Feature Story

Be on alert for blood gas traps

The brass tacks of blood gas analyzers

September 2003
Anne Ford

Sure, blood gas laboratories vary, but in the realm of inspection deficiencies, they have quite a bit in common. A survey of about 700 labs last year found five Laboratory Accreditation Program checklist questions on which blood gas labs most often slipped up, with three to five percent of surveyed labs deficient for each question. C. Robert Baisden, MD, Gulf regional commissioner for the CAP Commission on Laboratory Accreditation, reviewed these common deficiencies during an accreditation program audioconference May 14.

Just over five percent of the laboratories surveyed, Dr. Baisden noted, came up short on checklist item BGL.22100: Is there documentation of at least annual review of all policies and procedures in the blood gas laboratory section by the current laboratory director or designee? If not, the lab has some paperwork to catch up on. The remedy lies in reviewing about half of all procedures each month, dating and signing each procedure as reviewed, or implementing a sign-off sheet. The sign-off sheet should list the name and effective date of each procedure. The reviewer’s signature and the date of review must be at the level of each procedure. A single signature on a title page or index of all procedures is not sufficient. If the lab director delegates the review, there should be documentation of the authorized worker. Laboratories can indicate in writing the title, rather than the name, of the appointed reviewer, so they don’t have to make a revision every time personnel change.

“There’s no need to sign the front page of each procedure,” Dr. Baisden said, “although it’s acceptable to do so.” He warned, however, that it’s not uncommon to miss signing and dating a procedure or two using this strategy.

Written modus operandi are the solution as well for labs deficient on question BGL.27485: Does the laboratory have an action protocol when data from imprecision statistics change significantly from previous data? For 4.4 percent of the laboratories surveyed, the answer was no. Developing a written troubleshooting protocol for data that don’t measure up entails consulting the instrument manual or incorporating several steps into the lab’s testing or QC procedure: cleaning and recalibrating the instrument, checking the calibration fluid/gas integrity, tracking possible operator error, and, if necessary, retraining in sampling technique.

A similar number of labs were deficient on question BGL.27250: Is verification of the analytic measurement range (AMR) performed with matrix-appropriate materials of known analyte value appropriate to the AMR of the method system, and is the process documented? Most deficiencies in this area, Dr. Baisden said, arise because laboratories don’t perform AMR verification every six months. When verifying AMR, labs should use matrix-appropriate materials with target values that, at the minimum, are near the low, midpoint, and high values of the AMR. Laboratories should ask the manufacturer to supply material of extreme high and low values. Dr. Baisden added that labs must report values lower or higher than the verified AMR as “greater than” or “less than.”

If the laboratory has more than one method system for performing tests for a given analyte, are they checked against each other at least twice a year for correlation of patient results? Just over four percent of surveyed labs didn’t measure up on this question, BGL.27316. Dr. Baisden says laboratories can perform this correlation by using a combination of patient specimens and QC and linearity materials.

The final checklist item labs commonly missed, BGL.29760, asked: Are at least two levels of quality control specimens analyzed at least every eight hours of operation for all parameters when patient specimens are tested? Dr. Baisden hopes that this deficiency will become less prevalent now that the latest checklist, issued March 31, has adopted the less stringent CLIA standard. This standard requires labs to analyze only one level of QC specimens for pH, pCO2, and pO2 at least every eight hours of operation when patient specimens are tested. Now that the checklist has been modified, Dr. Baisden says, “this should then no longer be a frequent deficiency.”

Anne Ford is CAPTODAY senior editor.