Much of the discussion around EMRs and EHRs these days focuses on achieving broad, long-term goals such as improved population health. But here’s some data suggesting that these systems can serve a far more immediate purpose – finding inpatients at imminent risk of death.
A study appearing in The American Journal of Medicine details how researchers from Arizona-based Banner Health created an algorithm looking for key indicators suggesting that patients were in immediate danger of death. It was set up to send an alert when patients met at least two of four systemic inflammatory response syndrome criteria, plus at least one over 14 acute organ dysfunction parameters. The algorithm was applied in real time to 312,214 patients across 24 hospitals in the Banner system.
Researchers found that the alert was able to identify the majority of high-risk patients within 48 hours of their admission to a hospital, allowing clinical staff to deliver early and targeted medical interventions.
This is not the first study to suggest that clinical data analysis can have a significant impact on patients’ health status. Research from last year on clinical decision support tools appearing in Generating Evidence & Methods to Improve Patient Outcomes found that such tools can be beefed up to help providers prevent stroke in vulnerable patients.
In that study, researchers from Ohio State University created the Stroke Prevention in Healthcare Delivery Environments tool to pull together and display data relevant to cardiovascular health. The idea behind the tool was to help clinicians have more effective discussions with patients and help address risk factors such as smoking and weight.
They found that the tool, which was tested at two outpatient settings at Ohio State University’s Wexner Medical Center, garnered a “high” level of satisfaction from providers. Also, patient outcomes improved in some areas, such as diabetes status and body mass index.
Despite their potential, few tools are in place today to achieve such immediate benefits as identifying inpatients at high risk of death. Certainly, clinicians are deluged with alerts, such as the ever-present med interaction warnings, but alerts analyzing specific patients’ clinical picture aren’t common. However, they should be. While drug warnings might irritate physicians, I can’t see them ignoring an alert warning them that the patient might die.
And I can hardly imagine a better use of EMR data than leveraging it to predict adverse events among sick inpatients. After all, few hospitals would spend dozens or hundreds of millions of dollars to implement the system which creates a repository that simply mimics paper records.
In addition to preventing adverse events, real-time EMR data analytics will also support the movement to value-based care. If the system can predict which patients are likely to develop expensive complications, physicians can do a better job of preventing them. While clinicians, understandably, aren’t thrilled will being told how to deliver care, they are trained to respond to problems and solve them.
I’m hoping to read more about technologies that leverage EMR data to solve day-to-day care problems. This is a huge opportunity.