QUALITY ASSURANCE IN THE CLINICAL CHEMISTRY LABORATORY Quality Control * [tis a system of ensuring accuracy and precision in the laboratory by including quality control reagents in every series of measurements. ‘It is a process of ensuring, that analytical results. are. correct. by-testingknown samples that resemble patient samples. It involves the process of monitoring the characteristics of the analytical processes and detects ‘analytical errors during testing, and ultimattely prevent the reporting of inaccurate patient test results + Itis one component of the quality assurance system, and is part of the performance monitoring that occurs after a test has been established. PARAMETERS OF QUALITY CONTROL: 1. Sensitivity. ‘= It is the ability of an analytical method to measure the smallest.concentration of the analyte of interest. 2. Specificity, " _ itis the ability of an analytical rnethod to measure only the analyte of interest. 3. Accuracy snore ‘ Itis the nearness or-closeness of the assayed value to the true or target value. itis estimated using 3 types of studies: recovery, intereference and patient sample comparison. = Recovery study determines how much of the analyte can be identified in the sample; interference study determines if specific compounds affect the laboratory tests like hemolysis, turbidity and icteric; sample comparison study is used to assess presence of error (inaccuracy) in actual patient sample. 4. Precision or reproducibility * It is the ability of an analytical method to give repeated results on the same sample that agree with one another. 5, Practicability * Its the degree by which amethod is easily repeated: 6. Reliability _ "It is the ability of an analytical method to maintain accuracy and precision over an extended period of time during which equipment, reagents and personnel may change. 7. Diagnostic sensitivity ‘+ itis the ability of the analytical method to detect the proportion of individuals with the disease. + Itindicates the ability of the test to generate more true-pasitive results and few false-negative. = Screening tests require high sensitivity so that no case is missed. Sensitivity (%) = 100_x the number of diseased individuals with a positive test Total number of diseased individuals tested 8. Diagnostic specificity It is the ability of the analytical method to detect the proportion of individuals. without-the disease. = _Itreflects the ability of the method to detect true-negatives with very few false-positives. * Confirmatory tests require high specificity to be certain of the diagnosis. 2}Specificity (36) = 100 x the number of individuals without the disease with a negative test Total number of individuals tested without the disease Note: 100% sensitivity and specificity indicate that the test or method detects every patient with the disease and that the test is negative for every patient without the disease. KINDS OF QUALITY CONTROL: 1 Intralab Quality Control (internal QC) * It involves the analyses of control samples together with the patient specimens. "It detects changes in performance between the present operation and the “stable” operation. + tis important forthe dally monitoring of accuracy and precision of analytical methods. Itdetects. both random and systematic errors in a daily basis. = Itallows identification of analytic errors within a one-week cycle. 2.Interlab Quality Control (External QC) "It involves proficiency testing programs that periodically provide samples of unknown concentrations to participating clinical laboratories. "itis important in of the analytical methods. = Its also used to determine state-of-the-art interlaboratory performance. ‘+ The College of American Pathologists (CAP) proficiency program is the gold standard for clinical laboratory external QC testing. ‘Conduct of the External QC Testing: = A series of unknown samples are sent to the laboratory from the reference laboratory or authorized program provider. = Unknown samples must be tested by the laboratorians who regularly perform analysis of patient specimensusing the same reagents and equipment for actual patient specimens, and the results are submitted to the program provider, preferably as soon as every analysis is done. Analysis of the unknown samples should be completed within the usual time as for the routine samples. = Unknown samples should be treated like a patient specimen to determine the true essence of accuracy. ‘= Results of the proficiency testing must not be shared with other laboratories“during the testing period” - comparison studies can be made after the testing cycle to identify areas for improvement. "Some proficiency tests are not quantitative but are qualitative, however for chemistry, it should be quantitative. If there is no available proficiency testing program for a certain analyte, it is required to implement a non-proficiency test scheme. Inten th = Each analyte has a define performance criteria (exemple, +/-3SD for peer mean), where laboratories using the same method are evaluated by comparing them with the group, "In external QC, difference of greater than 25D in the results indicates that a laboratory is not in agreement with the rest of the laboratories included in the program, "If in case a clinical laboratory failed to identify or resolve an error or discrepancy in the test process, the faciltity is at risk of continous operation and may be recommended for closure.Rationale ofthe ExtemaL Oc/Profency Testing: The ultimate goal of proficiency testing is to ensure our accurate. = Proficiency testing allows each laboratory to compare and evaluate test results or outcomes with those laboratories that use the same methods (reagents and equipment). = Data obtained from the proficiency testing can be used to continuously improve test performance, and also serve as troubleshooting guide when investigating analytic error. icians that patient results are Objectives of I 1. To check the stability of the machine. 2. To check the quality of reagents. 3. To check technical (operator) errors. = The accuracy of any assay depends on the control solutions, how they are originally constituted and how they remain stable overtime. = General chemistry assays used 2 levels of control solutions, while immunoassays used 3 levels. "To establish statistical quality control on a new instrument or on new lot numbers of control materials, the different levels of control material must be analyzed for 20 days. * For highly précised assays (with CV less than 1%) such as blood gases, analysis for 5 days is adequate, Control Limits (Control Values): = These are expected values represented by intervals of acceptable values with upper and lower limits. ‘+ If the expected (control) values are within the desired control limits, the clinicians are assured that the test results are accurate and precise. = Control limits are calculated from the mean and standard deviation (SD). "The ideal control/reference limit is between +/-2SD. = Use of a single lot for an extended period allows reliable interpretative criteria to be established Which will permit efficient identification of an assay problem. ‘+ When changing to a new lot number, laboratorians use the newly calculated mean value as the target mean but retain the previous SD value, but when more data are obtained, all values should be averaged to get the best estimates of the mean and SD. = Determination of the mean and SD for the unassayed controls is also advisable because this process improves the performance characteristics of statistical control procedures. ‘Characteristics of an Idea! OC Material: 1. Resernbles human sample. 2. Inexpensive and stable for long periods 3. No communicable diseases. 4. No matrix effects/known matrix effects. ‘5. With known analyte concentrations (assayed control). 6. Convenient packaging for easy dispensing and storage.Notes to Remember: > Quality control materials should resemble human sample and be available for a minimum of one year (same lot number) - different lot numbers of the same material have different concentrations which requires new estimates of the mean and standard deviation. > Human control materials are preferred but because of limited sources and biohazard considerations, bovine control materials are used, > Bovine-based AC material is not the choice for Immunochemistry, dye-binding and certain bilirubin assays, > QC materials should be the same matrix as the specimens being tested, for example, measurement. of glucose in serum should have control solutions that are prepared from serum. > Matrix effects are results of improper product manufacturing, use of unpurified human and nonhuman analyte additives and altered protein components. Control materials can be purchased with or without assayed values. Assayed controls are more expensive but can be used as external checks for accuracy. Reconstitution of lyophilized control materials must be properly done to avoid incorrect control values. > Stabilized frozen controls do not require reconstitution but may have different characterizations compared to actual patient specimens. vv * The test method must be compared always with a method of acceptable accuracy such as the standard reference method (gold standard). + tis recommended by Westgard et al and Clinical Laboratory Improvement Amendments (CLIA) | that 40 to 100 samples be run by each method in duplicate on the same day over 8 to 20 days, ideally within 4 hours, to determine its accuracy and precision. +f only 40 samples will be measured, daily analysis in duplicate of 2 to 5 specimens should be followed for atleast 8 days. = Duplicate analyses of each sample by each method (test method and reference method) are recommended, with the duplicate samples analyzed in different runs and in different order of analysis on the two runs (should be performed within 4 hours). ‘The rationale for performing repeated assays is to detect random errors that affect precision. = After analyses, samples with wide difference should be repeated to rule out technical errors as the source of variation, "The most important characteristic of method evaluation is to determine if the total error (random and systematic error) is less than the allowable error (E,). VARIATIONS: "Are errors encountered in the collection, preparation and measurementof samples, including transcription and releasing of laboratory results. ‘Twpes of Error: 1. Random Error = Itis present in allmeasurements; it isdue to chance. + Itisatype of error which varies from sample to sample. "Its the basis for varying differences between repeated measurements = variations in technique.= itis due to instrument, operator and environmental conditions (variations in techniques) such as pipetting error, mistabeling of samples, temperature fluctuation, and improper mixing of sample and reagent. 2. Systematic Error = Iti an error that influences observations consistently in one direction (constant difference). "It is detected as either positive or negative bias - often related to calbration problems, deterioration of reagents and control materials, improperly made standard solutions, contaminated solutions, unstable and inadequate reagent blanks, leaky ion selective electrode (\SE), failing instrumentation and poorly written procedures. = itis a measure of the agreement between the measured quantity and the true value. Constant Error ~ _ Itrefers to a difference between the target value and the assayed value, ~ itis independent of sample concentration. - it exists when there is a continual difference between the comparative method and the test method regardless of the concentration. b. Proportional/Slope/Percent Error ~ _itresults in greater deviation from the target value due to higher sample concentration, - _itexists when the difference between the test method and the comparative method values is proportional to the analyte concentration. 3.Clerical Error + It is the highest frequency of clerical errors occurs with the use of handwritten labels and ‘request forms. Notes to Remember: Indicators of analytic performance: internal QC, proficiency testing, accreditation, quality assurance monitoring and laboratory utilization > The first step in method evaluation is the precision study which estimates the random error. > To study imprecision or random error, 2 control solutions are run twice a day in a 10- to 20-day period ~ by testing multiple samples on different days, a better assessement of random error over time is achieved. > The total imprecision analysis is the most accurate measure of performance that would affect ‘the laboratory values a clinician might observe and reflects differences such asin the work of technologists, pipetting and temperature fluctuations of the analyzers,Pre-analytical Errors: a Ineamnactnatient identificatios 2. Improper patient preparation 3. Incorrect specimen collection 4. Mislabeled specimen 5. Incorrect order of draw 6. Incorrect used of tubes for blood collection 7. tneorrect anticoagulant to blood ratio (short draw) 8. Improper mixing of blood and anticoagulant 9. Incorrect specimen preservation 10. Mishandled specimen (transport and storage) 11. incorrectly interpreted/ordered laboratory test 12. Incomplete centrifugation 13. Incorrect data logsin Analytical Errors: 1. Incorrect sample and reagent volume 2eincorrect incubation of solution.. 3. Equipment/instrument malfunction 4. Improper calibration of equipment/calibration error 41. Unavailable or delayed laboratory results Notes to Remember: > > > vv Preanalysis refers to all the activities that take place before testing, such as test ordering and sample collection. The most. frequent_preanalytic errors include improperly filling the sample tube, «placing specimens in the wrong containers or preservatives, and selecting the incorrect test. The analysis stage consists of the laboratory activities that actually produce a result, such as running a sample on an automated analyzer. Postanalysis comprises patient reporting and result interpretation. The length of time elapsed between drawing and the separation of serum or plasma from the cells can be a factor in analytical testing. Nonlaboratory personnel were responsible for 29% of the errors with regard to laboratory results. Online computer input is the most error-free means of requesting laboratory tests. In one study, error rates were reported to range from 0.05% to 0.61%, and the distribution of. errors among the testing stages was similar, with most (32%-75%) occurring in the preanalytic stage and far fewer (13%-32%) in the analytic stage (Bonini et al, 2002 cited in Mc Pherson and Pincus 2017).Allowable Error (Es) ‘= It is determined for each test method and is expressed either in measurement units of the analyte (mmol/L) or percentages. It is based on the quantity of error that will negatively affect clinical decisions. = The total error (random, proportional, constant and systematic error) must be less than the Eor fixed limits for 2 method to be considered acceptable. = ifthe total error is greater than E,, corrections must be made to reduce the error or the method be rejected. = The determination of whether long-term precision is sufficient is based on the total i being less than one-third of the allowable error. STATISTICS ‘= It is the science of gathering, analyzing, interpreting and presenting data. 1, Mean — isa measure of central tendency. It is associated with symmetrical or normal distribution Mean = _3%. 2. Standard Deviation (SD)- is a measure of the dispersion of values from the mean. It helps describe the normal curve. A measure of the distribution range. Itis ‘theimost frequently used measure of variation 3. Coefficient of Variation (CV) - is a percentile expression of the mean; an index of precision cee sD) x 100 Mean 4.Nariance- is called the standard deviation squared; a measure of variability. It represents the difference between each value and the average of the data v (spy Terminotoai 1 inferential Statistics - are used to compare the means or standard deviations of two groups of data 2. Fest - is used to determine whether there isa statistically significant difference between the standard deviations of two groups of data 3. Median - is the value of the observation that divides the observations Into two groups, each containing equal number of observations. Its the midpoint of a distribution; 50* centile 4. Mode - is the most frequent observation; it is used to describe data with two centers bimodal). 5. Range ~is the simplest expression of spread or distribution; and lowest score in a data 6. Standard Deviation index (SD!) ~is the difference between the value of a data p value divided by the group's SD 8or od 7. tetest - is used to determine whether there is a statisticaly sig of two groups of data Notes to Remember. > Statistical analyses are used to determine the types and quantity of error that a method has, and to decide whether the test is still valid or unacceptable to make clinical decisions. > Parametric tests: t-test and analysis of variance (ANOVA) > 3 measures of spread or distribution: CV, SD and range > SD describes the distribution ofall values around the mean. > SD and variance represent the average distance from the center of the data (mean) and every value in the data set. > CVallows a laboratorian to compare SDs with different units. > The CV of analyzers described as having reproducible test results can be lower than 1%. > Degree of freedom (n-1) indicates the number of quantities free to vary. > ANOVA is used to analyze precision data to give estimates of the within-In run, between-run and total imprecision. > Method evaluation and statistical analysis are essential, but not sufficient to decide if a test is vai > The acceptable range is 95% confidence limit which is equivalent to * 25D. QUALITY CONTROL CHART ‘= Its used to observe values of control materials over time to determine reliability of the analytical method. "= Its utilized to observe and detect analytic errors such as inaccuracy and imprecision, 1. Gaussian Curve (Bell-Shaped Curve) = Itoccurs when the data set can be accurately described by the SD and the mean. ‘Its obtained by plotting the values from multiple analyses of a sample. = Its a population probability distribution that is symmetric about the mean. = It occurs when data elements are centered around the mean with most elements close to the mean. ‘It focuses on the distribution of errors from the analytical method rather than the values from a healthy or patient population. = The total area under the curve is 1.0 or 100%. 2. Cumulative Sum Graph (CUSUM) = Itcalculates the difference between QC results and the target means. = Common method: V-mask = _Itidentifies consistent bias problems; it requires computer implementation. = This plot will give the earliest indication of systematic errors (trend) and can be used with the 1a, rule, "It is very sensitive to small, persistent errors that commonly occur in the modern, low calibration-frequency analyzer. ‘= Results are out of control when the slope exceeds 45" or a decision (+ 2.7 SD) is exceeded.3. Youden/Twin Plot "It is used to compare results obtained on a high and low control serum from different laboratories. + It displays the results of the analyses by olotting the mean values for one specimen on the ordinate (y-axis) and the other specimen on the abscissa (x-axis). ‘= The points falling from a center but on the 45° line suggest a proportional error, and points falling frorn the center but not on the 45° line suggest a constant error. 4, Shewhart Levey-Jennings Chart "tis the most widely used QC chart in the clinical laboratory. ‘= Itallows the laboratorians to apply multiple rules without the aid of 2 computer. ‘itis a graphic representation of the acceptable limits of variation in the results of an analytical method. * Iteasity identifies random and systematic errors. Errors which can be observed in LJ ch: @. Trend "Its formed by control values that either increase or decrease for six consecutive days. = Main cause: Deterioration of reagents Shift ‘= Itis formed by control valves that distribute themselves on one side or either side of the mean for six consecutive days. = Shiftin the reference range is due to transient instrument differences, "Main cause: improper calibration of the instrument . Outliers = These are control values that are far from the main set of values. * These are highly deviating values. * These are caused by random cr systematic errors. 5. Westgard Control Chart ‘It recognizes that the use of simple upper and lower control limits is not enough to identify analytical problems. In Westgard, error detection rates can increase without increasing the false rejection rate. Westgard used the term control rule to indicate if the analytical process is out of control. 10nt the an Deseription/ Possible Analytical Error Hlustration Tae Teis weed as a rejection or warning rule when fone control result exceeds the mean 2D; for sereening purposes. as Tis observed when one control result exceeds a the mean + 38D; due to randoin error. shee Bae Tt is observed when the last 2 control results (or 2 results from the same min) exceed either is the mean + 2SD; due to aystematic error. es) te “The last four (or any fous) conscoutive contol] —, al results exceed either mean * ISD; due to Sl es fystematc eer. clef Nf ety BEV) | Re "The range or difercnce between the highest = 7 and lowest control result within an analytical = 1 fin exceeds 4! itis due to systematic error. LNAE | Khe SY We ce ek an arava an whe 6 | a consecutive control meanurements fall on one | “RT 7) [ie sid of the mean. ee ca We need to reject an analytical ran when seven control measurements “trend” in the same direction, te, get progressively higher or progressively lower = & We need to reject an analytical ran when & sete al consecutive control measurements fall on one side of the mean, Wo We need to reject an analytical ran when 9 consecutive control measurements fall on one side of the mean. 10. Tes observed when 10 consecutive results are on the same side of the mean; due to systematic error. 12s reject when 12 consecutive control ‘measurements fall on one side of the mean Bova | We need to reject an analytical nin when 2 out of 3 control measurements exceed the same ‘mean plus 25 or mean minus 2s control limit. Se We need to reject an analytical run when 3 consecutive control measurements exceed the same mean plus 1s or mean minus 18 control mit, Notes to Remember. > pa e In measuring systematic error or inaccuracy, Westgard et al recommend that at least 40 samples, and preferably 100 samples be run by comparison-of-methods experiment (test method and reference method). ‘The combinaticn of the control rules used in conjunction with @ control chart has been called the Multirule Shewhart procedure. Multirules establish criteria for deciding whether an analytic process is out of control. ‘The sensitivity of the multirule procedure can be increased to detect smaller systematic errors by increasing the number of observations considered. False rejections can happen because of the controi limits design and not actually identify 2 problem with the method. “Westgard rules" are generally used with 2 or 4 control measurements per run, which means ‘they are appropriate when two different control materials are measured 1 or 2 times per ‘material, which is the case in many chemistry applications (http://www. westgardicom/miticule ht). ‘Some alternative control rules are more suitable when three control materialsare analyzed, which is common for applications in hematology, coagulation, and immunoassays. 2General Interpretation of OC Result The acceptable reference limit is set at + 2SD. = The upper and lower reference limits define a specified percentage, usually 95%, of the values for a certain group of population, hence, 5% of the population will fall outside the reference = interval/limit in the absence of an abnormality or disease. = An analytical method is considered in control when there is symmetrical distribution of control, values around the mean and there are few control values outside the 2s control limits. "ifthe analytical test results are not within the + 2SD confidence limit, run a newset of controls and repeat specimen testing. = Acontrol value between 2s and 3s isa sign of a potential problem. = Acontrol value outside the 3s would require corrective action. "Some laboratories use the 2s as a warning limit and the 3s as an error limit. = Not every rule violation indicates that an analytic process is out control. * Continuous QC failure requires more trouble shooting - preparation of new reagents, recalibration, instrument maintenance and repair, and contacting the dealer/manufacturer for ‘technical support or service. = After the likely cause of the problem has been identified and corrected, a full set of control materials should be retested. Six Sigma (60) = It is a performance improvement program, in which the goal is to improve the process by eliminating variations or errors: improved performance, improved quality, Improved bottom line, improved customer satisfaction, and improved employee satisfaction. = It is a tool that can be used to reduce laboratory errors, increase productivity and improve quality in the clinical laboratory. = It measures the degree of variability or error in products or services through statististics and, quantitative parameters - process defects or errors are analyzed, potential causes are identified, and improvements are implemented. = Main Goal: To reduce the number of defects to near zero In Six Sigma, defects are "= anything that does not meet customer requirements—laboratory result error, delay reporting, or a quality control problem; and "generally measured per million opportunities (DPMO). Lean System > It is a system for reducing waste (“nonvalued activities”) especially in production or manufacturing processes. > twas conceptualized to improve the automobile industry in terms of the quality and efficiency of automobile production. > Itutilizes the 55 (Sort, Set in order, Shine, Standardize, and Sustain), and PDCA (Plan, Do, Check, and Act) systems to diminish costs by identifying daily work activities that do not directly add to the delivery of laboratory services in the most efficient or cost-effective ways. > It focuses on work flow actions in performing specific tasks, procedures, or other activities accomplished by critically reviewing each step in the process to determine where inefficiencies ‘can be eliminated. B> A “Lean Clinical Laboratory” utilizes fewer resources, reduces costs, enhances productivity, Promotes staff morale, and improves the quality of patient care (Rutledge, 2010 cited by Mc Pherson and Pincus, 2017). ® Example: relocating analytic equipment to an area that would require fewer steps, thus improving turnaround time; consolidating test panels to fewer instruments, eliminating the expense of maintaining multiple instruments and supplies; accesability to instruments and ‘materials such as pipettes, culture tablets, reagents, etc; and reallocating staff to maximize use and minimize wasteful downtime Terminologies: 1. Analytical Run ‘= _Itisa set of control and patient specimens assayed, evaluated and reported together. 2. Delta check = It is the most commonly used patient based-QC technique. ‘* Itrequires computerization of test data so that current results can be compared with past results. "= _Itis the difference between two consecutive measurements of the same analytes on the same individual. 43. Interference experiments ‘= These are used to measure systematic errors or inaccuracy caused by substances other than the analyte. "Interferences: hemoglobin, lipids, bilirubin, anticoagulants and preservatives 4. Linear Range/Dynamic Range "It is the concentration range over which the measured concentration Is equal to the actual concentration without modification of the method, 5. Physiologic Limit ‘+ Is sometimes referred to as absurd value. " _Ithelps detect sample contamination or dilution, inadequate sample volume, inadequate reagent volumes, sudden major problems with the method, or incorrect recording or transmission of the result. 6. Point Of Care Testing (POCT}/Decentralized Testing "itis a type of analytical testing performed outside the confines of the central laboratory, usually by nonlaboratorian personnel (nurses, respiratory therapists, etc). ‘* Most commonly used POCT: use of portable whole blood glucose meters for the management of patients with diabetes mellitus. 7. Quality Assurance (QA) ‘It can be envisioned as a tripod with program development, assessment and monitoring, and quality improvement forming the three legs. ‘Is a systematic action necessary to provide adequate confidence that laboratory services will satisfy the given medical needs for patient care. 14"= Philosophy of QA:Quality is important to customers. Quality can be assessed and monitored. Quality can be improved. Quality’s benefits exceed its costs. "Primary Goal of QA: To deliver quality services and products to customers. 8. Quality Patient Care = Itincludes effective test request forms, clear instruction for patient preparation and specimen handling, appropriate turn-around time for specimen processing, testing and result reporting, appropriate reference ranges and intelligent result reports. 9. Recovery experiment ‘= Itshows whether a method measures all the analytes or only part of it ‘= Itestimates inaccuracy or systematic error. 10. Predictive value ‘* Itdepends on sensitivity, specificity, and prevalence of the disease being test. '* When a test cutoff changes, its accuracy (sensitivity/specficity) andpredictive value also change. '* Bayes’ theorem (predictive value theory) describes the relationshipbetween posttest and pretest probability of disease or no diseasebased on the sensitivity and specificity of the test. a. Positive predictive value = Itis the probability that a positive test indicates disease ; it is the proportion of persons with a positive test who truly have the disease. +. Negative predictive value = Itis the probability that a negative test indicates absence of disease. It is the proportion of persons with a negative test who are truly without disease. 11, Reference Limit/Reference Interval/Reference Value = itis a value obtained by observation or measurement of a particular type of quantity on a reference individual. = Its a pair of medical decision points that extend the limits of test results for a certain healthy population. = It is the range of values into which 95% of nondiseased individuals will fall - this definition implies that 5% of nondiseased individuals can have laboratory results outside the reference range. = Itis the usual values for a healthy population that represents 95% central tendency. = itis usually established by the manufacturers of reagents or group of experts = Itis mostly determined using nonparametric statistics (CLSI recommended method). = Therapeutic drug targets or values are not derived from a heaithy population, but rather require unique physiologic conditions. 15Factors To Be Considered When Establishing Reference Intervals: 1. The composition of the reference population as to ege, gender, genetic and socioeconomic factors must be carefully evaluated and determined (many laboratory data depend on age and gender only). 2. The quantity of reference individuals 3. The inclusion and exclusion criteria for the required reference population must be created. 4. The physiologic and environmental conditions ofthe reference population (employment, obesity, lifestyle, habit, food and drug intake, et) 5. Categorize the potential reference individuals based on the criteria set (questionnaire). 6. The specimen-collection procedure, including preparation prior to testing must be standardized, 7. The analytical method used. ‘Notes to Remember: > Ideally, the laboratory should have age and sex-stratified reference values on all populations tested, > To derive reliable estimates of reference intervals, at least 120 individuals should be tested in each age and gender categories. > Only 20 reference individuals need to be sampled for analysis on the test instrument if the laboratorian determines that the test ins:rument and test subject population are similar to those described in the manufacturer's package insert. > For“verification” of already existing and established reference intervals, CLSI permits 20 subject specimens/individuals. The manufacturer's reference iimits may be accepted provided not more than 10% of the tested subjects fall outside the original reported imits; if greater than 10%, an additional 20 or more subjecis/samples should be analyzed. The reference values vary slightly depending upon method and specimen type. A critical or panic value is a laboratory resuitthat may represent a life-threatening situation that sometimes not readily identified. It should be communicated immediately to the clinicians for appropriate medical interventions. vi vv 16