Real-Time Data Analysis at the Point of Care in the ICU

Clinicians in the intensive care unit (ICU) often work with critically ill patients in a limited space, and need to make quick decisions at that point of care. The environment includes an increasing number of monitoring and life support devices that are capturing data, but not necessarily making it accessible or usable. Healthcare analytics technology is at the point now where it can be valuable in aggregating and streamlining data, even, in some cases, being programmed to be “smart” enough to help doctors and nurses increase the speed and quality of care. HfS recently heard of just this kind of activity under way—the development of “SmartICU” at a hospital in Spain, in partnership with NTT DATA and everis (an NTT DATA Group).

Virgen Del Rocio University Hospital is teaming its doctors and technologists with NTT DATA’s data scientists, to get better clinical results because the IT solutions are so easy to use that “IT becomes transparent.”

The clinical team works with NTT DATA to create use cases and build analytic solutions that use machine learning and real-time processing to create actionable recommendations in real-time. The doctors and hospital have an idea of predictions they want to make, and the NTT DATA team works with them to translate business requirements into analytic scenarios and algorithms.

At the outset of the one-year journey of developing the SmartICU solution^[1] with the hospital, the team identified data management and reporting as the fundamental obstacle. Data sources at the ICU go far beyond high-level device readings, when you consider unstructured information such as a nurse’s hourly measurements noted down manually. The teams thus spent a significant amount of time on data discovery, integration, reporting dashboard designing and customizing. Change management was crucial to ensure that all data is captured electronically going forward by clinicians and nurses. Doctors and nurses can also access all the data entered about a patient, at their fingertips, as needed, from any data capture point in the hospital, such as biomedical devices, the pharmacy, or labs.

With this foundation in place, the hospital and NTT DATA teams have turned their attention to the most value adding analytic scenarios to layer on top. For example, acute hypotensive episode (AHE) requires effective and prompt intervention or there can be a severe impact on the patient’s health. The hospital needs to be able to detect this condition before it happens. The NTT DATA team collected large amounts of data from patients and built algorithms to predict an AHE. The model will be integrated into the operational system to generate an alert when a patient needs intervention, before an emergency. 

While the technology is a key enabler, what is critical to driving the results is rooted in the collaboration between the clinicians, the hospital’s innovation group and IT staff, and NTT DATA’s data scientists and IT team. By identifying the medical outcomes to achieve and scenarios they want to address, the team can put the health IT and data to work and create a “smart” system that influences clinicians to impact medical results.