Editor's note: This month, MLO's clinical expert is Sandra L. Honigfort, BSMT(ASCP). Sandra is Laboratory Services Coordinator and Infection Prevention Coordinator at Paul Oliver Memorial Hospital in Frankfort, MI.
Q
CAP requires that we do instrument/method correlations at least twice a year. It does not specify how many samples to use. What are your recommendations as to the number of samples needed? Thanks.
A
When performing initial correlations, it is felt that 40 samples from across the reportable range is an adequate number of specimens to validate a new procedure. Subsequent correlations are to ensure that nothing has happened to either instrument during the previous six months that has changed the correlation coefficient to make it unacceptable. At this point QC and CAP survey peer results are available to document that the instruments are giving accurate results. For these correlations, then, the recommendation would be to start with 10 specimens throughout the reportable range. If these match the acceptable criteria (an easy and reliable criteria to use is the same criteria used for CAP survey peer comparison), then you are done. If the results do not match the acceptable criteria, troubleshooting the methods to find the problem is required. Once the problem is solved, run another 10 samples. If this does not resolve the problem, outside comparisons (sent to a reference lab) should be done to find which of the instruments is the more accurate, and the other should then be taken out of service until the issue can be resolved. This is especially important for analytes that providers follow closely and watch for small changes to verify therapy response. The main concern is that the difference in results from instrument to instrument should not be an influence on a provider's clinical decisions.
Q
I am trying to resolve an ongoing discussion here in our lab. The discussion concerns whether or not it is good laboratory practice to begin a manual test or procedure and expect another tech to take over in the middle and read/result that test. Some of us believe that this is fine when placing barcoded primary tubes on an instrument that will read, test and send the results to an LIS to be collated or verified. Any other tech can do this. But if one tech starts a manual test (RSV, influenza, mono, HIV, BHCG, urine drug screen, etc.), we think that the same tech should read the results and report them out. Other techs here do not feel that it is a problem to finish an assay that another tech set up, labeled, and inoculated manually. Do you have an opinion or any advice?
A
It may be preferable for techs to finish their own manual testing, but it is probably not feasible in a lot of situations. In today's healthcare climate, it is very important to keep overtime under strict control. Therefore a test may have to be started by one shift and finished by another. This requires that there be definite guidelines for performing manual testing. All devices should have positive identification that connects the device to the patient. Tests that are time-dependent need to have accurate timing devices set, in order for the tech finishing the test to read the result at the correct time. If all of these guidelines are followed without exception, there should not be a concern about one tech setting up a test and another tech finishing the testing. If you consider the case of Microbiology, this is done all the time. Some results go through several techs before the culture is finalized. If there are concerns that certain techs are not reliable to either inoculate a device or to read the result, you have a bigger problem that needs to be investigated!
Q
Can you explain the logic behind working up (doing an ID and sensitivity) all positive blood culture bottles for a patient which are drawn on the same day?
A
A There is no specific “logic”; it is more a matter of tradition. Many laboratories are breaking away from doing complete testing on all isolates for that very reason. In some cases laboratories will do Identification and refer the provider to previous isolates for the sensitivity patterns. Other laboratories will only do a preliminary Identification that indicates it is the same organism (Gram stain and a few immediate tests).
Most of the time blood cultures yield a single organism that is responsible for a patient's septic condition. However, with some of the more seriously immunocompromised patients, multiple organism sepsis has been reported. Therefore it is very important that isolates are carefully screened to make sure that all organisms are identified. Most providers will repeat blood cultures when multiple organisms are isolated in order to rule out contamination.
Recently there has been an increased interest in sepsis, and protocols for diagnosis and treatment are changing. The blood culture is still an important hallmark; however, for severe sepsis it is not a very timely test. Other markers such as lactic acid and venous PH are becoming more rapid diagnostic markers. There are also a number of rapid agglutination tests that are being considered for identifying the source of a septic event. The traditional blood culture may become obsolete as newer, more rapid technologies evolve.
MLO's Tips from the Clinical Experts department provides practical, up-to-date solutions to readers' technical and clinical issues from a panel of experts in various fields.
Readers may send questions to Dan Baer by e-mail at from the clinical experts [email protected].
