Interviews

The 2012 update of the biologic variation database was its most popular yet. It gathered far more downloads and sparked more discussion than previous editions. We put some of those user questions to the group of scientists who compiled the database.  Dr. Maria-Carmen Perich Alsina, PhD, President of the Analytical Quality Commission of the Spanish Society of Clinical Chemistry, and her colleagues provided some answers.

Questions on the Biologic Variation Database and Biologic Goals

Interview by Sten Westgard, MS
July 2012

Question: We would be interested to hear your view on what set of requirements a laboratory should use (desirable, optimal, minimum, minimal, etc) to evaluate their methods. And if a laboratory achieves the minimum, but not the desirable, how should they change how they handle test results?

Answer: According to the Stockholm hierarchy, laboratories should evaluate their methods using the best criterion possible; thus, if data on BV are known, the corresponding desirable specifications should be used. For analytes with very small BV components, as for example albumin, calcium, chloride, HbA1C, sodium, the minimum specifications (based on BV) could be used for evaluating methods; in the other extreme, for analytes with high BV values, such as triglycerides, enzymes, urine constituents, the optimum specifications should be used because they are easily attained.

Only when no data on BV are available, analytical methods should be evaluated against specifications based on the state of the art; in these cases, it would be better using the best approach possible, for example the 80 or 90 percentile of the peer-method group.

We recommend not using the Consensus Minimum Quality Specifications for method evaluation, but for the individual laboratory to correct its poor performance.

Question: If a laboratory can achieve the optimal, how should they do things differently than a lab that can only achieve desirable specifications?.

For analytes where the optimal specification is reached, the internal QC rule could be less stringent than for other analytes, and efforts of laboratory staff can be focused on those analytes that have difficulty to reach minimum specifications

Question: What should labs do when they find that methods on the marketplace do not meet even the minimum quality specification? I know for some tests, this is true.

First of all, laboratories should assure that their procedures are working in a correct way: they have to use written SOPs, they have to plan internal QC and participate in EQA programs obtaining good results, they have to follow a training protocol, and so on. In these cases, if they still can not attain the minimum quality specification, they have to press providers to produce better analytical systems. Also, EQA organizers have to investigate results stratified by methods, so to highlight those with poorer precision and accuracy.

Then, if the laboratory wishes to use QC Design (Sigma-metrics), should they restrict themselves to using just the desirable or optimal specifications?.

We believe that, of course, when data on BV are available, labs should using the desirable or optimum specifications; if it is not possible, they could use the minimum specification.

Question: When looking at method imprecision and method bias, is it acceptable to have a method with both the maximum imprecision and the maximum bias listed in the table?.

No, if one of the two components of analytical error is the maximum, the other should be reduced. When both components are of high order, the level of quality necessary to satisfy medical need is not fulfilled.

Question: I am wondering if you have any anecdotes from colleagues, etc., on how laboratories are really using the database. On Westgard Web, we can track the number of visits to the pages, for example, and in the last 3 years, the main database page has been visited more than 165,000 times. I have seen a few papers where they discuss using the database as a way to set quality requirements. But do you hear more about how labs use the database?

Because of the strong demand, we have permanently opened an itinerant course dealing with internal QC, quality specifications and practical applications of the BV database, which is performed in different Spanish places (2 or 3 times per year, since the last three years). Also, by the SEQC-EQA feedback system, laboratories make questions on internal QC planning, how to proceed when BV based specifications are surpassed, etc. We can realize that labs progressively stop using control limits written in the commercial pamphlets and learn on QC design.

Also, when a new IVD system has to be bought, one of the conditions that are gaining position is that analytical imprecision fulfills the desirable BV.

Moreover, in the Bio-Rad convocations of experts in laboratory quality we have seen many colleagues worldwide actually use our database in their daily routines.

Question: Are manufacturers are paying more attention to the databsae. Does the group have any contact with them? What is your advice to manufacturers?

Although the fact is that in Spain we have few manufacturers but mainly distributors, we have seen that providers contact our group in the Spanish Society of Clinical Biochemistry and Molecular Pathology (SEQC) to review what information should be added to the devices. Concepts such as commutability of calibrators and control materials, traceability to standard materials or methods, quality specifications for imprecision and bias are being incorporated when possible.

Our thanks again to Dr. Perich and her fellow scientists involved with the biologic variation database.