Cerner Advances Big Data Analytic Capabilities
Big Data Cruncher
Health care can only advance as the technology that supports it continues to progress. Cerner, a global leader in health care technology, is at the forefront of advancing the industry’s data analytic capabilities.
"The goal is to provide the end user with a consistent and acceptable performance experience"
Cerner develops electronic health records (EHR) and additional health information technology (IT) solutions designed to collect a patient’s health history, including diagnoses, medications, lab results, allergies, radiology images and more, in one comprehensive digital record.
There’s a good chance that your health record, or a family member’s, is stored in a Cerner Millennium® EHR system. With millions of patient records and billions of patient data points housed in Cerner’s software, the company is continuously working to advance its data storage and analytic capabilities. To stay on the leading edge, Cerner needed to engage a multi-parallel processing database (MPP) that could handle the petabytes of operational measurement information that the company collects from provider sites around the world. To identify the solution that met Cerner’s needs, it tested its response time measurement system (RTMS) tool with six different MPP supplier databases. A set of 19 criteria were evaluated from each MPP supplier, with performance, cost and concurrency as the top three factors. Cerner ultimately selected Hewlett-Packard Enterprise (HPE) Vertica to power both its operational monitoring platform, as well as clinical big data analytics platform, HealtheIntent.
Dan Woicke, director of enterprise systems management with Cerner, explains the decision to select and implement HPE Vertica and the impact it will have on the operational front.
Dan represents Cerner on HPE Vertica’s customer advisory board, a group of customers including Facebook, AT&T, eBay and others that provide input and contribute to HP’s roadmap.
Enhanced and Advanced Health Care Analytics
Cerner collects Millennium software statistics and operational data, such as operating system and layered software workloads and performance statistics from Cerner health care provider clients around the world. We collect performance data on many actions that a clinician completes within the Cerner health solutions, which is used to survey and improve the performance and stability of our solutions.
For example, Cerner records how long it takes for our solutions to process a clinician’s medication or lab order. The goal is to provide the end user with a consistent and acceptable performance experience.
Not only do we need to store this vast amount of data, but we need to perform analytics across the set and identify issues in near real-time.
Three years ago, we selected HPE Vertica to house these massive amounts of static data that we needed to query and report on quickly.
Being able to analyze workflows, such as the time it takes to place an order, allows us to identify and proactively address situations that may prevent clinicians from completing actions in an acceptable time. Analysis exposes opportunities for improving workflow, system tuning and even architecture design.
To give perspective on the volume of data that I am referencing, Cerner collects 1 billion RTMS transactions per day across live client EHR systems around the world. RTMS transactions make up the largest data set, but it is only one of approximately 300 operational data parameters that Cerner collects. The totality of operational data that we collect in HPE Vertica equates to more than one terabyte per day, which is then stored for six months.
With HPE Vertica’s MPP solution, it only takes us a couple of seconds to query that full data set, a process that used to take a great deal of time and effort with a traditional relational database management system. Summarizing the data and making it available to our clients is now a completely automated process.
Benefits to Health Care Providers Across the Globe
With this advanced functionality, we were able to create an internal process where we proactively measure the performance of Cerner clients. When performance variance is identified by our statistical models, our team analyzes and identifies the nature of the variance. We can then create and apply a mitigation plan and audit process across all clients to ensure the variance never occurs again.
In addition, we have established algorithms to compute across the RTMS data to determine retrospectively if clients are experiencing issues, what we call Performance Abnormality Flag (PAF) at Cerner. This means that even if a client had an issue, but never logged it or called for help, the team is now able to identify the PAF and research the problem.
Each RTMS transaction that we log is like a breadcrumb – the physician placed an order, the physician pulled up the catalog to search for a specific medication, the physician placed an X-ray order into the scratch pad and then signed the order. By tracking each step, we have a holistic view of the path that physicians take through our solutions, which we call a transactional workflow. We can take those transactional workflows and compare them against other physicians who are taking more or less time to complete the same process. When we identify opportunities for improvement based on best transactional workflows, we can engage our consulting team to connect with the physician and provide additional training to help increase efficiency.
The functionality that HPE Vertica provides Cerner enables us to take our data and truly enhance our analytic processes to produce improvements and efficiencies that directly impact our clients around the world.