient who always takes all her medications on time and never misses a physical.
On top of that, many software providers didn’t build their big-data analytics systems from scratch. Sometimes tools are slowly added to existing workflows with legacy software already in place, which can lead to confusion, interoperability problems, and oversights.
Small-Data Analytics Solutions
How can users of analytics tools get the best information to aid decision-making in real time? Part of cultivating smart small data is targeting the right data to begin with, and developing best practices for creating meaningful insights. This can mean tailoring the data to fit your particular purpose, including setting benchmarks, deciding on the right data to track, and establishing error-correction techniques (i.e., deciding how you want the system to handle errors so that there are no gaps in your data).
EHR systems that claim to offer data analytics may only provide useless, high-level metrics, little actionable-level detail, and use data that is neither reliable nor trustworthy. In contrast, our Performance Analytic Application provides an evidence-based management analytics suite that normalizes and enriches data from all key sources, including EHR systems, as well as satisfaction surveys, scheduling software, RCM vendors, and more. It embeds pre-determined best practices, requires no new infrastructure, and is accompanied by comprehensive customer support by business intelligence experts.
With our extensive expertise in Emergency Medicine and Hospital Medicine, we know what data is required to create actionable performance metrics, how to acquire that data, and present it in a user-friendly tool. Contact us for more information about how “small data” can work for you, or to schedule a 30-minute demo.