Many lab professionals think Quality in a Medical Laboratory is

1
Many lab professionals think Quality in a Medical
Laboratory is .

• Accurate
• Timely
• Reliable
  
• 
  
• Reproducible

RESULTS
2

• INSTITUTE OF BIOCHEMISTRY WELCOMES YOU ALL TO THE
  CME
• ON
• QUALITY ASSURANCE IN CLINICAL LABORATORY
• UP, CLOSE PERSONAL

3
8 rights make Quality in a Medical Laboratory is
.

• Choosing the Right test
• Collecting Right Specimen
• After Right patient preparation
• Testing by the Right method
• Reporting the Right based on
• Right Reference intervals at the
• Right time and at the
• Right Price

4
Where can you go wrong in a lab ?
5
Quality is .

• A subjective term - for which each person has his
  / her own definition
• Technically Quality can have two meanings
• A product or service that fulfills the defined
  and expected requirement
• stated and implied needs
• A service or product free from defects
  deficiencies

6
Stages of Quality - Hierarchy
QUALITY MANAGEMENT
QUALITY SYSTEM
QUALITY ASSURANCE
QUALITY CONTROL
7
WHO definition of QA QC is .

• QA - Includes Internal QC, External QA,
  pre-analytic phase, test standardization,
  post-analytic phase, management, and organization
  (WHO, 1992)
• QC - Internal quality control (IQC) set of
  procedures for continuously assessing laboratory
  work and the emergent results immediate effect,
  should actually control release of results (WHO,
  1981)

8
Quality Assurance Quality Control - difference
is .

• QA is for correction prevention of errors or
  defects in the entire lab
• QC is detection of errors and defects in testing
  process

9
Purpose of Internal External QC

• Internal QC
• For CONTINUOUS IMMEDIATE (DAILY) monitoring of
  the laboratory work and the emergent results in
  order to decide whether the results are reliable
  enough to be released to physicians (WHO, 1981)
• Measures Precision Repeatability of the systems
  methods in use in the lab.
• Illusion of short term accuracy of the lab
  results.
• Do not detect the accuracy or trueness of patient
  results over a longer term

10
Purpose of Internal External QC

• External QC
• For PERIODIC AND RETROSPECTIVE monitoring of lab
  results by an independent external agency to
  indicate to the laboratory and its staff the
  accuracy or bias in their systems methods
• Lab can know its shortcomings and change their
  Internal Quality Assurance procedures.

11
Why EQAS is necessary?

• Serves as an educational tool and help to
  monitor improve the performance of the lab
• Measures the accuracy or bias of its results and
  stability of methods Over a longer period of
  time in terms of years
• Mandatory requirement for applicant accredited
  labs
• Non participation or repeated failures in an
  EQAS or PT programme may result in temporary
  suspension or cancellation of accreditation for
  those non EQAS tests
• Gives the laboratories, both the management
  technical staff, added confidence in their
  patient test results

12
WHAT DOES IT IDENTIFY ?

• Identifies systematic kit reagent problems,
  water quality problems, analyte calibration
  stability and status, equipment performance
• Indicator of where to direct improvement efforts
• Identifies training needs of lab personnel
• Benchmarks the labs performance against others
• Early Warnings System for Problems

13
HOW SHOULD IT BE USED ?

• Should only be used for motivating staff not
  to punish them
• Inaccurate lab results are not due to
  technicians
• But due to failure of lab systems

14
HOW SHOULD IT BE USED ?

• EQAS / PT samples should be treated exactly as
  the patient samples Only then the correct
  situation in the lab can be found fixing the
  problem becomes easy
• Never
• Run the calibration on the day of reporting EQAS
  sample if it is not a scheduled /required
  calibration
• Repeat the EQAS samples where as the patient
  samples are tested only once and give the mean of
  multiple runs
• Ask a specific analyst run the EQAS / PT sample

15
REMEMBER EQAS .

• SUPPLEMENTS Internal Quality Control
• NEVER a SUBSTITUTE for Internal QC
• Both measure 2 different aspects of quality

16
INTERPRETATION OF EQAS REPORTS

• VIS Variance Index Score
• SDI Standard Deviation Index
• Z-Score Classical Robust

17
Variance Index Score

• First proposed by the United Kingdom National
  Quality Control Scheme (UKNEQUAS)
• CCV (Chosen Co-efficient of Variation) DV
  (Designated Value) used to calculate VIS
• CCV is just the Allowable Limit of Error for an
  analyte (TEa) - Sum of both imprecision and bias
• Set recommended by WHO after studying the
  performance of many Indian labs

18
CCV of some common Analytes
Glucose  7.5 Sodium 2.3
Urea 10 Potassium 5.0
Creatinine 10 Chloride 6.0
CK 7.3 AST 12.5
T.Bilirubin 19.2 ALT 17.3
T.Protein 7.5 ALP 15.5
Albumin 7.5 Amylase 15.5
Calcium 6.0 LDH 15.5
Uric acid 7.7 Phosphorus 7.8
Cholesterol 7.5 Bicarbonate 9.0
TGL 14 HDL- C 7.6
HDL 7.6 Iron 15

19
Calculation of VIS

• Designated Value DV       120 mg
• Participant's result             95 mg
•      Variation V     Participant's Result -
  Designated value   
• 
  ---------------------------------------
  X 100                                            
           Designated value
•                                  120-95        X 
  100        25    X 100                         
          120                              120   
  
•                                          
•             20.8
• Variance Index                 V     X   100   
    20.8     X    100      277       
                              CCV                   
      7.5
• VIS 277

20
How to read the EQAS results ?
Calculation of VIS

• Check the VIS OMVIS values for each parameter
  every month
• Check if your value is close to DV
• Closer it is lower will be your VIS better is
  your labs accuracy
• Remember If your VIS is lt 50 it is regarded and
  given as zero score
• Even if gt400, it is still given as 400 only

21
Interpretation of VIS

• VIS Performance
• lt 100 Very good
• 100 -150 good
• 150 -200 satisfactory room for improvement
• gt 200 Not acceptable
• If VIS of gt200 on two or more occasions for the
  same analyte, them check your standardization
  procedures calibration
• Indicates an accuracy problem (systematic error
  / bias )

22
Interpretation of VIS

• Check the monthly OMVIS.(Overall Mean VIS)
  cumulative performance over a period  
• OMVIS Performance
• lt 100 Very good - your result are very
  close to DV
• 150-200 Need to take care of those parameters
  for which the reported values are very different
  from the  DV for that particular method
• gt 250 You are probably reporting many
  wrong results you should take urgent steps to
  locate the problem and correct them

23
Calculation of VIS Method Mean

• The Method Mean' - Mean obtained from results
  of all participating labs following the same
  method including results of outliers
• The Designated Value is the value obtained
  after excluding results, from labs with same
  method, which are gt 3SD of Method Mean and
  recalculating the mean after eliminating the
  outliers - Mean of inliers only
• The 'Reference Mean' - Mean obtained at the
  organizing lab  after exposing the QC samples to
  ambient temperature (25-35 C) for a period of 7-
  days (transport time) and analysing them on five
  different days
• The reference mean is shown  against the method
  by which it was analysed in the organizing lab

24
Mean value obtained at the Reference Lab (CMC)
after exposing 5 vials to ambient Temp. for 9
days and analyzing them on 5 diff. days
25
Mean value obtained at the Reference Lab (CMC)
after exposing 5 vials to ambient Temp. for 9
days and analyzing them on 5 diff. days
26
Mean value of results from of all participating
labs with same method
27
Value given (DV) is the mean of all participating
labs for that method after excluding results from
labs outside 3SD of the Mean Value (of the
participating labs with the same method)
28
STANDARD DEVIATION INDEX

• Another Statistical tool assigned to the lab
  by the EQAS / PT provider on the performance of
  the lab for each analyte in a EQAS cycle
• A measure of relative inaccuracy / relative bias

29
What is normal or Gaussian distribution ?

• Out of 100 results
• 68.2 of values fall within 1SD
• 95.5 of values fall within 2SD
• 99.7 of values fall within 3SD

30
IS KNOWING SDI USEFUL ?

• Yes - A measure of the result around a mean
  among a group of values
• Since 95 of all results in a normal population
  fall within 2 SDs of the mean, 2 SD is
  considered an acceptable laboratory value
• Expressed in the units being measured

31
IS KNOWING SDI USEFUL ?

• The number of Standard Deviations that your
  labs mean differs from the Peer Group Mean
• . Difference is converted into SD units (SDI)
• The SDI indicates how large / small the
  difference is between your result and target
  value
• Simply said
• Difference between your result and group mean in
  terms of the number of standard deviations from
  the overall mean

32
IS KNOWING SDI USEFUL ?

• Actual magnitude of the difference in the units
  of the test may look too small or too large
• To figure the actual size of this inaccuracy /
  bias in concentration units, you need to multiply
  by the actual SDI by of the group SD.
• For e.g if the group mean is 102 mg/dL for TGL
  and group SD is 5 mg/dL and your SDI score is 1.2
  
• Actual quantitative difference is 1.2 x 5 6
  mg/dL

33
IS KNOWING SDI USEFUL ?

• Any SDI of 2.0 or greater in a EQAS cycle for
  any analyte deserves special concern indicates
  some for of systematic error
• Any test whose average SDI is 1.0 or greater
  deserves some special attention because your
  method shows a systematic difference from the
  group.
• Likely to lead to unacceptable results in
  future
• SDI up to 1.0 your performance is satisfactory
  your result is with 1 SD of the group you are
  with in the 68.7 of labs result whose values
  are close to mean

34
IS KNOWING SDI USEFUL ?

• if you observe SDIs such as -0.4, -0.2, -0.5 -
  0.5, - 0.5 and -1.0 (all negative) for an analyte
  in successive cycles, your method is generally
  running on the low side and is negatively biased,
  on average, by 0.6 SDI
• You are reporting precise pateint values but
  lower than the true value by 0.6 SD ()
• So better
• calibrate your instrument and analyte
• or requires instrument maintenance

35
Z-Score

• Classical Z Score same as SDI
• Can be used for internal quality control also

36
Z-Score

• Robust Z score statistic is used when the
  distribution of results of participating labs is
  not Gaussian (not normally distributed) and
  there are varied results / outliers
• Both accuracy and precision (repeatability as
  well as reproducibility) are assessed in terms of
  robust Z score - both within and between labs Z
  score (ZB ZW)
• The participant labs are asked to analyse the
  same sample TWICE and submit both results to the
  EQAS provider

37
Z-Score
38
Z-Score

• Robust Z score
• Normalised Labs result- Median result of all
  labs )
• Normalized IQR ( Inter Quartile Range)
• It is calculated based on the median value
  (central value) and the interquartile range
• All results from participating labs are arranged
  in an ascending manner (lowest to highest) the
  central value is taken as the median
• The 25th and 75th percentile values are
  calculated The interquartile range is the
  difference between the 75th 25th percentile
• The 25th quartile (Q1) is the value below which a
  quarter of the results lie. Similarly, the 75th
  quartile (Q3) is the value above which a quarter
  of the results lie.
• IQR Q3 - Q1
• Normalized inter quartile range (NIQR) NIQR
  IQR x 0.7413 (a constant)

39
Z-Score

• The between laboratory Z-score (ZB)
• (S - median ) / (IQR x 0.7413)
• S (A B)/square root of (2) standardised
  sum of the two results for a laboratory (where A
  and B are results of two samples of the same
  test). Within laboratory Z-score (ZW)
• (D- median / (IQR X 0.7413) D (A - B)/square
  root of (2) standardized difference between
  the two results for a laboratory While testing
  two specimens in an EQAS / PT programme by a lab
  performance both ZW and ZB should be considered
  simultaneously for assessing the performance

40
Z-Score

• Interpretation of Robust Z-scores Z score less
  than 2 - Satisfactory Z score 2 but less than 3
  - Questionable perfromanceZ score more than 3
  Unsatisfactory
• Both ZB ZW should be lt 2
• Using only one Z-score may misleading
• ZW lt 2 ZB gt 2 Higher bias i.e low
  reproducibility
• ZB lt 2 ZW gt 2 low precision (i.e., low
  repeatability)
• For assessing laboratory's technical competency,
  both ZW and ZB value should be low at the same
  time.

41

• EXERCISES ON INTERPRETATION OF EQAS REPORTS
• Dr.V.K.Ramadesikan

42
Why Analysis of EQAS reports is important ?

• True benefit of EQAS /PT proficiency testing,
  lies in carefully critically evaluating your
  results and report
• Proper analysis of EQAS testing results can
  reveal problems even before failure in EQAS or
  even an adverse patient result
• Potential problems can be recognised by
  recognising patterns from graphs
• Review both the present cycle results as well
  as performance from previous cycle for the same
  analyte
• Graphs of Z score or SDI or VIS or deviation
  for trends
• Otherwise trends will be missed

43
How to recognise problems ?

• Results may consistently be different from the
  target peer group mean - All results on one side
  of the mean (may be close to mean or at variable
  distances from mean) Systematic Error
• Majority of results are close to the target
  value but some show larger deviation s on one
  side or both side of mean Random Error
• Reasons and hence corrective action differs

44
How to recognise problems ?

• Random error is an error / mistake / inaccuracy
  that has no set pattern. Its occurrence cannot be
  predicted
• Results on an average are close to target mean.
  Few results show large deviations on either side
  of target
• Detected by sudden, undue deviation /SDI /Z
  Score gt 3.5
• Indicates labs imprecision / poor
  reproducibility
• Easily identified values are far beyond the
  usual e.g SDI from 1.5 to 4.0 6.0

45
How to recognise problems ?

• Systematic error set pattern of error /
  mistake. Its occurrence can be explained and
  corrected
• Constant difference between the participant
  labs value and group mean All results lie on one
  side of mean
• Indicates labs bias or inaccuracy but good
  precision
• A progressive increase in deviation on the same
  side shift or stabilizes after a gradual
  increase trends in

46
Sources of Random Errors

• Random errors are blunders
• Transcription errors result not correctly
  transcribed from the instrument to workbook or PT
  sheet or from the Workbook to the system
• Misplaced decimal points e.g serum potassium
  58.2 instead of 5.82
• Result was entered in the wrong instrument or
  method group on the result form of PT provider.
• Lab might have changed the method or instrument
  recently but not updated with PT provider

47
Sources of Random Errors

• Mislabeling errors, interchanging results of PT
  specimens, misplacing specimens in analyser rack,
  inappropriate reagents and standards
• Result of some other analyte entered in the PT
  form of the provider in the system
• Wrong units. Result in one unit in the
  instrument but not converted to unit of EQAS
  provider e.g ug /mL instead of ng/L , mg/dL
  instead of g/L
• Result found on evaluation report from PT
  provider not matching with the result on the
  report form mistake of PT provider
• Calculation errors (conversion from one unit to
  another)

48
Sources of Systematic Errors

• Recalibration if not done earlier esp if a new
  lot of reagent has been used or if the open vial
  stability of the reagent is doubtful
• Instrument maintenance major part needs
  servicing or replacement (optics, alignment,
  incubation temperature failure, Internal QC
  values having a bias
• Recent instrument malfunction, instrument
  failure soon after PT specimen testing
• Assay settings (sample volume, reagent volume,
  delay time incubation time no of readings to be
  taken etc. )
• improper reconstitution of QC materials, water
  quality, not following manufacturers / PT
  providers instructions while reconstituting,
  storing or handling reagents / QC materials

49
What type of Error is indicated in green and red
?
50
Systematic Error
51
SAMPLE EQAS MONTHLY REPORT
Your result
52
SAMPLE EQAS MONTHLY REPORT
53
SAMPLE EQAS END OF CYCLE REPORT
54
SAMPLE EQAS END OF CYCLE REPORT
55
(No Transcript)
56
SAMPLE CAP EQAS REPORT
57
SAMPLE CAP EQAS REPORT
58
SAMPLE CAP EQAS REPORT
59
SAMPLE EQAS REPORT

• Transcription error
• Misplacing of specimens or interchanging results
  of PT specimens

U 1 31.3 909.51 U3
U 2 13.1 28.31 U1
U 3 1031.9 14.8 U2
60
ALT 226 U/L
Mean 445
Result - 226
Mean 445
SDI 4.34
SDI 4.34
61
Analytical error Calculation error PT was
repeated with 1 in 2 dilution as results were
high but the result was not multiplied by 2
Diluted result was sent
62
SAMPLE EQAS REPORT
63
SAMPLE EQAS REPORT

• Erratic Results Very poor precision and
  accuracy
• Problem Internal Controls wide SD due to
  various reasons
• Instrument not maintained calibrated reagents
  problems deterioration
• Analyte not calibrated
• Procedural problems

64
SAMPLE CAP EQAS REPORT
65
SAMPLE EQAS REPORT

• 1. Systemic error may have seeped in the system
  anytime after the last EQAS sample reporting
  look at daily controls for any shift in values
• 2. May be a random error Keep a watch on daily
  controls and follow during next EQAS
• 3. Wrong sample was tested
• 4. Wrong preanalytical stage reconstitution,
  storage etc look at values of other analytes
• Are they with in acceptable limits of SDI and in
  the same direction ?

66
SAMPLE CAP EQAS REPORT
67
SAMPLE EQAS REPORT

• Systematic error one showing a positive bias
  and other a negative bias
• Needs
• Recalibration for analyte
• Instrument maintenance
• Internal QC Value Mean needs to be reset

68
SAMPLE EQAS REPORT
69
SAMPLE EQAS REPORT

1. Fairly a good performance in terms of accuracy
   and precision
2. Values on either side of mean No Bias
3. Values within /- 0.5 SD