Editor's note: Katie Scheele, PMP, MBA, BSMT(ASCP), makes her debut as one of MLO's clinical experts. Now “retired and enjoying it,” Katie was a Project Manager and worked at Mayo Clinic in Arizona from 1997 to 2014, managing clinical, administrative, Information Technology, and process improvement projects.
Q: I have read a lot of information regarding Quality Control but have never seen this issue addressed. In the “olden days” you would send in your LJ charts for data comparison. Maybe three to four weeks later you would receive a group comparison report. You'd look at the data and perhaps make some mean (range) changes. Now, with instant reports, how often is “too often” to be making a mean change? Is one to two months too often to adjust the mean? Is there some type of general guideline on how often mean changes should be made?
A: Laboratory quality control (QC) is designed to detect, reduce, and correct deficiencies in a laboratory's internal analytical process prior to the release of patient results. QC is a measure of precision, or how well the measurement system reproduces the same result over time and under varying operating conditions.1-3
Specimens that are analyzed for QC purposes are known as control materials. To minimize variation, control materials should be available over an extended period of time. Minimal variation should exist and there should be sufficient material from the same lot number or serum pool for one year's testing.2-4
Statistical process control is a set of rules that is used to verify the reliability of patient results based on statistics calculated from the regular testing of control materials. Westgard rules and Levey-Jennings charts are the statistical tools typically used by clinical labs. The most fundamental statistics used by the laboratory are the mean [x] and standard deviation[s]. The mean, or average, is the sum of the control values divided by the total number of values. The standard deviation measures how far, on average, the numbers are from their mean.1-3
When a process is within control, approximately:
- 68% of all the QC values fall within ±1 standard deviation (1s);
- 95.5% of all QC values fall within ±2 standard deviations (2s);
- 4.5% of all data will be outside the ±2s limits when the analytical process is in control;
- 99.7% of all QC values are found to be within ±3 standard deviations (3s) of the mean.1-3
Any value outside of ±3s is considered to be associated with a significant error condition. For QC results, any positive or negative deviation away from the calculated mean is defined as random error. There is acceptable (or expected) random error as defined and quantified by the standard deviation (inside the ±3s limits. There is unacceptable (unexpected) random error that is any data point outside the expected population of data (outside the ±3s limits).5-6
The change in the mean may be gradual and demonstrated as a trend in control values; this indicates a gradual loss of reliability in the test system. In fact, trends are usually subtle. But a trend may be abrupt and demonstrated as a shift in control values that may result from the unexpected loss of system reliability.5-6
Also, consider if the change in the mean is statistically significant by calculating these two outputs of statistical testing.5-6
- P-value: The probability value indicates the probability of observing the difference if no difference exists.
- CI around difference: A confidence interval around a difference that does not cross zero also indicates statistical significance.
So how often is “too often” to change the mean and the range? It's all in the numbers.
- Westgard JO. Westgard QC. https://www.westgard.com. Accessed May 1, 2015.
- Laboratory Quality Control; Wikipedia. http://en.wikipedia.org/wiki/Laboratory_quality_control. Accessed May 1, 2015.
- Cooper G. Basic lessons in laboratory quality control. http://www.qcnet.com/Portals/50/PDFs/QCWorkbook2008_Jun08.pdf. Accessed May 1, 2015.
- Basar G. Quality control in the clinical laboratory. http://www.slideshare.net/drgomibasar/quality-control-in-clinical-laboratory. Accessed May 1, 2015.
- Sauro J. Measuring usability, customer services, and experience. http://www.measuringu.com/blog/statistically-significant.php. Accessed May 1, 2015.
- Statistical Significance; Wikipedia. http://en.wikipedia.org/wiki/Statistical_significance. Accessed May 1, 2015.
Q: I ran a glucose on one patient on both a gray top tube plasma and SST tube serum drawn at the same time and I got the following results:
- Gray top tube plasma glucose: 91mmol/dl (reference range 65-100).
- SST serum: 101mmol/dl (reference range 65-100).
On the serum result, our LIS highlighted it as ”High“ because it was greater than the upper limit of the reference range. My supervisor said there was no significant difference since it was not >10%. My supervisor indicates that we should report the gray top results. My worry is that just comparing two results without considering the reference ranges and interpretation may lead to giving wrong results. From what I know fluoride oxalate in gray top stabilizes glucose after two to three hours but won't stop it from falling (breaking down) in the first hours. And SST spanned and separated in time prevents glucose breakdown. Please advise me on this.
A: Because serum glucose metabolizes in collection tubes, serum glucose levels may decrease over time.1 If the blood is allowed to clot, the glucose in the sample gets metabolized by the blood cells unless the cells are separated. If there are higher numbers of red or white blood cells, there is excessive glycolysis in the sample with substantial reduction of glucose level. This occurs if the sample is not processed immediately and can lead to a faulty result.2
Collection tubes have been designed to, at least partially, block glucose metabolism by red blood cells in blood collection tubes that may not be analyzed immediately after blood collection. In a comparison of glucose values on different tube types, the results suggested that red-top tubes with serum separator (SST) or grey-top tubes with a fluoride glycolysis inhibitor may be used for reproducible glucose determinations.1
Several sources recommend the use of grey top tubes for analyzing glucose levels.3-5 The anti-glycolytic properties in grey top tubes prevent the blood cells from using the glucose in the sample. The grey top tube contains potassium oxalate as an anticoagulant and sodium fluoride as a preservative that is used to preserve glucose in whole blood. Oxalate and EDTA are anticoagulants which prevent the blood from clotting; sodium fluoride is a stabilizer. The resulting supernatant fluid is plasma. Sodium fluoride is an anti-glycolytic only, and the resulting fluid is serum.6
Grey top tubes should be at least 65 percent to 80 percent full, gently mixed by inverting five or six times, and placed immediately at 4°C for up to 30 minutes before spinning in a centrifuge. Failure to mix may result in the sample clotting.<sup>3-5 SST tubes should be at least 50 percent to 80 percent full and should be mixed. SST tubes must be allowed to clot at 4°C for 30 minutes to 60 minutes before centrifuging.4
In conclusion, grey top tubes processed appropriately are the tube of choice for blood glucose analysis, but SST tubes can also be used with reproducible results.
- Li G, Cabanero M, Wang Z et al.: Comparison of glucose determinations on blood samples collected in three types of tubes. http://www.ncbi.nlm.nih.gov/pubmed/23884222. Accessed May 1, 2015.
- Mandal A. Blood sugar glucose measurement. http://www.news-medical.net/health/Blood-Sugar-Glucose-Measurement.aspx. Accessed May 1, 2015.
- Mayo Medical Laboratories. http://www.mayomedicallaboratories.com/customer-service/faq/specimen/collection-tubes. Accessed May 1, 2015.
- Blood Collection Tubes – Molecular Diagnostic Services. http://www.mds-usa.com/collection.html. Accessed May 1, 2015.
- Austin Community College. http://www.austincc.edu/kotrla/phb_gray. Accessed May 1, 2015.
MLO’s “Tips from the clinical experts” column provides practical, up-to-date solutions to readers’ technical and clinical issues from experts in various fields. Readers may send questions to firstname.lastname@example.org.