current status of verification practices in clinical biochemistry in spain

DOI 10.1515/cclm-2012-0659 Clin Chem Lab Med 2013; 51(9): 1739–1746

Rubén Gómez-Rioja*, Virtudes Álvarez, Montserrat Ventura, M. Jesús Alsina, Núria Barba, Mariano Cortés, María Antonia Llopis, Cecilia Martínez and Mercè Ibarz

Current status of verification practices in clinical biochemistry in Spain

Abstract

Background: Verification uses logical algorithms to detect potential errors before laboratory results are released to the clinician. Even though verification is one of the main processes in all laboratories, there is a lack of standardi-zation mainly in the algorithms used and the criteria and verification limits applied. A survey in clinical laborato-ries in Spain was conducted in order to assess the verifica-tion process, particularly the use of autoverification.Methods: Questionnaires were sent to the laboratories involved in the External Quality Assurance Program organized by the Spanish Society of Clinical Biochemis-try and Molecular Pathology. Seven common biochemi-cal parameters were included (glucose, cholesterol, triglycerides, creatinine, potassium, calcium, and alanine aminotransferase).Results: Completed questionnaires were received from 85 laboratories. Nearly all the laboratories reported using the following seven verification criteria: internal quality control, instrument warnings, sample deterioration, ref-erence limits, clinical data, concordance between param-eters, and verification of results. The use of all verification criteria varied according to the type of verification (auto-matic, technical, or medical). Verification limits for these parameters are similar to biological reference ranges. Delta Check was used in 24% of laboratories. Most labo-ratories (64%) reported using autoverification systems. Autoverification use was related to laboratory size, own-ership, and type of laboratory information system, but amount of use (percentage of test autoverified) was not related to laboratory size.Conclusions: A total of 36% of Spanish laboratories do not use autoverification, despite the general implementation of laboratory information systems, most of them, with autoverification ability. Criteria and rules for seven rou-tine biochemical tests were obtained.

Keywords: autoverification; clinical authorization; posta-nalytical phase; results verification.

*Corresponding author: Rubén Gómez-Rioja, University Hospital La Paz, Paseo de la Castellana 261, 28046 Madrid, Spain, E-mail: [email protected] Álvarez: Clinical Laboratory L’Hospitalet (ICS), L’Hospitalet de Llobregat, Barcelona, SpainM. Jesús Alsina and María Antonia Llopis: Clinical Laboratory Barcelones Nord i Vallés Oriental (ICS), Badalona, Barcelona, SpainNúria Barba: Laboratory Catlab, Viladecavalls, Barcelona, SpainMariano Cortés and Cecilia Martínez: University Hospital Santa Creu i Sant Pau, Barcelona, SpainMercè Ibarz: University Hospital Arnau de Vilanova (ICS), Lleida, SpainRubén Gómez-Rioja, Virtudes Álvarez, Montserrat Ventura, M. Jesús Alsina, Núria Barba, Mariano Cortés, María Antonia Llopis, Cecilia Martínez, and Mercè Ibarz: Commission on Quality Assurance in the Preanalytical phase of the Spanish Society of Clinical Biochemistry (SEQC)

IntroductionVerification is one of the main processes in laboratory medicine. It involves different professionals in multiple steps of analytical and postanalytical phases. Labora-tory test results are checked for intrinsic consistency and congruence related to available patient and sample information (clinical and demographic data, as well as quality indicators from analytical and extra-analytical processes).

Verification of clinical laboratory reports is useful for error detection and also for improving the information provided to clinicians, as written comments can be added to the final report, or additional tests may be carried out when appropriate. Although comments should only be written by medical staff, many of the verification pro-cesses can be done by technical staff. For this reason, a distinction is usually made between technical verifica-tion, which involves the analytical and preanalytical processes (quality control, instrument warnings, etc.), and medical verification, which requires pathophysi-ological expertise. Most laboratory information systems (LIS) differentiate between these two processes. Techni-cal verification mainly seeks incongruence of the results with the aim to differentiate between preanalytical errors

Brought to you by | Purdue University LibrariesAuthenticated | 128.210.126.199Download Date | 9/30/13 3:29 PM

1740 Gómez-Rioja et al.: Verification practices in clinical biochemistry in Spain

and unexpected biological variations. Medical verifica-tion provides clinical decisions about health condition, therapeutic effects or prognostic information, resulting in interpretative comments.

A part of this process can be automated using com-puter algorithms to detect potential errors and to select complex cases that require special review of laboratory results before they are released to the clinician (autoverifi-cation). Preanalytical, analytical, and postanalytical data can all be used to estimate the probability of a specific result in a specific patient. This process can be associated with automatic verification procedures (repetition of the same test under the same or different conditions or carry-ing out other tests that increase the discriminating power of the algorithm) and automatic authorization (release of the report with or without specific comments, or report retention pending manual revision). Automation of this process ensures standardization and improves laboratory efficiency: “Automation of verification, validation and delivery is a fundamental step for a timely, appropriate and reliable release of laboratory reports”. [1].

Even though some of these autoverification systems have existed for more than 20 years, published data on the use of these systems in laboratories are limited. However, one such study was a survey carried out by researchers in the United Kingdom to assess adherence to the rec-ommendations of the Royal College of Pathologists on clinical authorization and reporting of results [2]. Results from that survey, published in 2003, showed that British pathologists reviewed an average of 30% of test results at their laboratories (range, 5%–100%). The American College of Pathologists (CAP) published a checklist that indicates what items should be included in the autoveri-fication process. This allows the accreditation of these processes in accordance with the Clinical Laboratory Improvement Amendments (CLIA) regulations. The CLIA emphasizes the need to document the rules used and to validate these rules before implementation and periodi-cally thereafter [3]. In 2006, the Clinical and Laboratory Standards Institute (CLSI) published the AUTO-10 guide with specific recommendations for the development and implementation of autoverification systems in laborato-ries [4]. This guide defines the autoverification process as “the automated actions performed by a computer system related to the release of test results to the medical record using criteria and logic established, documented, and tested by the medical staff of the laboratory”. The UNE-ISO 15189 guide [5] makes no direct reference to autoverifica-tion as a laboratory tool, but does include the requirement that computer systems used for processing results must be validated prior to use and must also guarantee data

confidentiality and security. This document (“Recom-mendations for the protection of laboratory information systems”) states that the algorithms used for calculations using patient data should be revised periodically and all results should be compared to a predefined range in order to detect incongruous results before the report is released to clinicians. Despite the existence of these guidelines and the 20-year history of autoverification systems, there is a lack of standardization mainly in the algorithms used and the criteria and verification limits applied.

The use of autoverification systems is usually related to laboratory workloads, but other factors, such as degree of mechanization and computerization of the laboratory, availability of medical staff, and the type of patients of the laboratory, probably play a role in how verification is performed.

Given this information gap, our institution, the Spanish Society of Clinical Biochemistry and Molecular Pathology (SEQC), decided to carry out a survey to deter-mine how the verification process is currently practiced in clinical laboratories in Spain, with a focus on autoveri-fication systems, and its degree of harmonization with the additional aim of performing an intercomparison program among Spanish laboratories in the future.

Materials and methodsWe created a 13-question survey divided into three distinct sec-tions (see Supplementary Data, which accompanies the article at http://www.degruyter.com/view/j/cclm.2013.51.issue-9/issue-files/cclm.2013.51.issue-9.xml). Questionnaires were sent to 646 labo-ratories participating in the External Quality Assurance Program organized by the SEQC. The questionnaires were sent by e-mail to the laboratories, thereby allowing them to complete and return the survey electronically.

The survey was sent together with a technical note that pro-vided definitions for the terms used in the survey (see Supplemen-tary Data). It is important to note that the common term of the veri-fication process in Spain is “validation”. For this reason, the terms “verification” and “validation” were defined as synonymous in the survey. However, the correct term as defined by ISO must be “verifi-cation: confirmation, through provision of objective evidence, that specified requirements have been fulfilled”, and for this reason we have chosen to use this term in this paper.

The first section of the questionnaire requested general informa-tion about the laboratory, the number of determinations performed per year for each of the seven biochemical parameters (glucose, cho-lesterol, triglycerides, creatinine, potassium, calcium, and alanine aminotransferase), setting (primary care or hospital), laboratory ownership (public or private), the LIS used, and type of verification performed.

The following two questions were included to classify laborato-ries based on its verification techniques:


http://www.degruyter.com/view/j/cclm.2013.51.issue-9/issue-files/cclm.2013.51.issue-9.xml

http://www.degruyter.com/view/j/cclm.2013.51.issue-9/issue-files/cclm.2013.51.issue-9.xml

Gómez-Rioja et al.: Verification practices in clinical biochemistry in Spain 1741

1. Are some results automatically verificated through the LIS? (yes or no)

2. Which of the following statements best defines your laboratory activity?

a. All tests are verificated by medical staff.b. Some tests are automatically verificated through the LIS and the

rest are verificated by medical staff.c. Some tests are verificated by technical staff and the rest are

verificated by medical staff (autoverification is not used).d. Some tests are automatically verificated through the LIS, others

by technical staff, and other by medical staff.

These questions allowed us to establish four exclusive groups according to the verification approach:Laboratories not using autoverification:

– With only medical verification (a). – With technical and medical verification (c).

Laboratories using autoverification: – Together with manual medical verification (b). – Together with manual technical and medical verification (d).

Both questions were found to have internal consistency, except in one case in which a laboratory reported use of autoverification but only manual verification. The response of that laboratory was there-fore excluded from the evaluation.

Laboratories were asked if they used common criteria for each type of verification (technical verification, medical verification, or autoverification). The questionnaire included the following list of the most typical criteria:a. Verification of fulfillment of quality control requirements

established by the laboratory for each analytical run.b. Verification of the absence of instrument warnings.c. Verification of the absence of signs of sample deterioration

(hemolysis, blood clots, etc.).d. Comparison of results with limits established by the laboratory

based on the reference range, critical limits, or other verification limits.

e. Comparison of results with previous results from the same patient (Delta Check).

f. Comparison of mean results of patients between analytical runs.g. Interpretation of the result based on clinical, demographic, or

requester data.h. Concordance between clinical parameters belonging to the

same request.i. Verification of results (repetition of the test in the same or

different instrument, with or without dilution).j. Other (free text).

The second section of the survey was related to the specific verifi-cation criteria [limits of automatic or manual verification (for adult males) and the maximum acceptable difference with respect to previous results (Delta Check)] for each of the seven biochemical parameters. Limits of manual verification are those limits applied when laboratories are not using automatic verification. Additional questions in this section were included to assess the use of sample deterioration criteria and the patients’ clinical data for verifica-tion. A final question was included about laboratory policy regard-ing the generation of additional testing in cases with unusual test results.

For all of the criteria above, laboratories were asked to indicate if it was possible to evaluate these issues through the verification screen of the LIS in order for us to assess the feasibility of using these criteria for verification.

The third section of questions focused mainly on the use of au-toverification in the laboratory, and included questions on the per-centage of biochemistry results automatically verificated, and also if explanatory comments were generated. Participants were also asked to provide the name of the commercial autoverification program used (if any) in addition to the LIS. They were also asked about how the autoverification rules were validated.

Statistical analysisThe results were evaluated by grouping the laboratories accord-ing to their characteristics: size, setting, ownership, and type of verification. The χ2-test, using the SPSS 15.0 Statistical Software Package (SPSS, Inc., Chicago, IL, USA), was used to study the re-lationship between the laboratory characteristics and the verifica-tion procedures.

ResultsOut of 646 laboratories surveyed, 85 completed the ques-tionnaire (13.2% response rate). Of those 85 laboratories, 54 (64%) reported using autoverification systems. The most common verification procedure was autoverifica-tion, followed by medical verification. The least common approach was mixed manual verification (technical and medical) (Table 1).

A significant correlation was observed between the use of autoverification and laboratory size, ownership and type of LIS. Autoverification was used by 91% of large laboratories (those analyzing > 200,000 creatinines/year) versus only 38% of small laboratories (laboratories ana-lyzing < 20,000 creatinines/year) (p = 0.002) (Figure 1). Autoverification is more common in public laboratories than in private laboratories (77% vs. 37%, respectively) (p = 0.002) and in laboratories that use the OMEGA (Roche Diagnostics) LIS (90%) versus those using a different type of LIS (50%) (p = 0.001).

Nearly all (98%) laboratories reported using the fol-lowing 7 of the 11 general verification criteria listed on the questionnaire at least in one stage of the verifica-tion process (automatic, technical, or medical): inter-nal quality control, instrument warnings, deterioration of the sample, reference limits, clinical data, concord-ance between parameters, and confirmation of results (repeated test). The use of the other four criteria (limits of critical results notification, Delta Check, other verifica-tion limits, or patient mean test values) was less common



(89%, 64%, 49%, and 22%, respectively) (Figure 2A). No significant differences in the use of these criteria regard-ing classification criteria (the use of autoverification, size, setting, or ownership) were found. Although the survey allowed respondents to list other criteria used for verifica-tion, no further criteria were reported.

Figure 2B shows the percentage of use of all verifica-tion criteria for each type of verification method (auto-matic, technical, or medical). For automatic verification, the most commonly used criteria were reference limits, critical limits, and confirmation of results. For technical verification, the most common criteria were revision of

38%

57%

86%91%

64%

0%

25%

50%

75%

100%

Small Medium Big Very big Total

Figure 1 Percentage of laboratories using autoverification accord-ing to the size of the laboratory.

Table 1 Characteristics of participant laboratories.

Scope of care (setting) Primary+hospital setting (65.5%) Hospital setting (23.5%) Primary (10.6%)Ownership Public (69.4%) Private (22.4%) Public+private (7.1%)Size (creatinines/year) Small < 20,000 (24.7%) Medium 20,000–100,000 (37.7%) Big 100,000–200,000 (24.7%) Very big > 200,000 (12.9%)Laboratory information system OMEGA (Roche Diagnostics) (35.3%) MODULAB (Izasa) (22.4%) Other (42.3%)Middleware Yes (51.8%) No (48.2%)Type of verification Medical verification without autoverification (22.4%) Medical+technical verification without autoverification (14.1%) Autoverification+medical verification (42.4%) Autoverification+medical+technical verification (21.2%)

sample deterioration, instrument warnings, and quality control. Finally, for medical verification, clinical data of patients, overall agreement of results, and verification of results were the most common criteria.

Table 2 shows the limits for automatic or manual verification for all of the seven biochemical param-eters, expressed as median and interquartile ranges. In all cases, these median values of automatic verification limits were equal to, or slightly larger than, those used in manual verification, although similar to reference limits in both cases.

Twenty laboratories (24%) reported data on Delta Check utilization in verification. Table 3 provides the number of laboratories using Delta Check for each parameter, the limit used, expressed as median and inter-quartile range, and the acceptable period of time elapsed since the last determination. Four of the laboratories (5%) indicated that Delta Check limits were based on the refer-ence change value (RCV) of the laboratory. Median values in these four laboratories were similar to those observed in the whole group except for triglycerides and alanine aminotransferase, parameters that have bigger biological variability.

The survey included a question about the use of infor-mation provided by requesting physicians on the test requisition form for the seven biochemical parameters included in the study. Of the responding laboratories, 84% reported taking into account diagnostic information, 88% used demographic data, and 81% considered the requesting department.

Many laboratories have developed protocols to gener-ate new tests based on the results obtained. For glucose and potassium, 11% of laboratories reported adding new tests and 40% of laboratories did so for creatinine (Table 4).

The information presented on the verification screen of the LIS varies between laboratories (Table 5). Data for quality control and instrument warnings were easily avail-able in < 50% of laboratories, despite its application on authorization processes.

Laboratories were also asked to estimate (in quar-tiles) the percentage of all tests that were autoverificated. In 31 cases, this percentage was 0% because only human verification was used. In the other 54 laboratories, the amount of autoverificated tests was evenly distributed, with reported autoverification rates as follows: < 25% of the tests (9 laboratories; 17%); between 25% and 50% (15 laboratories; 28%); between 50% and 75% (15 laborato-ries; 28%), and more than 75% (11 laboratories; 20%). No statistical association between the percentage of autover-ificated tests and laboratory size was observed.



100.0% 100.0% 100.0% 97.6% 89.4%

49.4%63.5%

22.4%

97.6% 97.6% 97.6%

0%20%40%60%80%

100%A

BAt least in one type of verification

0%

20%

40%

60%

80%

100%

Autoverification

Quality

contr

ol

Instru

ment w

arning

s

Deterio

ration

Referen

ce lim

its

Critica

l limits

Verific

ation

limits

Delta c

heck

Patien

ts mea

ns

Clinica

l data

Agreem

ent

Confirm

ation

Quality

contr

ol

Instru

ment w

arning

s

Deterio

ration

Referen

ce lim

its

Critica

l limits

Verific

ation

limits

Delta c

heck

Patien

ts mea

ns

Clinica

l data

Agreem

ent

Confirm

ation

Technical validation Medical validation

Figure 2 Relationship between types of verification and the criteria used. (A) Percentage of laboratories using all criteria at least in one verification type. (B) Percentage of laboratories using all criteria for every verification type.

Table 2 Verification limits.

Parameter Units Autoverification limits Manual verification limits

n Low High n Low High

Glucose mmol/L 56 3.6 (2.8–3.9) 6.9 (6.1–10.4) 27 3.6 (3.6–3.9) 6.1 (6.1–11.1)Cholesterol mmol/L 49 2.6 (1.8–3.1) 6.2 (5.2–7.5) 26 2.8 (1.6–3.1) 5.7 (5.2–6.7)Triglyceride mmol/L 50 0.4 (0.3–0.6) 2.3 (1.8–3.4) 26 0.5 (0.2–0.5) 2.3 (1.7–3.7)Creatinine μmol/L 53 44 (35–53) 115 (106–140) 27 53 (27–54) 115 (111–177)Potassium mmol/L 56 3.5 (3.03–3.5) 5.3 (5.1–5.5) 26 3.5 (3.0–3.55) 5.1 (5.0–5.85)Calcium mmol/L 55 2.05 (1.90–2.13) 2.63 (2.60–2.75) 26 2.05 (1.90–2.13) 2.63 (2.58–2.75)ALT IU/L 50 4 (2–5) 53 (40–53) 24 5 (0–6.5) 68.5 (41–100)

ALT, alanine aminotransferase; n, number of laboratories reporting validation limits; low, median and interquartile range of the low limits; high, median and interquartile range of the high limits.

Most (70%) laboratories that use autoverification per-formed an initial validation while 42% reported carrying out an annual check of the system. None of the labora-tories reported use of a commercial verification system; instead, all laboratories used their LIS. Approximately three-quarters (74%) of laboratories incorporate interpre-tative comments during the authorization of results.

DiscussionTo our knowledge the value and therefore the necessity of verification of laboratory results has not been the subject of any trial, but all laboratories consider verification a basic process for its activity, as part of the interpretation of results and especially as a method to detect preanalytical



The high workload in current laboratories and the need to standardize verification processes have prompted many laboratories to implement computerized verification systems. Autoverification systems allow a reliable, fast, and standardized review, selecting cases that may be assessed individually. Although several commercial verification pro-grams have been shown to have good sensitivity and the ability to markedly improve consistency of the process [7–9], none are widely used in Spain, where the LIS commonly incorporate autoverification tools. Despite the widespread use of autoverification systems, no interlaboratory compar-isons have been carried out to assess differences between criteria used in verification rules. A first step on the way to introduce specifications must be to detail the state of the art of the verification process. Surveys like this one can help to understand how verification is done in laboratories.

Autovalidation systems have been shown to improve laboratory productivity [6, 10] and may also improve the quality of results. According to the CLSI guide [4], the use of these systems increases the consistency of the verification process, decreases overall processing time and the need for manual intervention, and allows the use of increas-ingly complex verification algorithms, and can provide guidance to develop protocols for technical staff. The CLSI guide lists a series of requirements that all autoverification systems must meet. These guidelines are useful for com-panies interested in designing an autoverification system because they emphasize the importance of complying with the essential aspects of laboratory reports: accuracy, reli-ability, confidentiality, and flexibility in providing results based on needs (i.e., an ability to handle urgent requests). Nevertheless, it does not provide information on the crite-ria to be applied for specific parameters.

Table 3 Delta Check use.

Parameter Delta Check based on laboratory criteria

Delta Check = RCV

n %Delta Days n %Delta

Glucose 18 16 (10–20) 30 (8–90) 4 16 (16–19)Cholesterol 20 15 (10–20) 65 (26–135) 4 16 (15–18)Triglycerides 19 20 (5–75) 90 (3–999) 4 58 (19–72)Creatinine 19 15 (10–20) 30 (4.5–90) 4 16 (15–18)Potassium 17 12 (12–15) 30 (4–90) 4 14 (13–15)Calcium 15 8 (8–10) 30 (4–90) 4 7 (6–8)ALT 19 25 (10–68) 30 (6–135) 4 71 (67–75)

ALT, alanine aminotransferase; n, number of laboratories using Delta Check; %Delta, Delta Check value in percentage, expressed as median and interquartile range; days, acceptable period of time elapsed since the last determination; RCV, reference change value.

Table 4 Addition of protocolized tests according to the results obtained.

Parameter % Laboratories adding tests

Added tests (in order of frequency)

Glucose 11 Glycosylated hemoglobinUrinary glucose

Cholesterol 35 HDL cholesterolLDL cholesterolTriglyceridesApolipoprotein

Triglycerides 26 LDL cholesterolCreatinine 40 Urea

Estimated glomerular filtration rateSodium/potassium

Potassium 11 SodiumCalcium 25 Protein/albumin

PhosphorusParathormone

ALT 32 ASTGGT/ALP/bilirubinSerology

AST, aspartate aminotransferase; GGT, γ-glutamyl transpeptidase; ALT, alanine aminotransferase; ALP, alkaline phosphatase.

Table 5 Available information in the screen during authorization.

% Laboratories

Reference values 95Previous results 94Complete results of the analysis 89Information from test request form 79Repetitions 69Critical values 61Preanalytical errors 57Instrument warning flags 44Electronic medical records 35Quality control 28Delta Check 22Other limits 19Patients’ mean values 12

Percentage of laboratories having this information available in the authorization screen.

errors that are not detected by internal or external quality control [6]. Most preanalytical errors occur outside the laboratory, but if they are undetected can seriously affect patient safety. An identification error gives metrologically perfect results with dangerous consequences.

Within the total quality process, verification is a product quality control, similar to the final test of manu-factured goods. Although systems of quality assurance are designed to improve long-term quality, verification can help to find coarse preanalytical errors, such as identifica-tion errors, or deteriorated specimens.



In our survey, approximately 36% of Spanish labo-ratories do not use an automatic verification tool, even though it might be available in their information systems (all laboratories that responded to this survey reported having a LIS with autoverification tools). The laborato-ries (64%) participating in the survey that reported using autoverification systems were primarily laboratories with high workloads. This percentage is similar to the infor-mation reported in the United Kingdom [2]. Although we found a relation between laboratory size and autoverifica-tion implementation, we failed to find a relation between the size of the laboratory and the amount of workload that was autoverificated. The reason why autoverification is more common in users of OMEGA SIL could be due to its increased use in large laboratories, but no statistical asso-ciation was found (p = 0.13).

The fact that nearly all (98%) laboratories included the same seven verification criteria (of the 11 listed on the questionnaire) indicates a high degree of control of this process in Spanish laboratories. However, it should be noted that 11% of these laboratories do not review critical values, one of the essential criteria [11], due to the impor-tant clinical repercussion that these results could have, and only 49% of laboratories declare using specific verifi-cation limits. Biological reference intervals can be useful for pathophysiological interpretation but are impractical to locate preanalytical errors resulting in outliers, either in the sense of life-threatening (critical warning) or statisti-cally unexpected results.

The median values of the limits established for automatic or manual verification for each of the seven biochemical parameters are very similar to biological refer ence limits. This means at least a 5% rejection based on the autoverification rule for each magnitude, depend-ing on the laboratory setting, mainly primary care in our survey. The use of other verification criteria (Delta Check, instrument warnings, etc.) could modulate auto-verification degree. It seems evident that results should be checked if pathological results deviate significantly from the patient’s clinical profile. However, experience has shown that the number of revisions may be excessive and generate unnecessary repeated analyses in laboratories with a high workload. For instance, in a recent survey of over 25,000 retests selected by decision rules, the initial result was considered incorrect in only 2.6% of cases [12].

The use of Delta Check in autoverification is less common than the practice of reviewing results that fall outside the reference range, even though Delta Check may actually be more sensitive in detecting preanalytical errors as well as analytical and postanalytical errors [13]. Delta Check seems to be a good tool to control these three

processes. Another benefit of Delta Check is that signifi-cant changes in patients’ results can be detected (assum-ing that the RCV is used as the limit in automatic verifi-cation). However, very few laboratories in our study used RCVs as Delta Check limits. We found a large range in the period of time used to measure Delta Check among labora-tories, probably due to lack of specific recommendations and to the absence of scientific evidence. More scientific data are needed to standardize verification limits and Delta Check values.

The use of verification criteria as a function of pro-fessional status seems fairly coherent given the various professional categories in Spanish laboratories. Technical staff (graduates of non-university vocational education programs) are responsible for reviewing internal quality control results, instrumentation warnings, and sample deterioration, whereas medical staff (most of whom have completed a 4-year specialized accredited training after obtaining a degree in health sciences) are responsible for verifying the consistency of results according to the patient’s clinical data.

A large percentage (between 81% and 88%) of labora-tories considered the patient’s clinical and/or demographic data, as well as the clinical department requesting the tests. This is possible because most laboratories are highly computerized in Spain and thus these data are easy acces-sible. Nonetheless, one problem that arises during medical verification is that it is not always possible to access all necessary information. As a result, the laboratory cannot fully control the process. For example, although 100% of laboratories reported the use of information on quality control during the verification process, only 28% reported finding this information “easily accessible” on the LIS. The same happens for Delta Check or instrument warnings. This could limit efficiency on the verification process.

The use of algorithms or protocols for performing additional tests based on the results varies depending on the parameter studied. Between 11% and 40% of the laboratories generate new tests. However, because the possibility of carrying out such actions depends on the versatility of the LIS, this is more common in large laboratories than in small ones. Reflex testing could help laboratories to reduce the number of tests neces-sary and to improve efficacy on critical value reports [14, 15].

A high percentage of laboratories added comments to the results during the medical verification process, although it should be noted that manual clinical authori-zation is not always necessary, as some comments can be automatically incorporated through the autoverification process.



As our results have shown, many laboratories make an effort to assure the quality of these autoverification tools, with nearly three-quarters performing an initial validation and 42% performing an annual audit. Like-wise, responding laboratories expressed a strong interest in participating in a comparative study, with 88% report-ing that an intercomparison program would be useful and 93% willing to participate.

Limitations of this study could be the selection of target laboratories, this is because only laboratories par-ticipating in the SEQC External Quality Control Program were invited to reply. There is probably a quality bias in the participating laboratories, most of which are highly interested in participating in an intercomparison study. Moreover, the response rate was low (13%), similar to other voluntary surveys conducted by our Society. Nevertheless,

the results are consistent and we believe that the results obtained could be extrapolated to other European coun-tries. Inclusion of verification and interpretation in exter-nal quality assurance programs could contribute to the harmonization of criteria used in this process.

Conflict of interest statement

Authors’ conflict of interest disclosure: The authors stated that there are no conflicts of interest regarding the publication of this article.Research funding: None declared.Employment or leadership: None declared.Honorarium: None declared.

Received September 30, 2012; accepted February 28, 2013; previ-ously published online April 11, 2013

References1. Plebani M, Lippi G. Improving the post-analytical phase. Clin

Chem Lab Med 2010;48:435–6.2. Prinsloo PJ, Gray TA. A survey of laboratory practice in the

clinical authorization and reporting of results. Ann Clin Biochem 2003;40:149–55.

3. Miller J. Postanalytical data review. In: Ward-Cook K, Lehmann G, Schoeff L, Williams R, editors. Clinical diagnostic technology. The whole testing process, vol. 3: the postanalytical phase. Washington: AACC Press, 2006;37–54.

4. Autoverification of clinical laboratory test results; approved guideline. CLSI document AUTO10-A (ISBN 1-56238-620-4). Wayne, PA: Clinical and Laboratory Standards Institute, 2006.

5. ISO 15189:2007. Medical laboratories – particular requirements for quality and competence. Geneva: ISO, 2007.

6. Torke N, Boral L, Nguyen T, Perri A, Chakrin A. Process improvement and operational efficiency autoverification through test result. Clin Chem 2005;51:2406–8.

7. De Smet D, Nanos N, Martens GA, Smet L, Damiaens S, Van Moer G, et al. Automated validation of clinical laboratory test results: do we get what we expect? Clin Chem Lab Med 2009;47:379–80.

8. Guidi GC, Poli G, Bassi A, Giobelli L, Benetollo PP, Lippi G. Development and implementation of an automatic system for verification, validation and delivery of laboratory test results. Clin Chem Lab Med 2009;47:1355–60.

9. Oosterhuis WP, Ulenkate HJ, Goldschmidt MH. Evaluation of Lab Respond, a new automated validation system for clinical laboratory test results. Clin Chem 2000; 46:1811–7.

10. Pearlman ES, Bilello L, Stauffer J, Kamarinos A, Miele R, Wolfert MS. Implications of autoverification for the clinical laboratory. Clin Leadersh Manag Rev 2002;16:237–9.

11. Crolla LJ, Westgard JO. Evaluation of rule-based autoverification protocols. Clin Leadersh Manag Rev 2003;17:268–72.

12. Deetz C, Nolan D, Scott M. An examination of the usefulness of repeat testing practices in a large hospital clinical chemistry laboratory. J Clin Pathol 2012;137:20–5.

13. Risk management techniques to identify and control laboratory error sources; approved guideline – 2nd ed. CLSI document EP18-A2 (ISBN 1-56238-712-X). Wayne, PA: Clinical and Laboratory Standards Institute, 2009.

14. Barron J, Ng C, Aspin L, Robinson LJ, Smith G. Reflex testing to define action limits for community-based requests. Ann Clin Biochem 2012;49:337–40.

15. Wu AH, Ghani F, Prigent F, Petry C, Armstrong G, Graff L. Reflex testing II: evaluation of an algorithm for use of cardiac markers in the assessment of emergency department patients with chest pain. Clin Chim Acta 1999;288:97–109.


DOI 10.1515/cclm-2012-0659Ad Clin Chem Lab Med 2013; 51(9): Ad1–Ad4

Supplemental data

Quality assurance and accreditation of laboratories committeecommission on quality assurance in the preanalytical phase

Survey on the state of the validation process in the laboratories

Definitions

Validation: verification that the specified requirements are adequate for an intended use (VIM: International Vocabulary of Metrology)Verification: provision of objective evidence that a given item fulfills specified requirements (VIM).In practical terms, validation and verification are consid-ered synonymous.Autovalidation or autoverification: automatic verification of results performed by the laboratory information system (LIS), if the results fulfill a set of rules specified in the computer system.

Technical verification: part of the authorization process performed by the technical staff of the laboratory (gradu-ates of non-university vocational education programs).Medical verification: part of the authorization process performed by the medical staff of the laboratory (people that have completed a 4-year specialized accredited train-ing after obtaining a degree in health sciences).Delta Check: rule of automatic validation based on the acceptable difference between two consecutive results of the same patient.Reference value of change (VRC): Significant difference between two consecutive results, which can be biologi-cally relevant.VRC = 2.77(CVa2+CVi)2)1/2

Participant code:Name:Email:

1 Characteristics of the laboratory1.1 -Activity.

Magnitudes in serum No. determinations/year

GlucoseCholesterolTriglyceridesCreatininePotassiumCalciumAlanine aminotransferase (ALT)


Ad2 Gómez-Rioja et al.: Verification practices in clinical biochemistry in Spain

1.2 -Setting:Primary careHospitalMixed

1.3 -Ownership:PublicPrivateMixed

1.4 Laboratory Information System (LIS). Indicate name and provider.

1.5. Intermediate computer system between the analyzers and the laboratory computer system, involved in the valida-tion results (Middleware) (e.g., PSM, CentraLink…). Indicate name and provider.

2. Are some results automatically verificated through the LIS?YesNot

3. Which of the following statements best defines your laboratory activity?1. All tests are verificated by medical staff.2. Some tests are automatically verificated through the LIS and the rest are validated by medical staff.3. Some tests are validated by technical staff and the rest are verificated by medical staff (autoverification is not

used).4. Some tests are automatically verificated through the LIS, others by technical staff, and other by medical

staff.

Participant code: .SUE

4. Criteria used for verification in each case. (Please, indicate “YES/NO/not applicable” in the boxes.) Regard-ing autoverification, it refers that the computer system verifies or applies the criterion automatically as part of the automatic verification rule. In the case of technical or medical verification, refers that the person authorizing personally performs the action.

Criteria Technical verification

Autoverification Medical verification

a. Verification of fulfillment of quality control requirements established by the laboratory for each analytical run.

b. Verification of the absence of instrument warnings.c. Verification of the absence of signs of sample deterioration (hemolysis, blood clots, …).d. Comparison of results with limits established by the laboratory based on: Reference range Critical limits Other validation limitse. Comparison of results with previous results from the same patient (Delta Check).f. Comparison of mean results of patients between analytical runs.g. Interpretation of the result based on clinical, demographic or requester data.h. Concordance between clinical parameters belonging to the same request.i. Verification of results (repetition of the test in the same or different instrument, with or

without dilution).j. Other: indicate which


Gómez-Rioja et al.: Verification practices in clinical biochemistry in Spain Ad3


5. Limits established in your laboratory for each verification criteria and for each of these magnitudes. (If not applicable, please leave empty).

Magnitude Glucose Cholesterol Triglycerides Creatinine Potassium Calcium ALT

a. Units:b. Limits of autoverification. Limits applied by the LIS outside which does not perform automatic validation. If this limit depends on sex or

age, please indicate the limit used for adult, male population.LowHighc. Limits of manual verification. If you are not using automatic verification, indicate the limits used for medical verification. If this limit

depends on sex or age, please indicate the limit used for adult, male population.LowHighd. Maximum acceptable difference with previous results (Delta Check)PercentageAbsolute valueDoes the Delta Check value correspond to the reference change value (RCV) in your lab? Please, indicate Yes or Noe. Do you apply the calculation of Delta Check only if the time elapsed since the last request is below an established limit? If yes, please

indicate the limit applied for each of the previous magnitudesTime (days)f. Do you notify interferences based on serum index? If yes, please indicate the cut-off limit for interference. To be able to compare limits

between participants, please indicate the correspondence of the index reported by the analyzer to mg/dL of hemoglobin, triglycerides and bilirubin, in each case

Hemolysis (mg/dL serum hemoglobin)Lipemia (mg/dL triglycerides)Icterus (mg/dL bilirubin)


Magnitude Glucose Cholesterol Triglycerides Creatinine Potassium Calcium ALT

g. Is the information provided in the request form used during the verification process? Diagnosis or treatment Demographic data (age, sex, or others) Petitioner datah. Do you generate complementary tests based on the results of some of these magnitudes? Please Indicate which Added tests

6. Which data are available or easily accessible on the LIS verification screen (please tick the appropriate box)

Yes Not

a. Visualization of whole results of the requested tests (biochemistry, hematology, microbiology, etc.)b. Status of the quality control for the corresponding analytical seriesc. Instrument warnings generated for the resultd. Outcome of repetitions/dilutionse. Information on preanalytical incidentsf. Indication of the limits of verification for Reference Critical


Ad4 Gómez-Rioja et al.: Verification practices in clinical biochemistry in Spain

Other validation limitsg. Delta Checkh. Possibility of reviewing previous analyses of the patienti. Mean results by analytical seriesj. Clinical information Available in the request Access to the patient’s medical history

7. Which percentage of biochemistry results are autoverificated in your laboratory?0% < 25%25%–50%50%–75% > 75%

8. Do you generate interpretative comments during verification?YesNot

9. Are you using some commercial autoverification program? (Valab or similar)Yes (please, indicate name and provider)No

10. If autoverification is incorporated in the LIS of your laboratory, did you perform an initial validation of its performance?YesNo

11. If autoverification is incorporated in the LIS of your laboratory, do you perform a periodic audit of its performance?YesNo

12. Do you think that it would be useful to have an intercomparison program for the automatic verification system?YesNo

13. If yes, would you be interested in participating in the pilot program?YesNo

Thank you for yours truly, collaboration


current status of verification practices in clinical biochemistry in spain

Documents