Printer Friendly

When duplicate testing leads to different results.

When duplicate testing leads to different results

You are walking down the hospital corridor to the coffee shop when you see a group of attending physicians standing in the hall. Suddenly an eminent gastroenterologist rushes up to you and loudly announces, "I've finally been able to prove that your lab is lousing up the test results"--an outburst heard by every physician within a 20-foot radius.

You invite him to your coffee to discuss the problem. He replies, "Can't; too busy. But let me tell you that I drew duplicate specimens from one of my patients and sent one to your laboratory and one to a referral lab. The results didn't begin to agree. Why can't you straighten out your laboratory so we can depend on it?" With that, he stomps off down the hall.

Does this scene sound familiar? Unfortunately, it happens all the time. Rarely do physicians stop in to discuss the problem. Even if they did, no matter what you said, their answer would be the same: "You got result A. The referral lab got result B. They are obviously correct."

* On the wrong track. Although drawing duplicate specimens for the purpose of comparative studies sounds like a legitimate test for accuracy, that's precisely why it's so deceptive. It is extremely difficult to prepare stable, reproducible specimens that will give comparable results after shipping. Anyone familiar with national survey programs with national survey programs or companies that sell reference or control specimens will testify to that fact. (1-4)

What should you say to the annoyed physician? You can't simply tell him he's wrong. There is a logical way to show him why.

* Choose your weapon. You can use several approaches to defend your lab's integrity. Let's assume that your hospital laboratory is next to his office. Although the referral lab provides rapid pickup, it is located some distance away. If the actual condition is reversed--the referral lab is closer to the doctor--the time factor will be different, but not much else.

Ask yourself the following 10 questions. Use them as a guide to identify where the variance occurred. Some of the answers will provide evidence for you to take to your distrustful physician.

1. Was it our error? Start with the obvious: the chance that your lab made a mistake. Was the specimen mislabeled? Was it handled in such a way that results became unacceptable? To investigate the possibilities, talk with your staff and find out how the specimen was treated. If no gross clerical errors are found, quality control charts show no deviation, reference standards are correct, and control samples give consistent results, your lab is probably not the problem.

2. With whom are we being compared? As stated earlier, a referral laboratory did the duplicate test. Although these institutions usually like to be called reference laboratories, most are not. A reference lab resembles the National Bureau of Standards in that its staff members are acknowledged experts in a certain area. Most send-in laboratories are considered referral labs. These large labs attract a high volume of specimens for unusual tests.

The staffs of some referral labs have unique skills in defined areas and can perform certain specialized procedures, such as hormone analysis, that entitle them to be considered reference labs in those categories alone. Other referral labs have no more expertise than your own, and some have less.

If the referral laboratory is a POL, it may be well run under rigid quality control or managed so poorly that its results are grossly inaccurate. Having full knowledge about local labs will allow you to determine with whom your laboratory is being compared.

3. Did the physician get the facts straight? Almost all of us have confused two patients at least once. If confronted, the physician may admit that he had actually confused one set of results for another. Such a confession is likely to be made, if at all, in your office rather than in front of the executive committee. By asking for the patient's name and the date the specimen was tested, you may be able to check the hospital chart yourself. (To save face, the physician may add: "But your lab has been making plenty of other mistakes takes recently.")

4. Were the methodologies the same? Most laboratory tests can be performed by several methods; the choice depends on each institution's preferred techniques and instruments. Results from different ones may not be identical. Glance at the reference intervals reported by each laboratory to obtain a clue about whether the methods used differed in any way.

5. Were the instruments and methods calibrated by the same acceptable standards? Even if both laboratories used the same methodologies, the final values may have varied somewhat if different standards were employed.

6. How much variance was there between the two methods? Even when the same specimen is runtwice on the same instrument, a small variance may occur that no instrumental analysis will explain. Yet the tests in question were performed in two different environments. Therefore, see whether the two values are within the acceptable tolerance of the method.

Perhaps the variance, although representing different numbers, has no clinical significance and can be ignored. [5] Amazingly, many physicians expect results to be identical and express concern when they are not.

7. How were the specimens drawn? You may be tempted to assume that evacuated tubes were used for specimen containers, the same amount of anticoagulant was used for all, and the specimens were drawn serially. But were they? Was the first tube drawn at the beginning of the venipuncture and the second as an afterthought after three or four others had been drawn?

Consider this possibility: The initial specimen was drawn. The results were so startling to the clinician that he had a second specimen drawn and sent to the referral lab. Results are bound to vary with the passage of time.

Leaving the tourniquet in place for three minutes or longer during the phlebotomy causes many standard test results to vary. [6] The position of the patient (supine, sitting, or standing) or any movement or exercise can also induce values to differ. [7] Patients' values fluctuate on different days and even at various times of day. [8-10]

Were the specimens drawn from a venipuncture placed close to an intravenous line in place to administer fluids or medications? This happens now and then and is such an obvious contraindication that it may not have been considered. Further, if the speed of administration changed between the drawing of the two specimens, results will not be consistent--although both will still be wrong.

8. Were the specimens drawn in the same type of container and with the same anticoagulant? Although though it is almost inconceivable that someone would draw a specimen in two different anticoagulants and expect the values to be identical, it could happen.

Some tests done with certain instruments on specimens that are lipemic, hemolyzed, or jaundiced can be inaccurate. [11] If the technologist performing the work doesn't recognize this and fails to report this discrepancy, the results will seem not to agree for other reasons.

9. How were the specimens handled? Was one sent to the laboratory immediately while the other was carried in a coat pocket, where it was kept warm? Or was one put in a mailbox overnight, perhaps to freeze, or left until the next morning's mail pickup at the nurse's station?

Was the test so unusual that the specimen had to be mailed to a specialty lab? If so, was it sent by overnight delivery or regular mail, causing a delay of three to four days? If sent by air, was it on a weekday or weekend? Some transportation services do not deliver specimens on the weekend, but store Friday's batch until Monday. A specimen could conceivably spen d the entire weekend in a refrigerator or at room temperature, depending on how it was labeled and the routine of the laboratory in question.

10. Were the specimens taken from a normal patient? If the patient was ill, the blood specimen may have continued too many white cells, too few red cells, or some other abnormality that can cause rapid changes during the time it was stored.

* Think before you speak. Assess all these factors and discuss any questions with the physician. Only then can it be said that the specimens were parallel and one of the sets of results was inaccurate.

Some lab directors have a disquieting tendency to use the term "lab error" to imply acceptable variance, while others consider "error" a synonym for "mistake." Avoid the word. Variances and discrepancies between laboratories are a given; errors are not.

* Where to focus. Once you have isolated the problem and evaluated the magnitude of the discrepancy, switch the discussion to your laboratory's performance on external proficiency testing. Is the lab inspected regularly? If so, by whom? Stress that your lab is meeting the highest requirements of professional, state, and Federal organizations. Is the referral lab inspected and certified in the same manner? If not--if that lab consists of a nurse working in a POL, for example--such factors as quality control, proficiency testing, and experienced technologists may be absent. Tell the physician about it.

By using these 10 questions in your search for the missing link, you may avoid nasty confrontations and confirm your laboratory's accuracy to any member of the clinical staff who questions it.

[1] Grannis, G.F. Studies on the reliability of constituent target values established in a large interlaboratory survey. Clin. Chem. 22: 1027-1036, 1976.

[2] Gilbert, R.K. The accuracy of clinical laboratory studies by comparison with definitive methods. am. J. Clin. Pathol. 70: 450-470, 1978.

[3] Gilbert, R.K. A comparison of participant mean values of duplicate specimens in the CAP chemistry survey program. Am. J. Clin. Pathol. 66: 184-192, 1976.

[4] Ross, J.W., and Fraser, M.D. The effect of analyte and analyte concentration upon precision estimates in clinical chemistry. Am. J. Clin. Pathol. 66: 193-202, 1976.

[5] Grannis, G.F.: Gruemer, H.D.; Lott, J.A.: et al. Proficiency evaluation of clinical chemistry laboratories. Clin. Chem. 18: 222-236, 1970.

[6] Statland, B.E.; Bokeland, H.; and Winkel, P. Factors contributing to intra-individual variation of serum constituents: IV. Effect of posture and tourniquet application on variation of serum constituents in healthy subjects. Clin. Chem. 20: 1513-1519, 1974.

[7] Statland, B.E.; Winkel , P.; and bokeland, H. Factors contributing to intra-individual variation of serum constituents: II. Effect of exercise and diet on variation of serum constituents in healthy subjects. Clin. Chem. 19: 1380-1383, 1973.

[8] Statland, B.E.; Winkel, P.; and Bokeland, H. Factors contributing to intra-individual variation of serum constituents: I. Within-day variation of serum constituents in healthy subjects. Clin. Chem. 19: 1374-1379, 1973.

[9] Winkel, P.: Statland, B.E.; and Bokeland, H. Factors contributing to intra-individual variation of serum constituents: V. Short term day-to-day and within-hour variation of serum constituents in healthy subjects. Clin. Chem. 20: 1520-1527. 1974

[10] Bokeland, H.; Winkel, P.; and Statland, B.E. Factors contributing to intra-individual variation of serum constituents: III. Use of randomized duplicates to evaluate sources of analytical error. Clin. Chem. 20: 1507-1512, 1974.

[11] Peters, R.W.; Myhre, B.A.; Ellis, R.R.: et al. Unnecessary albumin therapy in a jaundiced patient. Transfusion 14:61-62, 1974.

Dr. Myhre and Dr. Fu are professors and Dr. Louderback is clinical assistant professor of pathology, University of California School of Medicine, Los Angeles. Dr. Mhyre is chief of clinical pathology and Dr. Fu is head of biochemistry at Harbor-UCLA Medical Center, Torrance, Calif.
COPYRIGHT 1990 Nelson Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1990 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Myhre, Byron A.; Fu, Paul; Louderback, Allan
Publication:Medical Laboratory Observer
Date:Oct 1, 1990
Previous Article:Dealing with know-it-alls - even if you are one yourself.
Next Article:Development of a peripheral smear differential review.

Related Articles
Problem solving and decision making with proficiency data; analysis of proficiency surveys can help a lab troubleshoot snags in its procedures.
A single-instrument approach to TDM toxicology.
Selecting instruments for the Stat lab.
Use of a spreadsheet for method evaluation statistics.
Tracking hematology QC with a concise visual display.
Tips on evaluating and monitoring quality.
Duplicate orders.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters