As the role of technology becomes ever more critical in modern healthcare, physicians need to know that the IT systems they use will be available whenever and wherever they need them.
By analyzing billions of real-time monitoring metrics at speed, Cerner can rapidly identify and address IT performance issues, and develop predictive models to warn of potential future problems.
Supportsthe delivery of exceptional healthcare with seamless IT service provisioning
Increasesresponsiveness by reducing the time it takes to diagnose performance issues
Fostersinnovation by building predictive models to detect problems before they arise
Business challenge story
Picking up distress signals
Technology continues to transform the healthcare industry, as providers invest in new ways to leverage patient data and help clinicians make better decisions at the point of care. But what happens if, at a critical moment, an important system—for example, an electronic medical records [EMR] system on a doctor’s iPad—becomes slow and unresponsive, or stops working completely?
Adam Splitter, Knowledge Architect within the Technology Improvement Department at Cerner Corporation (Cerner), responds: “As a provider of IT solutions to the healthcare industry, the consequences of an outage are always uppermost in our minds.
To minimize downtime, Cerner must constantly monitor operational data from its systems for signs of performance issues, and react quickly to fix any problems that it detects. With ever-larger streams of monitoring data being generated by servers, network components and end-user devices, the company needed to find a better way to make sense of it all.
“Every day, my team looks at around 500 different attributes collected from all our systems, adding up to two terabytes across our client base,” says Adam Splitter. “We continue to bring on new data sets weekly, so the load will only get larger with time.
“If an EMR solution fails, for example, then clinicians may need to go back to writing their patient notes with pen and paper. They waste time writing everything down, and then even more time inputting the data once the system comes back online. Worse still, they may find it more difficult to get the information they need about each patient’s condition, which can slow them down and prevent them from completing their patient visits in a timely way.”
“To help us react to performance issues more quickly, we needed to be able to recognize them sooner. The tools we were using at the time could not crunch through the data fast enough to have a real impact on our service delivery—so we began looking for a new approach.”
Diving in for deeper insights
To enable analysis of operational data at greater speed and depth, Cerner embraced stream computing technology from IBM.
“IBM® Streams was right for us because we wanted an enterprise solution,” comments Adam Splitter. “We usually work with open source tools, and we like to build things for ourselves, but in this case, we really wanted to shortcut the process and get results fast. The high quality of the documentation and support from IBM gave us a big head start on what might otherwise have been an 18-month project.
“Once we had settled on IBM Streams, we were able to plug in the statistical models developed by our data scientists and embark on a rapid proof of concept, which went very well. From there, we were able to industrialize the solution in just a few months.”
Cerner uses IBM Streams to pull data from its Apache Kafka data pipeline, and apply sophisticated algorithms to transform each new record in real time. Adam Splitter adds: “The straightforward integration between IBM Streams and Apache Kafka is a big advantage for us, and allows us to treat Streams as our data ingestion platform. Because IBM Streams already includes some very useful out-of-the-box connectors, we don’t need to reinvent the wheel and build our own interfaces.”
The company is currently rolling out an application for IBM Streams into production, based on two successful pilot programs.
“The statistical model we’re currently working with uses data from our RTMS [real-time monitoring system] daily timers to detect performance abnormality flags [PAFs],” states Adam Splitter. “Using IBM Streams, we can gather, process, benchmark, and analyze data from all our 1.2 billion RTMS daily timers and apply our PAF algorithm to gain much faster, more detailed insights into system performance.”
Enabling life-saving servicesBy honing in on performance issues at greater speed and with greater accuracy, Cerner can take action to minimize the impact of hardware and software problems on its clients’ services.
“Our new statistical techniques are helping us to provide greater service continuity with less effort,” says Adam Splitter. “We can now process PAF data in just ten minutes rather than 24 hours, a huge improvement. Since we now have more notice when performance starts to degrade, we can rapidly take measures such as adding resources or rolling out fixes to correct the issue.
“And considering we saw 25 fewer outages per quarter as soon as we introduced our PAF model, we are excited to see how much impact the new, much faster analysis will have on our uptime levels. Ideally, we’re hoping to get to a point where we can resolve any issues before our clients complain—or even before they notice!”
Moreover, the solution is unlocking greater innovation at Cerner, enabling the company to take a more agile approach to the development of analytical models.
“Previously, it could take us 18 months to develop, test, roll out and gain user adoption for a new monitoring solution,” recalls Adam Splitter. “With IBM Streams, we can eradicate a lot of that delay and release new solutions faster than ever before. As a result, we can be a lot more innovative and creative when coming up with new ideas, a key advantage in the competitive healthcare IT market.”
Looking to the future, Cerner has plans to introduce additional elements of automation into its service continuity processes, and gain more predictive insight into PAFs.
Adam Splitter concludes: “As we dive deeper into our operational data with IBM Streams, we want to move from looking at lagging indicators to leading indicators. We want to identify the patterns in the data that indicate that an outage or abnormality is going to happen in the near future. Ultimately, we would like to build a ‘dictionary’ of abnormalities that will help us predict when an outage is coming, and take action to prevent it before it happens.
“We’re excited about the future, and we’re pleased with what we have already achieved. IBM Streams is helping us to deliver ever greater levels of availability to our clients, with less manual intervention—allowing clinicians to get the most out of the technology at their disposal, and devote their full attention to the patients under their care.”
About Cerner Corporation
Cerner Corporation provides leading-edge health information technology solutions. It delivers the technology that connects people and systems at more than 18,000 facilities worldwide, and a range of services that support the clinical, financial and operations needs of organizations of every size. Headquartered in Missouri, US, Cerner employs more than 24,000 people globally and achieves annual revenues of USD 4.4 billion.
Take the next step
IBM Watson Data Platform offers a complete portfolio of data and analytics services providing unique and seamless product integrations to build apps faster and gain new insights easier with flexible deployment and pricing options. For more information about how IBM Watson Data Platform can help businesses solve tough big data problems rapidly and cost-effectively, please visit ibm.com/analytics/us/en/watson-data-platform/.
View more client stories or learn more about IBM Cloud.