We focus on the core principles of the automation industry in medical manufacturing and its potential growth and future development.

shutterstock_1874898838

The evolution of automation has predominately been around the ability to assess devices visually and functionally. (Credit: Elpisterra/Shutterstock.com)


Automation has undergone technological updates over the years, leading to more flexibility in what manufacturers can build. But have the economics behind setting up an automated product line changed that much in the past few decades? Peter Littlejohns speaks to Ken McClannon, technical business unit director for Jabil Healthcare Pharmaceutical Delivery Systems group, who explains how the core principles of why we automate have remained the same since he started working in the field in the mid-1990s.

The word manufacturing no doubt evokes images of mechanical titans churning out items on a conveyor belt. But ask the average person what they think the sector looked like 30 years ago and you can bet they’ll describe more grubby jumpsuits, hard hats and safety glasses, with human hands still carrying out a portion of the labour.

It’s true that medical device production now looks different to back then, but according to Ken McClannon, technical business unit director for Jabil Healthcare Pharmaceutical Delivery Systems group, the comparison isn’t as night and day as some might expect.

“I’m 25 years working [in manufacturing] now,” he says. “I’ve been working on fully automated production lines for 25 years, so for me, there was no ‘before automation’. The automation has evolved over that time, but the principles of high-volume automatic production are still the same as they were when I started.”

In McClannon’s area of the Jabil business, the inhalers, epinephrine auto-injectors and other drug-delivery systems churned out in their millions are never touched by a human hand until they reach the end customer. Far from a slow build up in technological capability to reach this high-volume production though, Jabil, like many other contract manufacturers, merely pivoted their already established automation capability from one industry to another out of necessity.

“We were making mobile phones for companies like Nokia and Motorola. The first flip phone Nokia ever produced, we manufactured,” says McClannon. The catalyst for change came when consumer electronics started to become a fashion accessory. With a near-constant stream of competing models released into the market, the necessity to innovate meant products that once remained desirable for years could become obsolete in one.

“If you’re lucky, you have six months to put the manufacturing process in place for it,” McClannon explains. “Even if you wanted to automate that, it doesn’t make sense too because the phone isn’t going to be around long enough to give you any return on your investment.”

As the process required to manufacture ever more sophisticated consumer electronics migrated from an automated business in western European high-cost countries to a manual process in countries with low labour costs, McClannon says contract manufacturers looked to the medical device industry as an alternative market. “What product category lends itself to automation and has the longevity and the volume that would justify manufacturing it in western Europe or the US? Medical devices.”

A reason to automate

There’s a lesson to be learned from contract manufacturers’ pivot to producing medical devices, and it’s about the limitations, as well as the advantages, of automation. Although the technology behind it has improved, the reasons behind pursuing it have not. The three big reasons are safety, quality and cost, with the last one often the main driver.

“The decision to automate or not, if you’re trying to cut cost, still comes down to product volume, longevity and complexity,” says McClannon. “If you’ve got decent volume, the product is going to be around for a reasonable period of time and it’s not too complex, you can get a good return on your investment.”

Despite all of the shiny new technology powering automated production lines and the constant buzz around what’s often called the Fourth Industrial Revolution, or Industry 4.0 – the same challenges that existed 25 years ago are still barriers today.

“We make the electronics for a lot of medical devices, and our electronic production factories have automated lines for producing the PCB assembly,” says McClannon. “But putting something like a defibrillator together or a diagnostic testing machine is still very manual.”

It’s not that automating this part of the process would be impossible, the technology is certainly there to enable it. But there must be an economic reason to do so. “The volume of these products doesn’t justify putting automation in place, and neither does the complexity of their assembly,” says McClannon.

On the other hand, products with fewer parts, like those in Jabil’s drug delivery devices business are still being produced based on cost-benefit decisions made more than more than a decade ago. “We have one automated product in one of our sites in Ireland that’s been running for 22 years,” says McClannon. “I can’t think of another product category that would run for that long. Coca Cola maybe?”

A more flexible automation system leads to more scalability

This sort of consistency is ideal for a product that needs to be the same every time to ensure there’s a low risk to patients. The longevity of an automation line can present its own problems, however, as the parts required to maintain it improve and become more sophisticated.

In the context of that 22-year product line run by Jabil, McClannon explains: “Computer hardware 20 years ago is completely different to what it is now. You’d have more in your mobile phone now than in a desktop PC from 20 years ago. So, how you maintain the [automation] equipment and controls is very important. You don’t want to end up on eBay looking for antique spare parts.”

Another challenge in the pursuit of bringing a product to market is planning the scale of output required to meet demand. It requires a prediction of what the peak capacity is going to be before the product is launched, before a company has any idea if it’s going to get regulatory approval and before it has any feel for market acceptance. After that, it’s a contract manufacturer like Jabil that will build the capability to produce it.

“In the early years, as you’re ramping up that capacity, your line is somewhat underutilised,” says McClannon. “Then you get to your sweet spot in the middle, where the line is the right size and it’s giving the right quality and cost savings.” The last phase of a product life cycle is its end of life, where demand is significantly reduced.

It’s within the first and last phases that scalable automation is desirable. But adjusting a process to scale output up and down is easier said than done in a fixed automation system and requires engineering adjustments that carry their own cost.

For Jabil’s automated line that’s been running 22 years, the product it continues to churn out is in its end-of-life phase, and rather than scale production down on its current machinery, McClannon says the company is building a smaller line to replace it.

“You could argue that if the automation was scalable, you could reduce the volume and wouldn’t need to buy a new line,” he says. “But after 22 years running, it’s so worn out that, due to product supply risk, we’re better off with new equipment. It’s like running a car into the ground – there comes a point where it will let you down.” A scalable production line is easier to achieve with a more flexible automation system, rather than the fixed automation machinery McClannon is most used to.

But those alternatives come with higher investment costs due to more sophisticated technology, and also output lower volumes. When much of the reason to automate is the rate of production bringing down the cost per unit, the decision of whether to make the process scalable becomes “a question of economics, not technology”, McClannon adds.

automation
A scalable production line is easier to achieve with a more flexible automation system, but that comes with higher investment costs due to more sophisticated technology and also outputs lower volumes. (Credit: i viewfinder/Shutterstock.com)

The cost of technology for automation

Creative problem-solvers are coming up with new medical devices all the time, but if McClannon’s experience is anything to go by, their go-to-market vision is often far from the reality they encounter. He says start-up companies often have had brilliant ideas for wearable insulin pumps and have poured significant time and money into developing them, but then expect manufacturers to establish the capacity to produce them in as little as a year for a fraction of the cost.

“They completely underestimate what’s involved when you get up to that level of product complexity. I’ve had situations where a company has budgeted a certain amount, and we’ve come back and said it’s four times that. That’s been the difference between them having the money to launch and not being able to, and I’ve seen companies fail because of it.”

What adds to that frustration for start-ups is that the tightly regulated nature of the medical device industry necessitates a level of product standardisation that forces companies to choose automation because manual processes can’t give the same level of quality on a repeatable basis.

McClannon says there’s also an issue of device makers wanting the benefits of new technology without realising it has disadvantages – the most obvious being cost. The shell of a wearable pump, for instance, needs to be water-tight, and one of the ways you can achieve that type of seal is laser welding.

But while “everybody goes ‘woah, laser welding, sexy new technology, love it’”, McClannon says it’s “really expensive and has a really slow cycle time, so it kills efficiency”. It’s this cost and the often-unproven nature of new technologies that lead McClannon to avoid them, specifically telling Jabil’s designers to stick with what’s proven.

“Anytime our designers are talking to me about design for assembly, I tell them the more you can stay with the mainstream technologies, the more reliable and economic [the manufacturing] is to implement. Healthcare doesn’t want to be at the cutting edge of technology, we want to be second, not first.”

 

Quality and assurance

Automated medical device manufacturing may not have undergone any sea-change moments in McClannon’s 19 years at Jabil. But he originally started his career designing vision systems – a key component of product quality and assurance – and says much of the evolution in automation has been around the ability to assess devices visually and functionally.

“When I was in vision systems, I would sit at my desk with a pile of maths books on one side and a pile of C programming books on the other side.” In these early days, McClannon would create the algorithms required so that an automated vision system could measure, for instance, the diameter of an object.

He’d then translate the algorithm into code for the machine to understand. The growth of the automation industry has since led to companies that specialise in medical device manufacturing, and huge libraries of algorithms that can be adjusted simply by setting the custom dimensions of parts.

The result is far greater reliability, McClannon says, with one of the automation lines he oversees having 48 vision systems to check a product made of ten parts. “To put 48 vision systems on a line 20 years ago would have taken a bunch of PhD guys probably years of development.”

It’s quality and assurance McClannon expects to see the most development in the future too. There’s already a growing trend of serialisation allowing for products to be tracked through the supply chain.

But he believes adding the capability to collect data as parts run through the automation process could be a valuable next step in determining if errors have occurred. “I’ve seen a situation where a new device has been developed and had a 0.1% functional failure rate,” he says.

“At that failure rate, the FDA wouldn’t have approved it.” What followed was a four-month systematic root cause investigation in which 72 experiments were performed to eventually narrow down the issue and fix it. “If we were automatically collecting data, we could have run statistical tools on it that would have spotted the anomaly very quickly and we wouldn’t have had the stoppage.”

This article first appeared in Medical Device Developments Vol. 2 2021. The full publication can be viewed online here.