COMPANY CONFIDENTIAL. DO NOT DISCLOSE.
Data-led innovation grows value in precision medicine
Siemens Healthineers uses a leading-edge statistical approach to increase value for its customers by streamlining production, transforming testing processes and delivering on a quality promise
_______________________________________________________________________________________
Challenge: Create efficiencies in troubleshooting to optimize development and grow the company’s portfolio of healthcare technologies.
Solution: Use JMP to process a large volume of data and implement statistically-driven experimentation, thereby reducing testing time and preempting potential issues.
Results: The embrace of statistical methods like design of experiments (DOE) has hastened development across Siemens Healthineers. In one case, engineers reduced testing parameters from 41 variables to 7, saving precious testing time and resources and identifying faulty components swiftly.
_______________________________________________________________________________________
Data analytics has revolutionized the healthcare field. And though research and development have always been the backbone of modern medical technology, as the industry has moved towards full digitization, the volume of data available has led to a digital revolution in healthcare processes. Analytics technology – and new statistical methods in engineering – have helped both the public and private sector to cope with the increasing demands of society’s healthcare needs. This trend is only set to grow in the near future as more and more organizations embrace the power of big data.
While analytics has significant potential to revolutionize daily healthcare operations and directly fight disease, to date, the biggest strides in healthcare research have been made in identifying efficiencies and reducing costs – vital goals in an increasingly competitive field burdened by strident regulatory benchmarks. One of data analytics’ most ardent early adopters, Siemens Healthcare Diagnostics, has used data to stay one step ahead of the game by anticipating needs and potential problems.
Drawing data from 5 million patients daily
Siemens Healthineers, a division of Siemens Healthcare Diagnostics, works with healthcare providers to increase value by expanding precision medicine, transforming care delivery, improving the patient experience and digitalizing healthcare services. The company’s broad portfolio encompasses leading-edge, high-quality medical technologies that solve some of clients’ most pressing challenges in hospital systems around the world. Moreover, Siemens products are estimated to touch nearly five million patients globally every day.
Data is at the center of the operation. Siemens Healthineers generates a vast volume of data and insights, derived from a base of over 600,000 digital systems. While swathes of the business are dedicated to developing new products, the reengineering of existing solutions is just as vital, and Siemens employs data teams that look at R&D entries to develop tests for swift and actionable troubleshooting. Siemens’ forward-looking culture of analytics makes it one of the most innovative in the world, and examples of data-driven improvements abound. Like many across the organization, Harald Ackermann, a manager at Siemens Healthcare Diagnostics, uses a statistical approach to engineering on a daily basis; one case in particular stands out as illustrative of the benefits to this strategy.
Large data sets, many variables, limited time to test
When the Siemens Healthineers production team noticed a critical inconsistency thwarting the scalability of production efforts, Ackermann’s team was called in to assist in identifying the root cause. “They had a production problem with a component they made for a specific diagnostic assay,” Ackermann recalls. There were 40 to 45 different components used in different sub-processes for this particular product and each component had between 25 to 30 different lots to be investigated. “We had a lot data: reaction data, source materials, and more that were looked at. As you can imagine in medical technology production, everything's fine and then you start testing a new reaction and it doesn't work anymore.”
Ackermann and his team set to work sorting through the data – when the volume of entry points is so high, and the sources so varied, formatting the data before analysis can be a mammoth job in itself. “In the beginning it was difficult because the data wasn't formatted well,” says Ackermann. “We had to get it into a certain format which took some time. But JMP helped me to really understand where the cause of the problem was.”
Ackermann had been using JMP software from SAS for more than 25 years over the course of his career with Siemens Healthcare Diagnostics. As such, he was ideally placed to help the team of engineers best utilize a range of advanced statistical methods made possible by the tool’s dynamic suite of applications for design of experiments (DOE), quality engineering and Six Sigma.
By Ackermann’s estimate, without DOE capability in JMP, it would have taken one person working full-time for more than a week to test each component necessary to root out the source of the production issue. “If you look at 41 components, it might take half a year to investigate it all,” he adds. “The fewer experiments you have to do, the more time you save.” By applying a design of experiments approach to manage a multivariate environment and identify an optimal design, Ackermann says the testing was completed in just three months.
Old JMP hand that he may be, Ackermann says that learning about new functions and utilities available in JMP 14 was something of a revelation. “While I was investigating these marketing data, I used the marketing module in JMP,” he says. “I also found another solution that actually led me very close to the root cause of the problem. For me it was completely surprising – I didn't expect it.”
Ultimately, Ackermann was able to trace the inconsistency to one raw material supplier whose quality issues were only amplified by the material’s incorporation into Siemens’ end technology. “Because of the analysis in JMP,” Ackerman says, “we reduced the amount of studies that needed to be done from the 41 that we initially investigated to only 7. We saved a lot of time.”
An expanding use of optimal DOE and quality engineering methods facilitates faster troubleshooting
It’s vital for engineers to be able to troubleshoot quickly and efficiently, as these solutions are usually being delivered to products already in the market. Ackermann says that by making the testing as targeted as possible with effective solutions like those he delivered, engineers bring fixes to existing customers as rapidly as possible. “Instead of looking at all the different possibilities that could have caused the problem, you want to make sure in the beginning that you can limit the number of tests to only those that are required, not simply everything that can be done," he explains.
This type of collaborative, data-driven approach is the backbone of what Ackermann and his team do: “If you run into a problem in production you always want to have a fast solution,” he explains. “You want to identify the cause very quickly…. So the quality control crew is responsible for fixing the problem, and they have my team as a support as they run the tests. We have to make the components, and we have to investigate the cause. The earlier we use JMP, the faster we might be able to really solve a problem.”
Ackermann explains that he uses these solutions, native to JMP, in reagents, cost, process problems and in products themselves. "The process can be easily transferred to any kind of issue we have in production, if it's related to a reagent cause,” he explains. “The categorical analysis system in JMP helped me to reduce the number of tests that were required.”
Ackermann is about to move to a new project within Siemens Healthineers, involving acids. “We'll be using JMP for pretty much all the data we produce. And we'll be using it to design experiments, like using urea to look at optimizing reagent components.” He says that most lab heads at Siemens Healthineers are statistically proficient and use JMP, but that lab technicians are not quite there yet – something he intends to challenge: “We always plan to do more, so that we can educate everyone in the organization to use JMP.” And in so doing, he explains, accelerate innovation cycles and move Siemens’ most ground-breaking new technologies from the workbench to the hands of medical professionals faster than ever before.
