COMPANY CONFIDENTIAL. DO NOT DISCLOSE.
A smarter factory speeds production of breakthrough diabetes monitoring technology
Technologists at Abbott Diabetes Care have developed an integrated operations dashboard that gets to the root of production improvement
______________________________________________________________________________
Challenge: With demand for a smartphone-enabled diabetes monitoring device outpacing production, technologists at Abbott Diabetes Care sought faster, more accurate ways to implement production improvements and ultimately improve yield.
Solution: Create a direct link between Abbott’s manufacturing database and JMP, bypassing third-party software to obtain raw data and immediately begin manipulation and analysis. Technologists reduced repeat analyses using custom JSL scripts and designed a top-level user interface that enables users to extract, manipulate and present meaningful data.
Results: Ultimately, this solution crystallized into a multi-faceted operations dashboard that allowed for quick, real-time analysis by users with a range of expertise. This innovation, made possible by JMP, has reduced analysis time from hours to minutes.
______________________________________________________________________________
Today, more than 425 million people worldwide are living with diabetes. Health providers have long understood that the key to managing the disease lies in consistent and accurate blood glucose monitoring. Yet, until recently, the only available way for diabetes sufferers to accomplish this was to draw blood from a finger prick between four and ten times a day. The intrusive, often painful nature of this method not only interferes with patients’ quality of life, it could also cause inaccurate or incomplete readings.
All that changed in 2014. That’s when Abbott Diabetes Care announced it had developed the breakthrough FreeStyle Libre, a wearable glucose monitoring device that can be scanned with a smartphone. Not only is the device able to take data continuously – leading to a more complete and accurate picture of blood sugar fluctuations – it enables patients to discretely and painlessly check their glucose levels at any time.
Though the device has been released in only a few countries to date, the widespread success of FreeStyle Libre technology has quickly resulted in soaring demand among diabetes sufferers around the world. As a result, Abbott’s 700-person production facility in the United Kingdom – the exclusive manufacturing site for FreeStyle Libre – now faces the formidable challenge of scaling up their operations to meet the need.
Finding improvement opportunities hiding in the data
“So many people want the device,” says Abbott technologist Giacomo Cascone. “But unfortunately we can only make a finite number at a time.” Cascone, together with fellow technologist Tom Bennett, is now spearheading Abbott’s efforts to optimize production as part of a technical services group that uses analytics to identify improvement opportunities and ultimately produce more devices.
Bennett, a biomedical engineer by training, and Cascone, a chemist, recall they were initially tasked only with an open challenge – increase yield, fast – and a software tool neither one knew how to use. That tool was JMP, and it soon came to play an outsize role in their day-to-day operations.
It was in learning to use the software that Bennett and Cascone began to explore the facility’s vast reams of production data. Understanding what they had to work with was a crucial first step and with its visualization focus, JMP enabled exploratory data analysis that helped the pair sift through the noise. “With Libre, there's so much data being generated from the manufacturing process. We very quickly realized how important all of that is,” says Bennett. “But,” Cascone adds, “It was such a vast amount of data, the question became: how do you present it in a way in which you can see what's really going on?”
An operations dashboard is born
As they expanded their knowledge of JMP to include the data visualization platform Graph Builder, Bennett and Cascone grew to understand the tool’s potential. “That's where all the visual aspects of JMP really come into their own,” Cascone says. “That was kind of our initial journey with JMP, wasn't it?” recalls Bennett. “Using Graph Builder for trial-and-error. Then once we’d nailed something down, we’d say ‘That's exactly what I'm looking for.’ And that was when we started looking at the scripting behind Graph Builder.”
Adding a script to run a familiar analytical process in JMP would only make the steps faster still – not to mention that automation would make charts more shareable with others outside of the technical services group. The natural next step was for Bennett and Cascone to learn how to write custom scripts in JMP Scripting Language (JSL).
“We started to use JSL a lot,” Cascone says, “because with JSL, we could save those scripts and send them to other colleagues who could run them without having to do all the trial-and-error up front. You increase efficiency within the team as well, just getting that data quicker.” Saved scripts and journals grew into a small library and the pair began to develop them into a full-fledged automated operations dashboard.
The pair had a breakthrough moment when they discovered they could connect JMP directly to the database, bypassing cumbersome third-party software. This allowed them to further refine the dashboard, responding to management questions with the press of a button. This level of automation has also made the tool more approachable for those with little to no expertise in data analysis. Says Bennett: “Our dashboard and button boxes mean that now [even new employees] can run charts from day one.”
“Eventually, what began with separate plot outputs evolved so that we could get them all to run under one button and be presented in one window from a single data table,” says Bennett. “Now, when you highlight something in one process, it will highlight it in the other process as well to track performance over the entire manufacturing process.”
From hours to minutes – with fewer defects
“Our operations dashboard has become a background tool,” Cascone says. “There's no database analysis needed – you can just run it in the background while you're talking. It'll churn away and pop up the plots that you need to demonstrate the point you're trying to make.”
Instead of 2 to 3 hours, each chart the team produces now requires only 30 minutes. “It's probably equivalent to one or two more headcount per week in terms of the actual work we do,” Cascone shares. “It's 30 or 40 hours of work a week [that we’re saving] just by having all these functions automated.”
More impressive still, the dashboard ultimately plays a role in quality validation and verification, helping the team to identify and resolve potential arisings in advance. “Because we're analyzing so much data so frequently, we can now spot problems before they happen,” Cascone says. And having an automated system like the operations dashboard helps the team to monitor their data even more closely, resolving issues proactively and therefore achieving the higher yields with which they were first tasked upon joining the company.
Unlocking the potential of cyber-connected manufacturing systems
Abbott’s operations dashboard is a sophisticated tool that in very little time, has transformed the facility under a unified data strategy that centralizes all kinds of parameters affecting production output – while allowing employees to quickly and proactively diagnose problems. Imagine, Bennett and Cascone urge, the potential of this system once it is connected in cyberspace with the factory itself.
Ultimately, the pair say they would like to sift data instantaneously, with JMP connected live to the manufacturing systems. “We already know how we would implement [such a system] with JMP and how we could get it to update live – and hold that information. So we're slightly ahead,” says Bennett. “We've got a nice, long wish list, but at the moment it is still just that – a wish list.”
Bennett and Cascone have an opportunity to get their wish soon with advances in cyber-connected systems, which will likely augment Abbott’s efforts to produce more FreeStyle Libre devices. And more people living with diabetes around the world will get access to the devices they need to stay safe and healthy.
“We're in a really good position,” reflects Bennett. “In addition to challenging the norms with our product, we can challenge the norms, as well, in our role. People may have been set in their ways for the last 10 years in manufacturing, but we can now push ahead and say, ‘we're going to do it differently.’”
