Printer Friendly

Tips on evaluating and monitoring quality.

Site visits and well-planned specimen submissions provide insight into the operation of labs to which you're considering sending out tests. Regularly evaluate the labs you decide to use.

When a clinical laboratory sends out a test, the lab director is responsible for making sure the referral laboratory chosen will deliver top-quality work. This article will present suggestions for making the best possible choice in a referral lab and, thereafter, monitoring to assure good performance.

Marketing literature alone, while convenient to consult, can be deceptive. Brochures and advertisements do not always distinguish ideals and goals from actual performance. Colleagues who recommend labs they have used are often the best resources. Nevertheless, their suggestions should be validated before being acted upon.

* Visit the lab. If you can plan to work with a particular referral lab for a long time, to send a large volume of tests there, or to expect it to provide special services for your physician clients, a site visit will be worth the effort. Through direct observation you will obtain a great deal of information unavailable in any other way.

[paragraph] Staff. For a good working relationship with a referral lab, you must have confidence in the people who will be doing the testing. Get it through direct contact with your counterparts during site visits. The director of the primary lab should make such trips, perhaps accompanied by physician clients who are interested in special areas of testing.

Spend time with the referral lab's bench personnel as they perform an entire assay procedure. First, however, do some preparatory reading. A brief study of immunoblotting procedures will enhance your inspection of the immunology section. Are staff members well trained? Does their work reflect a high level of stress? Do they pay proper attention to quality control?

Does the staff include a disproportionate number of laboratory assistants who lack the expertise to make value judgments? Must you fight your way through marketing managers, sales representatives, and product developers to reach the bench staff? Who is responsible for turnaround, and does that person actually help clients or merely pacify them?

Your are looking for specific information: from the assay operator about QC data, from the scientific director about assay design and validation, from the medical director about the referral lab's experience in having its assays used in clinical situations.

[paragraph] Resources. When making a site visit, see if the referral laboratory's backup resources seem able to handle the inevitable breakdowns. Ask whether such a breakdown has occurred and how it was handled.

When you return to your lab, review the results of your site visit with physician clients who will make use of the referral laboratory's special abilities.

[paragraph] Shipping. Are specimens shipped to the referral laboratory at room temperature or frozen? Ask how accuracy is guaranteed in terms of specimen handling, preservative materials used, specimen containers, and packaging specimens for transport. It's imperative that the specimen under scrutiny in the lab be the identical one drawn at venipuncture.

[paragraph] Accuracy. It's up to the referral lab to provide evidence of test accuracy. Is there sufficient proof for each assay? Does the reported value actually represent what it is claimed to represent? For example, has the range of expected values in healthy individuals been contrasted with the range for individuals who have the disease in question? Does the low reported range of thyroid-stimulating hormone (TSH) immunometric assays clearly distinguish between euthyroid and hyperthyroid patients? What clinical correlations have been used to establish this distinction?

The possibility of obtaining false-positive results brings renewed concern whenever a new technology is introduced. For example, a false-positive result in a prenatal screening test for the cystic fibrosis gene performed with polymerase chain reaction (PCR) technology causes unwarranted anxiety in the patient and could even lead to unnecessary termination of a pregnancy.

Are the values obtained for a particular test in the referral laboratory on a par with values you are accustomed to obtaining for the same test in other labs? For cutting-edge tests, widely accepted standards may not have been established yet. In such cases it is essential to discover how the standards used in the referral lab will compare with those your physician clients are accustomed to.

[paragraph] Precision. Is an assay as run by a given operator on a specific day an acceptable one? To answer the question, check the following points. Is a sufficient number of standards used to establish standard curves properly? If duplicates or even triplicates are necessary to place points on the curve, are they being done? Confirm that the standard curve is well evaluated in terms of such constants as slope and initial point. Ascertain that a new curve is generated for each batch assayed.

Does the precision of a determination vary along the full curve? Although the precision for varying points along a standard curve may have been determined at some time during the development of the assay, does the operator presume these precision factors are obtained every time a serum specimen is applied to the curve? Or are specimens run in batches, so that a precision factor is generated for each new specimen? If duplicates are run and the mean is reported, does the operator observe acceptable limits of coincidence between duplicates?

[paragraph] Reproducibility. While it is important to set up a precise assay, even more vital is the assurance that results will be consistent week after week. QC pools--human sera containing previously determined concentrations of the analyte being measured--should be run in consecutive assays to provide data on reproducibility. These data should be provided to the primary laboratory on request.

[paragraph] Sensitivity. Does the assay measure in the range that is useful to elucidate the clinical problem under investigation? Does the test measure a low enough value for its intended purpose? For example, many adrenocorticotropic hormone (ACTH) assays can distinguish very high values from moderately high values but are insufficiently sensitive in the clinical ranges required for patient testing.

[paragraph] Specificity. Evaluate the sophistication of the test design. The best are general screening procedures that incorporate an optimal number of cross-reactions, false positives, and false negatives. Less practical tests consist of exacting series of column chromatographs and extractions that reduce the result to an unquestionable value more useful in research. Serum cortisol, for example, can be determined in several ways: with the fluorometric method, noted for its high blank values; with the competitive protein-binding method, which can, under unusual circumstances, measure steroids other than cortisol; or with the antibody method, which is highly specific. A specialist will know exactly what test to request. A knowledgeable endocrinologist, for example, will request an immunoassay with a specific antibody. The director of a primary lab, on the other hand, will look for a general sense of excellence--a well-run show.

Suggested questions to ask during site visits, derived from a National Committee for Clinical Laboratory Standards document, are listed in Figure I.

* Long-distance evaluation. If a site visit isn't feasible, try to gauge the quality of the referral lab by telephone or mail. Send specimens from patients who have conditions that would lead you to expect certain values. The pregnancy test of a woman known to be pregnant should be positive every time. Testing the referral lab's quality in this way usually involves making an aliquot of one pool and sending samples to the lab variously labeled and under different conditions.

Sending split specimens to several different referral laboratories to compare their results is rarely satisfactory, since you won't be able to resolve any discrepancies from the resulting data alone. A better method is to send one referral lab five aliquots of the same specimen at the same time, labeled as taken from five different pseudonymous patients. The mean value of the specimens should compare favorably with standard values for that assay published in peer review journals.

A highly reliable way to assess reproducibility is to send a referral lab specimens from the same pool on five different (but not necessarily consecutive) days. If you do this, consider the stability of the aliquots. When freezing is required for stability's sake, the freeze-and-thaw sequence can alter values.

The final common denominator in judging referral labs is clinical correlation. A physician client specializing in the area under investigation can provide specimens from patients who have clinical syndromes clearly associated with high, low, or medium values. To test a luteinizing hormone (LH) radioimmunoassay, for example, send serum from a postmenopausal woman. The test should reveal high LH. Serum from a patient whose pituitary gland has been surgically removed should produce low LH values.

Whatever your findings from sending trial specimens, speak with the referral lab staff about them. Include in your discussion the operator who performed the assay, the scientific director, and the medical director of the lab. Since any good lab will want to known when an assay was bad, discuss any poor results identified.

* Periodic reassurance. Once you have selected a referral lab and begin to use it, reevaluate its quality routinely. You can do this in several ways:

[paragraph] Quarterly report. Obtain reproducibility data on the repetitive performance of quality control pools from assay to assay to provide sufficient data for determination of standard deviation and coefficient of variation. Data in the high, medium, and low ranges of each assay should be available on request at least every three months.

[paragraph] Physician feedback. Communicate regularly with at least one local physician who is preeminent in the topic at hand. It's better to correlate results with expectations than to answer complaints later.

[paragraph] Spot checks. Send specimens of a known range to the referral lab. After receiving the results, discuss your findings with your contacts at the referral lab.

* Goals and feedback. Look for a well-trained staff whose work does not reflect a high level of stress. Assays should be precise, provable, and consistent over time. Each test should be specific, measuring the best range for the clinical problem that concerns your physician clients.

The staff of the referral laboratory are your colleagues and should be your partners in problem solving. If you're not entirely happy with what you saw on your site visit or discovered by mailing specimens to their lab, say so. They should eagerly listen to your suggestions for improvement--and then, if possible, make them. Jerald C. Nelson, M.D. The author is professor of medicine and associate professor of pathology, Lorna Linda (Calif.) University School of Medicine, and medical director for endocrinology, Nichols Institute Reference Laboratories, San Juan Capistrano, Calif.
COPYRIGHT 1991 Nelson Publishing
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1991 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Using Referral Labs Efficiently, part 2
Author:Nelson, Jerald C.
Publication:Medical Laboratory Observer
Date:Jul 1, 1991
Previous Article:How ethical dilemmas induce stress.
Next Article:Starting a user group for a chemistry analyzer.

Related Articles
Caution advised in relations with laboratory referral sources.
What ever happened to service?
Congress, HCFA race toward new lab regulation.
When duplicate testing leads to different results.
How to get the best service from a referral laboratory.
Imperative skills for today's lab manager.
Making the grade on the long road to QI.
Automating laboratory send-out tasks.
A quality approach to referral lab contracting.
Choosing a reference laboratory?

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