WS/T 616-2018Autoverification of Clinical Laboratory Quantitative Test Results (English PDF)

4.1 Data source

4.1.1 Before inspection

4.1.1.1 Patient Information

The patient's descriptive information can be used as a judgment parameter in the operation of the program, including age, gender, inspection department, clinical diagnosis, Medication, unique patient identification code, etc.

4.1.1.2 Sample Information

Sample related information may include, but is not limited to.

a) sample type;

b) acquisition time;

c) the collection site;

d) receiving time;

e) sample traits (eg hemolysis, lipemia, jaundice, presence or absence of clots, etc.);

f) Others (such as remarks filled in at the time of application, such as before dialysis, after dialysis, etc.).

4.1.2 Inspection

The object of the program operation can also be derived from information related to the detection system status, including.

a) calibration status;

b) indoor quality control;

c) instrument alarm information;

d) reagent related information (such as reagent validity period, open bottle stability period);

e) Methods Analyze performance related information and others (eg analytical measurement range, sample dilution factor, interference, etc.).

4.1.3 After inspection

The program can also use data from the following sources.

a) the result warning symbol (for example, a warning symbol indicating that the biochemical reaction substrate is exhausted);

b) the previous measurement result of the same test item of the same patient (for the difference check);

c) the results of other test items in the sample (for correlation or logical relationship analysis between different items, such as LDL cholesterol the sum of alcohol and high-density lipoprotein cholesterol should not be greater than the total cholesterol), or the test results of the same sample of the patient (eg the urinary albumin is abnormally abnormal, and its total urinary protein should not be normal;

d) Results of different patients on the same test instrument (eg for calculating the moving average of certain items).

4.1.4 Data selection

The laboratory can select the appropriate data for the analysis of the automated review process based on the workflow of the medical institution.

4.2 Data Analysis

4.2.1 Analysis of patient information

The program combines the test results with patient information (age, gender, inspection department, diagnosis, etc.) for analysis.

4.2.2 Analysis of sample information

The program should be able to identify samples that are not in compliance with the requirements for collection and inspection. For fully automated inspection systems with automated sample preparation, the sequence should be able to identify abnormal sample traits that affect the test results (eg, by determining serum index or image judgment system to identify hemolysis, lipemia, yellow specimen).

4.2.3 Detection system status analysis

The program should be able to identify various types of warning symbols transmitted or generated by the inspection system, LIS, middleware, etc. that are relevant to the accuracy of the test results (eg the result is beyond the warning range of analysis and measurement range, interference, and quality control out of control).

4.2.4 Numerical comparison

Numerical comparisons should include, but are not limited to, the following.

a) The program compares the test results with the set range. Commonly used ranges include biological reference intervals, analytical measurement ranges, and reportable Scope, critical value, medical decision level, etc.

b) The laboratory may also use other methods to customize the range for comparison (eg, based on percentiles of patient data distribution).

When the laboratory is customizing the scope of comparison, it should fully assess the medical risks based on the characteristics of the medical institution and the level of medical technology.

Ask the clinician for advice.

c) Different comparison ranges can be set for different populations, such as physical examination population and disease population, outpatient and inpatient.

The scope of comparison of the same project may also vary from patient to department, such as serum creatinine from general surgery and nephrology samples.

The circumference can be set separately.

d) Data comparison should be able to identify impossible results (eg alanine aminotransferase results are negative or non-numeric).

Note. The program should pay special attention to the identification of critical values.

4.2.5 Difference check

The difference check includes but is not limited to the following aspects.

a) The automatic audit procedure should have a difference check function to identify specimen confusion between different patients, manual input data errors, instruments Analyze problems in the process, etc.

b) The laboratory shall set the time interval and acceptance criteria for the difference comparison for the selected project. Time interval is used to control how long the program is Historical results within the range are searched and compared. Acceptance criteria can be set to percent deviation and/or absolute deviation.

c) The laboratory can analyze the changes in the patient's historical results and combine the clinical experience to set the acceptance range of the difference comparison in the application process.

The necessary adjustments can be made.

4.2.6 Logical relationship and correlation analysis

Compare the results of different projects in a certain way, and the comparison results should meet the logical requirements (for example, total bilirubin should not be less than straight with bilirubin, a certain type of white blood cell count should not exceed the total number of white blood cells); the test results can also be combined with clinical diagnosis for compliance Analysis (eg, whether serum creatinine levels in patients with chronic renal insufficiency are consistent with diagnosis); associations between projects can also be analyzed (eg Relationship between serum creatinine levels and urea, major anions and cations, whole blood hemoglobin concentration and red blood cell count). Laboratory can Define logical relationships and association analysis rules based on their own characteristics. See Appendix A for commonly used logical relationships and correlation analysis.

4.2.7 Consistency Analysis

The procedure should be able to compare the actual report items with the medical application items and identify minor, multiple, and incorrect items.

4.3 Flow chart drawing

The flow chart is used to describe the steps and functions of the automatic audit program to perform data analysis. The flow chart should reflect as much as possible the impact of the test results Identification of problems before, during and after the test related to reliability and reliability may be problematic. The flow chart design can be simple or complex. automatic See Appendix B for an example of the review process design.

4.4 Report issuance

4.4.1 Automatic issuance

When the result of the automatic auditing process meets all the preset rules, the report is directly issued by the LIS, and no manual intervention is performed. By Reports issued by the dynamic review process should have an easily identifiable mark. The laboratory shall have provisions specifying how to determine the automatic issuance of inspection reports.

Inspector.

4.4.2 Manual issuance

When the automatic review program judges that the result does not meet the preset rule, the program marks the sample, and the report will be retained and manually Necessary information check, sample trait check, retest, dilution, etc., and then contacted with clinical staff (such as emergency reports, Unqualified specimen report, etc.). The automatic audit procedure should be able to record the reasons for the failed audit, or further prompt manual retesting, dilution, etc.

operating.

The contents and format of the automatically issued and manually issued inspection reports shall comply with the laboratory's requirements for inspection reports.

5.1 General

Verification of the automated audit process to verify that its performance meets the requirements of the laboratory audit test report is a must for the application of the automated audit process Prerequisites.

5.2 Personnel requirements

It is recommended that the intermediate and above titles with rich clinical laboratory experience be verified by those who have worked in relevant fields for at least 5 years.

Work to encourage testing physicians and clinicians to participate in the validation process.

5.3 Verification content and method

5.3.1 Pre-implementation verification

Regardless of whether you develop or introduce software with an automated review process, the lab should respond to the program's functionality, before the formal implementation of the automated review.

The parameters and rules are verified to confirm that they meet the requirements of the laboratory. The content and requirements of the verification shall include but are not limited to.

a) All functions, rules and parameters involved in the automatic audit process should be verified to ensure that the performance of the program is consistent with the laboratory Requirements for review of results;

Example 1.

The program requires that the items with unqualified quality control cannot pass the automatic review. If the first indoor quality control of the serum total protein is out of control and not corrected, all the transmissions are the total protein results of the information system were not automatically reviewed.

Example 2.

If the program requires the identification of an instrument-specific alarm symbol, the data with the alarm symbol should be transmitted to the information system to confirm that it is recognized by the program.

Example 3.

If the procedure sets the serum total bilirubin below the direct bilirubin and cannot pass the audit, then the eligible data should be transmitted to the system to confirm that it cannot pass.

Review.

Example 4.

If the serum albumin level is set to 25g/L to 50g/L in the program, it can be verified that the results in this range can be approved, and higher than or Results below this range have not been reviewed.

b) each test item and each sample type involved in the automatic audit process should be verified;

c) that each item in the verification inspection report is subject to automatic determination;

d) there should be no more or less reports on the automatic audit report;

e) The verification procedure identifies which item in the report has not passed the automatic review and its reasons;

f) The recommended verification time is not less than 3 months and/or the number of reports is not less than 50,000.

Note. If there is no data that can trigger the automatic audit procedure rule within the above verification period or the number of report cases, it can be measured by setting the simulation data.

Try to identify if the program can meet the relevant requirements. If no direct bilirubin is greater than total bilirubin during the verification process, a simulation can be included.

A report of the situation to check if the automated review process identifies such an error report.

5.3.2 Verification if necessary

In the process of using the automatic auditing program, if equipment update, automatic audit parameter change, information system upgrade, etc. may occur, Changes to the automated audit function should be verified to ensure compliance is met before continuing to use. Verified content, number of reports, and time limit it can be determined by the laboratory based on the changes.

5.3.3 Periodic verification

The designated personnel will review the reports that have passed the automatic review, and the results of the review should be consistent with the results of the automatic review. Regular verification week is recommended the period is 1 year, the verification time is not less than 10 working days and/or the number of reports is not less than 5000.

5.4 Verify the correct rate requirement

The correct rate of verification should be 100%. If the correct rate does not reach 100%, the procedure should be modified, the parameters adjusted, or Rules and verify the problem.

6.1 General

The laboratory should periodically review the automated review process to ensure that its functionality continues to meet the requirements. If during the application automatic review process when the test report and the clinical non-conformity (which can be derived from clinical complaints and consultation) increase, the review should be initiated. If identified during the review process the limitations of the automatic audit procedure or parameters and rules shall be modified and reconfirmed/verified as appropriate.

6.2 Personnel requirements

The management of the department should plan and supervise the assessment, and designate personnel with more than 5 years of relevant professional work experience to complete, encourage the inspection physician and clinicians participate in the review of the automated review process.

6.3 Review content

6.3.1 Periodic review

The laboratory should regularly review the automatic approval pass rate, and analyze the reasons for not passing the assessment, evaluate its impact on the diagnosis and treatment, and take it when it fails.

The necessity and effectiveness of the measures can be adjusted by adjusting parameters and regulations on the premise of ensuring that the performance of the test and analysis meets the requirements and guarantees medical safety.

The program is optimized to increase the automatic approval pass rate. It is recommended to review it once a year.

6.3.2 Review as necessary

When an automatic audit report is found in a clinical complaint or consultation, the relationship between it and the automatic audit procedure and parameters and rules should be analyzed.

Do not medical safety hazards, modify and re-verify if necessary.

7.1 Establishing a system

The laboratory shall establish corresponding procedures and systems, and stipulate the procedures for establishing, verifying, reviewing, and authorizing the automatic review procedures, and the corresponding responsibilities and authorities.

7.2 Licensing

The formal implementation of an automated audit should be authorized by the laboratory manager. The laboratory manager can authorize the relevant personnel to the automatic review process After verification, a report proving the reliability of the automatic review procedure shall be submitted, and the automatic audit will be enabled after the manager signs and confirms.

7.3 Emergency procedures

Laboratory managers should confirm the existence and enforceability of emergency procedures. The automatic audit procedure should have a similar switch function to ensure that it is necessary the automatic review feature can be turned off at any time. In the case that the automatic review process is enabled, the specified filter can be forced according to the preset filter conditions.

Some or some samples do not perform an automatic review. When the automatic review process is deactivated, sufficient personnel should be ensured to complete the inspection report review.

Appendix A

(informative appendix) Test report review common logical relationship and correlation analysis example Examples of common logical relationships and correlation analysis for inspection report review are shown in Table A.1.

Table A.1 Example of Common Logical Relationships and Association Analysis for Inspection Report Audit

Serial number content 1 total cholesterol should be greater than the sum of low-density lipoprotein cholesterol and high-density lipoprotein cholesterol 2 total bilirubin should be greater than direct bilirubin 3 total protein should be greater than albumin 4 total prostate specific antigen should be greater than free prostate specific antigen

5 Creatine kinase total activity should be greater than creatine kinase isoenzyme MB activity

6 The total number of white blood cells should be greater than a certain classification of white blood cell counts

7 The sum of the percentages of white blood cell classification should be 100%

8 The ratio of hemoglobin concentration to red blood cell count exceeds a certain range, prompting manual review

9 The ratio of serum creatinine and urea exceeds a certain range, prompting manual review.

10 The difference between the main anion and the cation (anion gap) exceeds a certain range, prompting manual review Thyroxine (T4), free T4, triiodothyronine (T3), free T3 and thyroid stimulating hormone or manually below the upper and lower limits of the biological reference interval Appendix B (informative appendix) Automated review process design example an example of an automated review process design is shown in Figure B.1.

Figure B.1 Example of automatic audit process design

references

[1] Narayan T, Leonard B, Tracy N, et al. Process improvement and operational efficiency through test result Autoverification [J]. Clinical Chemistry, 2005, 51(12). 2406-2408.

[2] Clinical and Laboratory Standard Institute. Autoverification of clinical laboratory test results. Approved Guideline [R].CLSI document AUTO10-A.Wayne, PA. CLSI, 2006.

[3] Guidi GC, Poli G, Bassi A, et al. Development and implementation of an automatic system for verification, Validation and delivery of laboratory test results [J]. Clin Chem Lab Med, 2009,47(11). 1355-1360.

[4] Shih MC, Chang HM, Tien N, et al. Building and Validating an Autoverification System in the Clinical Chemistry Laboratory[J].Laboratory Medicine, 2011,42(11). 668-673.

[5] Fraser CG.Making better use of differences in serial laboratory results[J].Ann Clin Biochem, 2012,49(Pt

1). 1-3.

[6] Jones JB.A strategic informatics approach to autoverification [J]. Clin Lab Med, 2013, 33(1). 161-181.

[7] Krasowski MD, Davis SR, Drees D, et al. Autoverification in a core clinical chemistry laboratory at an Academic medical center[J].J Pathol Inform, 2014,5. 13.