By Julie Shaw, Lauri Infanti, and Paul Yip
The Centre for Proficiency Testing Point-of-Care proficiency testing program includes two surveys which ship in May and October for testing and reporting on glucose meters. The last two May surveys (2017 and 2018) have included questions designed to gain insight into patterns of practice related to quality assurance for POC glucose testing.
POC glucose testing is the most common POC test performed in hospitals and is important for monitoring of glycemic control in patients. More recently, tight glycemic control for hospitalized patients has been shown to lead to improved outcomes, particularly for surgical patients where tighter glycemic control, with less hyperglycemic events, has been linked to a reduction in surgical site infections.
Ongoing quality assurance activities related to POC glucose meters are key to ensuring the accuracy of results. These activities include but are not limited to: the regular comparison of POC results with those from the central laboratory (if the institution has a central laboratory), validation of reagents and validation of new/replacement meters. These are also the IQMH requirements for medical laboratory accreditation.1
Clinical users need to be made aware of any differences in results that may be expected with the introduction of a new lot number of glucose strips. For example, as in some cases, subtle changes can lead to different treatment decisions depending on the treatment protocols in place.
Quality control and precision goals
In May 2017 and 2018, questions were distributed to IQMH participants enrolled in the glucose meter survey revealing that many laboratories are using manufacturer targets for quality control, including the mean and range. However, it is simplistic to treat the range as QC error limits since the values are not specific for individual reagent strip lots. This can be a limitation of the POC meters themselves or of the middleware used to manage the meters. The values may be set by default, and extra effort is required to adjust the target mean and SDs. The default limits may be quite wide, which can cause errors to be missed when individual meters are considered. Where possible, goals based on actual meter performance are preferable.
Comparing meter performance to the main laboratory
Ongoing validation is required to monitor the performance of the meters, particularly with new test strip lot numbers, or following a period of extended use (e.g. at least every six months). Bias goals will depend on the clinical needs; for example, small differences in values may lead to different treatment decisions for hypoglycemia in neonates. Clinicians are not necessarily aware of the limitations in precision and accuracy at the low end of the measuring range with glucose meters.
Comparison of results from the meters to the central laboratory is crucial as clinicians will use these values interchangeably and need to be aware of potential differences. It is important to educate clinicians on potential bias between the laboratory and the meters, as this aids in their clinical decision-making. The neonatal hypoglycemia example demonstrates this. Communication of the potential size and direction of the bias will be important for action thresholds. Correction factors are usually not advised; while instead, unacceptable bias may require a lot to be rejected. Therefore, verification of glucose meter accuracy should include comparison to the central laboratory analyzer.
Total allowable error goals
In the responses to the same questionnaire, sites have indicated a variety of sources for their total allowable error goals.
The total allowable error goals
In the responses to the same questionnaire, sites have indicated a variety of sources for their total allowable error goals. The total allowable error goals for any POC test must ensure that the results are fit for purpose for clinical decision-making. An important point is that one of the current standards, ISO 15197:2013, specifies requirements of glucose meters for self-testing, with a performance requirement of ±15% (or 0.83 mmol/L) for the accuracy of testing blood. The current IQMH allowable performance limit for the proficiency testing survey is 20% (or 1.0 mmol/L). The latter is wider to reflect the number of users, nature of the matrix, and other factors related to the survey. Even tighter requirements have been specified in other standards including, CLSI POCT12-A3: Point-of-Care Blood Glucose Testing in Acute and Chronic Care Facilities: Approved Guideline – Third Edition.2 Each participant site must decide the clinical requirements for accuracy within its organization.
Training for contraindications — glucose meters are used in areas with critically ill patients
Those overseeing POC programs must consider where glucose meters will be used and whether they are robust for certain patient populations. Patient conditions or drug administration may impact specimen quality. The use of glucose meters in the critically ill population has been the subject of debate recently. Validation of the meters should include specimens from critically ill patients if the meters are to be implemented and used for testing on critically ill patients. The effects of certain interferences need to be understood (e.g. hematocrit, ascorbic acid, acetaminophen) as well as limitations for testing based on certain patient conditions (e.g. hypoperfusion where capillary specimens are not appropriate).
Clinical users of POC require ongoing education on limitations of POC glucose testing to ensure patient safety. Increasingly, training for POC is facilitated through online modules, which can pose challenges in linking theoretical concepts to practical clinical situations. CLSI has published a free white paper to assist users in the limitations of testing, POC17Ed1E:2016: Glucose Meters for Critically Ill Patients, 1st Edition.3
Critical glucose results practices
The questionnaire in the May 2018 survey included questions about policies for repeat testing when critical glucose results are obtained. A variety of practices were identified. Pre-analytical factors, such as taking specimens from IV lines or from patients with contaminated hands can result in erroneous glucose measurements. For example, acting upon a falsely elevated glucose concentration with insulin treatment is dangerous. Repeat of critically high glucose results can help catch these errors but will not identify all falsely elevated glucose results.
It is important for clinical staff performing POC glucose testing to follow the recommended procedure for testing: properly clean patient fingertips, wipe first drop away, or if taking a specimen from a line to flush what was infused previously. This will help prevent potentially dangerous measurement errors.
As the benefits of glycemic control on patient outcomes continue to grow, the use of glucose meters on complex patients in the hospital setting likely will increase in parallel. The laboratory plays a key role in the oversight of POC programs, and consideration of patients in different clinical areas is needed to ensure the accuracy of results.
1. Institute for Quality Management in Healthcare (IQMH). ISO 15189 Plus™ Point-of-Care Testing Accreditation Requirements. Version 7.1, April 2017.
2. CLSI. Point-of-Care Blood Glucose Testing in Acute and Chronic Care Facilities; Approved Guideline—Third Edition. CLSI document POCT12-A3. Wayne, PA: Clinical and Laboratory Standards Institute; 2013.
3. CLSI. Use of Glucose Meters for Critically Ill Patients. CLSI white paper POCT17. Wayne, PA: Clinical and Laboratory Standards Institute; 2016.