Calibration Information and Courses from MediaLab, Inc.

Several manufacturers offer semi-automated instruments (dipstick readers) for reading reagent strips. Use of an instrument removes the subjectivity of visually interpreting color changes on reagent strips, and assures that tests will be read at the correct time. Transcription errors will also be avoided if the instrument is interfaced with the laboratory information system. The technology employed is based on the principle of reflectance, with the amount of light reflected being inversely related to the concentration of substances present. An example of reflectance is the light which is scattered after light strikes an unpolished surface. Since each component on the dipstick produces a different color reaction, the light source for each test must be at the appropriate wavelength. This is accomplished either by using filters or monochromatic light sources. The percent reflectance is determined by dividing the test reflectance by the calibration reflectance and multiplying by 100. Algorithms are used to change the results obtained into a linear relationship with concentration of analyte.

Quality control procedures may include the following: monitoring variables within the laboratory, like water quality, glassware calibration, and instrument calibration, that may affect results defining protocols for performing quality control for each test procedure participating in proficiency testing programs graphing quality control results performing daily and periodic analysis of quality control data and charts troubleshooting quality control errors documenting quality control errors, including the steps taken to resolve the problem and verification of success performing routine maintenance of laboratory instruments keeping records

To ensure reliability of results, many other factors besides controls are monitored. These include: water quality analytic balance calibration glassware calibration centrifuge calibration thermometer calibration electric power stability heating bath temperatures refrigerator temperatures freezer temperatures expiration of reagents, standards, and controls instrument maintenance procedure manuals, including written step-by-step procedures of all tests performed by laboratory method selection, based on local population and clinician needs normal range verification based on the local population technical competencies of laboratory staff members

Technologists are primarily responsible for performing testing and reporting results. Other duties include:Performing only those tests authorized by the director and for which the technologist is licensed by specialty.Following the laboratory's procedure for specimen handling, running tests, reporting results, and maintaining recordsParticipating in proficiency testing and demonstrating that proficiency samples are tested in the same manner as patient samplesFollowing quality control and instrument calibration policiesDocumenting corrective action taken when results exceed the laboratory's acceptable performance valuesUsing professional judgment to ensure test validity, including recollecting and retesting samples that may be flawed or contaminated

Standards are also used to determine the linearity of the testing instrument. This is done by plotting a calibration or standard curve. Most testing instruments must be operated within a linear range. The Clinical and Laboratory Standards Institute or NCCLS defines linearity as “the measure of the degree to which a curve approximates a straight line.The examples to the right show linearity because a change along the x-axis shows a corresponding change along the y-axis, whether the x-value is low or high.

Linear calibration curves are more desirable because they result in the best accuracy and precision. Some testing methods, however, do have nonlinear calibration curves. When that is the case, more calibration standards are needed to achieve desirable precision. Regardless, linearity is not to be used as a tool for calibration verification, accuracy assessment, or establishing a reportable range.

David S. Plaut further refines the definition. He states that linearity is “a line that does not change slope between the high point and next lowest one and passes through zero.” The formulation of a calibration curve will vary according to the procedure, but it should be based on at least three known values—more if necessary.

A shift, on the other hand, describes a sudden change in the data mean that persists throughout further days of testing. Shifts can be caused for any number of reasons, including persistent instrument malfunction, loss of calibration, or improper reagents. Shifts are often easier to recognize than trends, as the data may be more clearly suspect. Again, shifts must be corrected as soon as possible, as patient data are most likely already invalid. In the example to the right, the large jump on day 5 followed by persistently high control values means that a shift has occurred.