In healthcare, “big data” has been a topic of discussion since electronic health records (EHRs) were first introduced, spurring questions like, how is data stored? What’s the best way to organize it? And perhaps most importantly, how can it be applied to improve patient care? Today, clear-cut answers remain elusive.
What is clear is that vast amount of information being collected can and should be put to good use because fundamentally, data can play a vital role in helping healthcare deliver the best care everywhere. Advanced analytics tools – specifically those that leverage machine learning (ML) and artificial intelligence (AI) – have the potential to demonstrate data’s utility in a way that does not further burden already-taxed care teams. These tools can also help address the problems of equitable care and access to care in resource limited areas.
Healthcare offers a unique use case for these technologies as data in this sector continues to grow exponentially. The human brain is limited in its ability to look at a given set of variables and understand how they interact, so there is a natural ceiling to the complexity of situations anyone – let alone a busy clinician – can successfully navigate.
In addition to the vast amount and different types of healthcare data being generated every day, the language of healthcare is largely unstandardized and unstructured, preventing widespread data sharing and restricting what can be understood across platforms and stakeholders.
Fortunately, advanced data analytics can evaluate a huge number of data inputs: lab results, white blood cell count, bilirubin, neutrophils, vital signs, medicine administration, concentrations and durations of meds, duration in hospital, demographics about the patients and demographics on a hospital—to name a few.
Machine learning can look at all these dimensions simultaneously mapping them to the answers needed by clinicians and other healthcare stakeholders, avoiding delays in treatment and improving outcomes for patients. In a clinical setting, the ability to extract these learnings can be lifesaving, providing an early warning or rapid diagnosis than would not be possible through manual processes.
The AI hype cycle would suggest applications for big data are seemingly endless yet, anytime multiple data sets, technology vendors, and human interventions are involved, there are bound to be challenges.
One of the greatest misses of healthcare IT, and the biggest hurdle for leveraging big data, is the lack of interoperability and compatibility among disparate systems, preventing it from delivering the most value it can for patients and providers. This becomes particularly apparent when health systems consolidate through mergers and acquisitions, creating siloed data that cannot be easily ported across providers or analyzed for a big picture view of the organization, or more importantly, for a seamless experience for patients.
Analytics tools sitting on top of EHR systems already help address this problem, pulling in multiple data sources and integrating them through an external system. Clinicians or administrators can then manipulate the data without altering anything in the original patient records. Applying AI to the data enables healthcare organizations to automate and scale the arduous process of mapping troves of fragmented data into something that makes sense and is cohesive.
For example, AI using natural language processing deciphers inconsistencies in EHR mappings for lab tests and then matches them to a commonly understood standard so resulting insights are actionable.
Another area where big data analytics is already delivering immense value is in infection prevention.
