Our Blog / the diabetes online community on accuracy

At a recent public meeting the FDA acknowledged that there are some 510(k) cleared BG meters and strips that do not meet the accuracy standards at which they were approved. There is no clear course of action to insure people with diabetes (PWD) are using blood glucose testing strips that meet regulatory requirements.

For PWD using insulin, meters are the tool used to regulate insulin dosing. Patients using meters that fail to meet accuracy standards increase the risk of overdosing on insulin. Hypoglycemia can result. Underdosing can result in hyperglycemia and even diabetic ketoacidosis.

In addition to acute adverse health outcomes, acceleration of long-term complications may result from inaccurate meters. This will result in higher medical cost to providers and lower quality of life and shortened life spans for patients.

The Centers for Medicare & Medicaid Services (CMS) auctions establish the price of test strips for Medicare and Medicaid patients. In turn, these prices impact private insurers’ reimbursement rates. These auctions are based on FDA 510k approved systems. Lower quality, less costly strips are likely to win these auctions, but there is no mechanism between FDA and CMS to maintain the quality of strips in the market.

It appears that foreign manufacturers are selling substandard strips to US consumers.

Success of the StripSafely campaign and lessons learned in bringing this issue to light may help future efforts to address other diabetes related public health issues, like pumps being part of CMS health guidelines for older Americans.

Every PWD should call for:

  • Ongoing testing of BG strips to assure compliance with regulatory accuracy standards.
  • Quality assurance on strips acquired through the distribution channels people use.
  • Acknowledgement that the accuracy of strips is a public health and safety issue.
  • Standards for accuracy of all strips sold in the USA improved to the latest ISO standards.
  • FDA and Congressional consideration of the adverse effects that can come from meter inaccuracy.
  • The creation of a process for CMS competitive bidding that consider the quality of what they are buying – not just price.
  • Better standards than those created in 2003.

Bennet Dunlap

One Comment

  1. Adrien de Croy

    I believe the problem starts higher up – at the ISO. I have purchased and reviewed ISO 15197:2013, and there are IMO problems with it which are at the root of the issue of strip (or more generally meter system) safety.

    The specifications for testing / evaluating metering systems are based on a user (e.g. clinician or diabetic) testing regime which does not consider system design, yet sets test limits in some key environmental parameters.

    This would be like an airline choosing a new plane for its fleet based solely on pilot evaluation, and worse still where the testing could only occur at a small range of the speeds and altitudes that would actually be used in service. How can we have any faith in the results of such tests? We can’t.

    As we know, different system designs can introduce or address inaccuracies based on a number of environmental factors, such as temperature, haematocrit, pH etc.

    Furthermore, patents and lawsuits force companies to try novel ways to do things like estimate haematocrit, which may only work in certain environmental conditions.

    Different designs behave differently, and therefore the design must be considered in the evaluation of the device. The ISO standard does not contemplate this, and even worse removes certain parameters from the evaluation completely. The most striking example of this is the removal of temperature as a tested parameter by specifying that all tests should take place at 23 +/- 5 degrees Celcius.

    This means it’s possible to pass clinical trials based on ISO 15197 without ever being tested for temperature sensitivity. We have seen here in New Zealand (and you all may soon see) the result of this, where manufacturers in an attempt to compete on strip price, design systems that do not measure strip temperature at all!!!. This is unthinkable, as there is never any guarantee that the meter internal thermal sensor will be at anywhere near the same temperature as the enzyme reaction, and the enzyme reaction is extremely temperature sensitive (factor of 3:1 from 10 to 25 deg Celcius). This can lead to readings that are many times higher than the real value when the meter is cold, which then leads to insulin overdosing.

    So the solution will require some changes to the specifications at the top, and in my view should require some system analysis to determine likely sensitivities, and explicit tests for these to allow proper evaluation of the device likely performance in the real world. At the very least the devices must be tested across the whole range of environmental conditions (e.g. temp, humidity) that is advertised by the manufacturer. A manufacturer that could only advertise that their device is usable at only 23 +/- 5 deg, would have much more difficulty competing.

    In my view it’s impossible to make an enzyme-based system safe for general use, if it does not properly account for temperature, and this requires accurately measuring or controlling the reaction temperature. So systems which do not do this, should not be considered even as candidates.


So, what do you think ?