An In-Depth Look at Diabetes Device Recalls (Part 1 of 4)

Multitudes of diabetes devices have been pulled from the market or flagged for possible flaws over the years, from glucose meters and test strips to insulin pumps, continuous glucose monitors and other products we rely on to stay healthy.

Of course, we have a big federal agency overseeing these regulated products, and an incredible amount of time and energy goes into responses to recalls when they come up. But with hundreds of such incidents recorded during just the past 15 years, that’s a lot of responsibility to carry and a lot of margin for error.

And with so much potential for danger, we have to wonder: Is this process working properly as it should, to keep us patients safe?

Simple answer: We don’t know. We've learned that data on FDA recalls is incredibly complicated to analyze, which makes it tough to reach any overarching conclusions.

We've spent the past several weeks examining FDA records and delving into various aspects of this issue -- industry, regulatory, policy, legal and of course patient side -- and frankly, we're left frustrated at finding more questions than answers.

But we did learn a whole lot, which we're happy to present to you all in a Four-Part Series on Diabetes Device Recalls, beginning today with a broad look at trends and policy. 

History & Trends

The FDA has historically been criticized for how it tracked recalls and trends in the medical device industry. That was detailed in an FDA report in 2014, which found the number of safety recalls had nearly doubled during the decade from 2003 to 2012, and many of those recalls were more serious than they had been 10 years ago. Another comprehensive review shows that hundreds of FDA recalls were recorded since 2002, and have increased in frequency in recent years.

Specific to diabetes, a few horror stories and mass recalls stand out:

• 359 million Abbott FreeStyle strips recalled in 2010 because they gave false low readings
• 62 million generic NovaMax strips recalled in July 2013, because they gave incorrect high readings that led people to dose insulin when they didn’t need it
• Most of the 2012-manufactured Animas 20/20 insulin pumps recalled in early 2013 after J&J found an issue that could lead to unintended insulin delivery. This followed another J&J recall of the LifeScan glucose meters, over concerns that the devices could simply shut down when encountering dangerously high levels of blood sugar
• 2.9 million Abbott FreeStyle strips worldwide recalled in April 2008 because of a technical issue causing unnecessary error messages on the meters
• Product liability lawsuits filed in the aftermath of 2013 insulin pump malfunctions

Those are the big news headlines, the ones that put us on edge, although most device recalls aren't quite so dramatic.

Regulatory policy experts and industry folk say you can’t draw conclusions based on those numbers alone, because the widespread recalls stem from better quality control measures put in place by manufacturers. In other words, they're being overly cautious by pulling the largest possible swath of product from market, in the case of any suspected flaw. Adding to that is heightened public awareness of recalls fueled by mass media and social media coverage, which can lead to more product complaints and reports and, ultimately, more recall notices.

“Recalls are interesting because a lot of times, people think that if there’s a recall, then something bad has happened. That’s not how we always see it. Many times a flaw is detected without any injury reports, and it's just proactive quality control,” says Dr. Courtney Lias, director of the FDA’s Division of Chemistry and Toxicology Devices within the agency’s Office of In Vitro Diagnostics and Radiological Devices. “Sure, there are times when a company has a lot of recalls because they weren’t doing something they should have been doing. But mostly, it’s because they are doing what they should be.”

Recalls By the Numbers

So, just how many recalls have there been on diabetes devices?

Seems like a simple question, right? Especially since the FDA is in charge of regulating medical devices and the federal agency keeps a searchable database of recall records dating back to late 2002.

Well, not so fast. Tracking the amount of recalls is no easy task.

That's largely because the FDA's recall database doesn't tag instances by type or disease. This is a separate database from the one FDA maintains listing reports of medical device complaints and possible defects, known as Adverse Event Reports or Medical Device Reports (MDRs). There are tens of thousands of those filed each year, and while the information is sometimes used to raise red flags, the actual product recalls are logged separately in their own dedicated database.

For starters, the fact that there are separate, unconnected databases for this type of information poses a problem for anyone trying to track patterns or connect dots on medical device safety issues.

Even those who work closely with the FDA admit that the system is not user-friendly, and there are many limitations to what analyses can be conducted. In either database, it's not nearly as simple as just searching the term "diabetes devices," because you can't. You can search by specific product or company names, but that's not efficient because it only brings up incidents associated with one manufacturer. Official product classification codes can be used to search, if you happen to know them. But even that is much more complicated than it should be. The code for insulin pumps, for example, turns out to be LZG (really, WTF?!)

Searching for broad terms like “diabetes,” “glucose,” and “insulin” turned out to be our best (and only) way to explore the recalls database. Those three terms alone produce 385 entries. It took us quite a bit of legwork to figure out which of those were duplicates and parse them down. We can't be sure that's all the diabetes device recalls, but we're pretty sure these terms make up the brunt of those in the database.

Eventually we could see that the data shows about 21% of the total FDA recorded recalls for those terms involved clinical diabetes devices that patients don’t use at home.

We kept searching for the outpatient products, but it ain't easy, let me tell you!

For example, you’d think the rash of insulin pump recalls from a decade ago would fall under the "insulin pump" category. But instead many are labeled "infusion pumps" and lumped with pumps that deliver other medicines, and some entries actually contain typos in the name -- like “isulin” pumps, so you're at a loss.

Of the 385 recalls we managed to find in the system, Roche had the highest number of entries at 73, while Abbott followed second with 62 recalls and J&J LifeScan with their OneTouch and Animas brands came in third at 53. Meanwhile, Medtronic registered 17 recalls while other companies had single digits. Generic and off-brand meters and strips showed 24 recalls, while insulin pens and needles had 35 listings.

Context & Timing

“You can’t just look at the number of recalls and make conclusions based on those numbers alone,” cautions the FDA’s Lias. “You have to take into context what the recall was, how it was found, and what else was going on with the company at the time. That’s hard to do from the outside, because you don’t have those conversations and context all the time.”

Take the recent Dexcom receiver recall that's been abuzz in the Diabetes Community, for instance.

On April 11, the FDA tagged the Dexcom G4 and G5 CGM receivers with its most serious Class 1 recall -- but only after the California CGM maker had, in mid-February, voluntarily issued a “customer notification” about its receivers not always alarming for low blood sugars. Only those products shipped before March 10 are affected, and the company is proactively replacing the receivers as needed -- up to 263,520 of them shipped out between October 2012 and March 2016. Whoa!

So the official FDA recall announcement was late to the game, maybe based on confusion? We found six different entries in the FDA database all on the same date relating to the same Dexcom technical problem.

It turns out that every individual recall notice is listed in the FDA database, meaning that if a company puts out multiple notices relating to the same problem -- say it affects different meter brands, or they have one for the U.S. versus internationally -- then the FDA database could show multiple entries for the same defect or potential problem.

This also highlights a dilemma of the system, that by the time the FDA issues its official recall notice, manufacturers have often already addressed the particular product problem -- but no one realizes this because neither the companies nor the FDA publicize the fix, even when the recall is officially listed as "terminated," in a line buried within the FDA database.

"Often in that time (lag before the FDA sends out a recall notice), we've pulled the product and fixed whatever the issue is. It's history for these companies already, but then later the FDA sounds the alarm and people think something’s wrong," says industry insider David Chadwick, Director of Regulatory Affairs and Regulatory Science at Indiana-based medical device manufacturer Cook Medical.

Chadwick has been observing and working with the FDA for decades and previously worked at insulin pump company Disetronic, which was acquired by Roche Diabetes Care in 2003 and paved the way for the Accu-Chek Spirit pump.

“Recall is a word that is used in the lay press so much and it's often misunderstood,” he says. “It's very hard in our current environment for a consumer to follow recalls and stay on top of any trends. I don't know how one can connect these dots and make it understandable. So much of it comes down to the language we use, knowing when we need to sound a four-star alert instead of just addressing a minor labeling issue."

Recall Confusion

Confusing is putting it mildly, so no wonder the patient community is bewildered by all this scattered information.

As noted, it's not uncommon to see a company send a notice to patients about a recall. Then, weeks or even months later, we see another notice from the FDA. It’s the exact same recall issue, but who would blame folks for thinking that there are now two recalls from the same company and wonder why something isn’t being done about it?

Regulatory expert Phil Phillips in Washington D.C., who's consulted exclusively with the JDRF primarily on Artificial Pancreas issues for the past eight years, says many consumers have a misguided view of medical device recalls. Before his consulting days, Phillips led the FDA's Office of Device Evaluation for roughly two decades.

He says medical device users all too often equate the recalls to what happens in the auto industry and with consumer products, where the problems are design and safety related. But with medical devices, a recall can be as basic as a label misprint or a specification that wasn't met in labeling the product -- because the labels are scrutinized extremely carefully in the medical world.

Echoing Lias' comments, he says. "If a company is in business long enough, sooner or later a recall situation will be encountered. When circumstances warrant a recall, conducting a recall actually suggests that the company is conscientious and not necessarily that the company has done a poor job."

There are varying classes of medical device recalls, with some being more serious than others, and a vast majority are not likely to result in patient injury or death. In addition, all these medical device recalls are considered "voluntary," meaning that the FDA suggests but does not enforce them, leaving companies to institute them.

Welcome to the convoluted world of medical device regulations.

FDA Ups and Downs

The FDA tells us that their recalls policy really hasn’t changed over the past decade or longer, although updated communication guidelines were released in October 2014 on how the medical device industry and the agency should interact when addressing recalls.

While the FDA doesn't enforce recalls itself, the agency does sometimes assist in the process. For example, in the past they've worked with manufacturers to remove faulty glucose meters from healthcare facilities, where those devices were giving false high readings. According to FDA data, 13 deaths occurred as a result of these faulty in-clinic meters between 1997-2009. The adverse event reporting system did contribute to this issue eventually being caught and dealt with.

But Lias says the FDA has been working internally to implement new and improved ways to spot trends and communicate with companies about such issues. She wasn’t able to provide details yet, but said those changes are already helping the agency identify issues faster and contact manufacturers in a timely manner.

From the industry side, Cook Medical's Chadwick says he does think the FDA has been doing better in overseeing this process in recent years, and the industry too has improved how it reports this data and responds to these issues.

Phillips, the former FDA device evaluation guy now consulting with JDRF, agrees, saying, "FDA has devoted increasing resources to recalls and has become much more involved in recall processes. Clearly, the system is much improved today compared to 1981 when I entered the agency. Also, recall information is shared more widely within FDA than ever before. This means there is an increased likelihood of FDA applying lessons learned from the post-market period in the pre-market review of new devices."

But not everyone is convinced that things are improving fast enough.

Medical device policy expert Richard Vincins, VP of Quality for consulting firm Emergo Group in Texas, has been publicly critical of the recall situation for several years now.

“What’s most interesting to me is that design controls were implemented 20 years ago to help prevent device recalls, yet there has been this increasing trend," he says. "Even though organizations may be more compliant today in terms of reporting of device recalls, I can still see a continued need to assure proper design and development processes are implemented."

Vincins is particularly concerned about quality assurance when it comes to medical software. "As medical devices become more complex, manufacturers must implement proper design controls, design change management, verification testing, and validation testing to help them save valuable dollars on the bottom line and reduce regulatory oversight," he says. Not to mention patient safety!

Total Industry Disclosure (Sometimes)

We asked several of the diabetes companies about the recall process, and they all say they have comprehensive "corrective action plan" templates in place for when product issues arise, and follow similar protocols in dealing with the FDA, healthcare professionals, distributors, and patients. Beyond that, most of these companies will provide only "canned" responses saying they hold patient safety as the highest priority and do everything they can to ensure quality products.

Interestingly, in this case Medtronic provided the most detailed response, with a lengthy statement from spokesperson Pam Reese that included the message: “Medtronic continually evaluates feedback we receive from customers through the 24-Hour Helpline and other channels about situations they encounter with our products. We use this information to determine when there might be an issue that we need to let our customers know about. Once we identify an issue, we work as swiftly as possible to identify the cause, correct the issue, implement the change, while notifying regulatory bodies and communicating with customers and their healthcare professional. However, there is no set timeline for how long this process might take as every situation may vary.”

While the federal law and FDA policy dictates that companies follow plans on communication with patients, the specifics can vary from blanket social media notifications and website posts to direct phone calls and certified letters to proactive pursuit of news coverage about the issue.

Lias reminds us that the FDA does not dictate what the companies must say to customers or how they contact people; it's role is just to assure that such communication takes place. However, if a recall issue continues or the FDA receives complaints from customers that they aren’t being notified properly, then the agency is able to step in and demand specific actions. A lot of companies do struggle with that process, Lias said.

We’ve covered the issue of communicating recalls to the patient community before, and found that (surprise, surprise!) opinions vary, but the linchpins seem to be getting the information out swiftly, and in a prominent place that anyone can easily find. Patients are most upset when they feel the information's been withheld for too long, or hidden on some obscure web page that's not easily searchable.  

State of the Recall Union

So in summary, the situation with diabetes device recalls is plagued by several core issues:

• The whole FDA recall process is dependent on a database that's clunky and hard to use -- which makes it difficult to analyze the data and learn from best practices over the years
• In part due to this clunky infrastructure, there's often a lag time between company notifications of recalls and the FDA's official announcement, which can lead to duplicate notifications and inevitable confusion
• Despite design and quality control efforts, the sheer numbers of industry recalls is increasing, not diminishing
• The FDA can essentially only make recommendations, leaving companies to choose their own plans of action and communication methods for recalls

No matter what the FDA says or how companies tout their action plans, it’s clear that by design this isn't very "patient centric." In the end, people with diabetes are often the ones left out of the loop by this disjointed system, that is supposedly meant to help us keep tabs on the medical devices that our lives depend on.

- - - - - - - - - - -

Written by Michael W. Hoskins and originally published at DiabetesMine