Our patient community has been up in arms for quite some time about glucose meter accuracy, or rather lack thereof. Fortunately, the patient-initiated StripSafely campaign seems to have struck a chord, and the FDA has recently -- at long last -- issued new "draft guidance" on accuracy standards. What surprised a lot of people, though, is how FDA chose to split their recommendations into two categories: one set of requirements for meters purchased over-the-counter and used by patients in everyday life, and another for meters used in clinical settings by healthcare professionals. And guess which is required to be more accurate?
You guessed it: +/-15% for home-use meters and a tighter +/-10% for clinical meters.
On top of that, the new guidance is only aimed at NEW products getting ready to undergo FDA evaluation for the first time; it doesn't address the accuracy of meters already on the market -- which disappointed many of us calling for better "post-market surveillance" of meters that may be doing a crap job after some years of use. Also, the new guidance is "non-binding" for manufacturers, so not mandatory. Huh?
Last week, the FDA hosted a teleconference led by Courtney Lias, Director of CDRH/Division of Chemistry and Toxicology, to address these concerns with patient advocates, physicians and interested members of the media.
NEWSFLASH: FDA Clears Dexcom Share Direct
Dexcom gets regulatory approval of its 'on-the-go' mobile apps for CGM data-sharing.
Snail Uses Insulin to Poison Fish
New study shows these slow-moving creatures use toxic form of insulin to capture prey.
A New Square Patch Insulin Pump
Israeli company developing new reusable square insulin pump that has Bluetooth for smartphone communication.
The main point Lias seemed to want to convey is that the FDA is trying to be practical here, by issuing what they believe to be reasonable and achievable guidelines.
"The technology isn't currently sufficient for manufacturers to meet +/-10% accuracy at the huge volume of lay use, but for the smaller volume of tests done in clinical settings, they can... In an ideal world, everyone would have a meter that's within 5-10% of the accuracy reference, but as it stands right now, if we called for that, manufacturers wouldn't be able to provide meters at all."
— Courtney Lias of FDA, on the agency's new draft guidances for glucose meter accuracy
The other major reason for separate guidances is the fact that meters in clinical settings are used on multiple patients, so the cleaning and disinfection requirements need to be a lot more stringent, she said. And the study data submitted for product approval of clinical meters has to include test measures of venous blood, arterial blood, blood from newborn babies, and all the other yucky ways that clinical meters are used. (OK, it's all yucky -- who wants to take a self-blood test, anyway? But I digress...)
Home Use Meters - What's New
Lias says the first portion of the new guidance addresses cleaning and disinfection instructions, which is all about preventing infectious diseases of course.
Next the FDA addresses user evaluations, calling for manufacturers to look at a sample size increased from just 100-150 samples to 350 samples -- plus an additional 50 high and 50 low blood sugar samples "because those are hard to catch," Lias says. This differs from current requirements, which include very little data in the hypoglycemic range.
And on the accuracy standards themselves: 95% of all this data must be within 15% of the "reference method" (for testing accuracy and precision) and a full 100% of the data must be within 20% of the method (to reduce the number of "outliers," aka crazy-off data that may skew the results).
"This is meant to be a mechanism for risk analysis — illustrating why the meter won't pose safety problems," Lias said.
She also said the guidance states that it's acceptable for home meters to measure only as low as 50 mg/dL, which surprised me. But she says "we'll get a more accurate meter that way."
Without getting too deep into the science, there are also new specifications on "interference factors" and hematocrit (red blood cell count) that can affect BG results -- all aimed at reducing the potential impact of testing errors.
FDA will also be taking a closer look at manufacturers' lot release methodology — collecting information about manufacturers' sites and the quality of their production. This will be accomplished through data collection and site inspections, Lias says.
Perhaps most importantly, FDA is calling for new labeling information on test strip vials; they must include information about the lot/production, and a description of the performance (accuracy data) on the outer box label so users can compare one meter to another. That's HUGE! They're finally recognizing that patients have concerns and want to make informed decisions!
The vials will also need to carry "a prominent warning" that these home strips are not meant for use in healthcare settings (i.e. concern over the spread of Hepatitis B).
Clinical Use Meters - What's New
Since we're talking about meters that are used by multiple patients here, the guidance naturally calls for more stringent cleaning and disinfection methods -- although this hasn't changed much from the current requirements, Lias says.
User evaluations are meant to reflect how these meters are employed in the clinic: using venus blood, arterial blood, blood from newborn babies, etc., and manufacturers now need to submit data on 350 individual patients per type, plus 50 high and low samples just as with the home meters.
The tighter accuracy standards here state that 95% of the data must be within 10% of the reference method, and almost 100% of the data must be within 15% of the method -- again to assure minimal outliers. Lisa reminded us that in the hospital, they're often trying to address PWDs just coming out of surgery, so precision accuracy is of utmost importance.
Perhaps the most important change here is that clinical glucose meters are now going to be prescription devices. This means clinics must obtain "waivers" to use them. This seems like an important safety check to be sure what's happening in the clinic/hospital is safe and good for PWDs.
Manufacturers of these clinical devices must also report data on interference factors, of course, and the acceptable hematocrit range is tighter than with home meters, Lias says.
What About Enforcement?
All good, but this is "non-binding guidance," so... huh? How will this be enforced? That was the first question on the conference call, submitted by the president of the American Association of Clinical Endocrinologists (AACE).
"This is pre-market guidance, so it's not about enforcement," Lias said. "Rather these are the requirements to convince us that a product is acceptable for clearance."
She also said it's not simply an "honor system." If the FDA suspects the integrity of data submitted, they can stop the evaluation and send out inspectors to the production site. "But that's not common, " Lisa says. "More common is that we get info based on idealized data and... that may loosen criteria. We need to make sure lot lease criteria are better defined."
(Note that manufacturers must sign a "truthful & accurate" statement when submitting their data, and if they lie, it's a criminal act.)
Jane Chiang, New Senior VP for Medical Affairs of the American Diabetes Association (ADA) asked about "harmonization" with the new European ISO accuracy standard, updated in May of 2013 and used in many countries around the world. Lias said the FDA has not yet recognized the 2013 update because they "don't believe it captures important concerns" (wow!). So they drafted their own new guidances instead.
However, she says FDA is still encouraging vendors to meet the ISO standard for marketing in other countries. And if vendors meet the new stricter FDA standards, they'll automatically cover the international standard too.
"The EU works more on a self-certification system for this type of device," Lias added.
And Post-Market Quality Checks?
What about post-market surveillance? A big focus of the patient-led Strip Safely campaign? Well, the FDA currently does not have a provision for monitoring meters after they've been on the market for 12 months, and though they are "looking into it," it doesn't appear that this will change any time soon.
StripSafely founder Bennet Dunlap asked point-blank: "How can manufacturers justify making new meters with all these tight requirements and still compete with all the cheap low-quality stuff flooding the market?"
The answer here was a bit nebulous. Lias referred to "market forces" and competition. "This will be an issue for a little while. Devices have a lifespan... they can stretch for several years, but will eventually go away," she said. One can also assume that PWD consumers will begin to rely on the new labeling requirements that show product performance data.
Separate But Equal?
One final item Lias mentioned on the call was that some folks at the recent Diabetes Technology Society (DTS) meeting on glucose meter accuracy recommended creating separate accuracy criteria for meters aimed at type 1s versus type 2s. Not going to happen, of course. "Every person with diabetes deserves an accurate meter," Lias said, and we sighed collectively. "But performance data on the outer box labels will now help people choose the meter right for them," she added.
New research just published by the American Association of Diabetes Educators (AADE) underscores the issue: 27 percent of patients with type 1 diabetes and roughly 9 percent with type 2 "had experienced health problems as a result of inaccurate glucose meter readings."
Meanwhile, we inquired with DTS founder Dr. David Klonoff, and learned there are "no updates yet" on a new steering committee that was supposed to be formed (backed by medtech industry group AdvaMedDx and others) to look at issues surrounding glucose meter accuracy and transparency in data reporting.
Patient Community Perspective
Both draft guidance documents remain in the open "public comment period" until April 7, 2014. The FDA wants community input! See below for links.
So what do the forces behind the patient-led advocacy initiative StripSafely think of the draft documents so far? Bennet says:
"I think the draft guidance is a very positive step. It is a reminder that there are many in FDA who are as driven to accuracy as the patient community.
"One of the most encouraging things is the draft is asking for a 'description of the lot release criteria.' I would love to see a similar ask for criteria to triggering medical device reports.
"StripSafely will have sample comments for the guidance docket after a more detailed review. In the meantime we look forward to trading ideas with the community about the strengths in the guidance and opportunities to suggest improvements. I strongly encourage all users to follow and engage in the conversation about the guidance. This conversation at DiabetesMine is certainly a solid start of that dialog."
SUBMIT YOUR FEEDBACK
Right -- we need to have our say!
Lias made a point of saying that the FDA wants to hear not only what people don't like and want changed in these documents, but also what you DO like — which helps them get a better picture of what the community wants overall.
Here are the links to the actual documents and spots to leave comments for each (by 4/7/14):
OTC (over-the-counter) Meters: http://www.regulations.gov/#!