clin chem 1

MAIN SOURCE: lecture/bishop 9e/rodriguez 2018 ESTRADA SY

QUALITY CONTROL
 Variations between vials should be minimal (so that
QUALITY CONTROL differences seen over time can be attributed to the
 Quality control – system for verifying and maintaining analytic method itself and not variation in the QCM)
desired level of quality in an individual test or process  Concentrations should span the clinically important
 includes techniques and procedures that monitor range of analyte at appropriate decision levels
performance parameters  range of L1 (normal) ≠ L2 (abormal)
 utilizes statistical procedures to monitor and  For chemistry assays, two levels of control (L1 & L2) are
evaluate analytical processes used
📌 QC also includes non-statistical check procedures 📌 Other assays e.g., immunoassays makes use of three
(linearity checks, reagent & standard checks, and
temperature monitors) Other notes:
 QC involves quantitative techniques that:  Assayed controls give expected target ranges
 monitor and detect sources of analytical error (including mean & SD)
 estimate magnitude of errors  Unassayed QC materials do not have assigned
 alert lab personnel analyte values provided by the manufacturer ∴ the
laboratory assigns the expected results to unassayed
PURPOSE OF QUALITY CONTROL QCM
(1) Detect analytical error  Reconstituted material are generally more turbid than
(2) Prevent reporting of incorrect patient result actual patient specimen
(3) Part of performance monitoring  Improper preparation and handling of QCM is the
main reason for QC failures in the laboratory
ASPECTS OF QUALITY CONTROL  QCM must be analyzed between 5-20 days
A. Analytical methods – definitive procedure that produces  special, highly precise assays: atl 5 days in a
a test result month
 methods used in measuring analytes;  routine, less precise assays: atl 20 days in a
equipment/machine month
 accurate and relevant
 effective and reproducible C. Comparing values – comparison of determined values
 analytical sensitivity (can measure smallest vs. expected values and corresponding interpretations
concentration)  QC charts are used to visualize QC data
 analytical specificity (can measure only the analyte  Control charts – graphically represent the observed
of interest) values of a control material over time in the context of
📌 QC should be incorporated in the test performed the upper and lower control limits in relation to the target
B. Control materials/QC materials – specimens analyzed value
for QC purposes  can detect errors in accuracy (systematic error) and
 should be the same matrix as the patient imprecision (random error) over time
specimens  Control limits – expected values represented by
a. Pooled control sera – mixture of residual patient intervals of accepted values with upper and lower limits
serum (of known concentration) with normal results  expressed as the mean ± SD
Disadvantages:  CL set at ±2SD = 95% confidence level
 Easily deteriorated  CL set at ±3SD = 99% confidence level
 Contamination  points falling outside the control limit suggest that
 Loss of potency (loss of stability of biochemical problems may be developing
substances)
 Exposure to pathogenic agents QUALITY CONTROL CHARTS
b. Commercially-prepared serum – purchased from  Levey-Jennings/Shewhart Chart – most commonly used
manufacturer  devoid of matrix effects & chart that allows observation of shifts, trends, and
pathogens multigard rules violations (random & systematic errors)
Prepared from either: GENERAL COMPONENTS OF THE LJ CHART
1) Human serum  x axis = time/runs/control observation number
2) Bovine-based products (cow, ox, buffalo)  y axis = control limits  derived from control values
📌 Liquid CM have the advantage of eliminating errors from the control runs of the previous month
caused by reconstitution Green totally accepted
CHARACTERISTICS OF IDEAL QCM Yellow warning
 Available in sufficient quantity (same lot number) Red rejection
 if not same lot number, troubleshoot: re-calibrate  USING THE LJ CHART
QC  test px samples (1) Obtain control values
 Last at least a year  control values are plotted
 Aliquoted in stable form  control values FROM LAST MONTH are used to
 lyophilized (through cryodesiccation/freeze-drying) calculate for the mean and SD to derive the control
that comes with specific diluents to reconstitute the limits
QCM 📌 Essentially, control values from this month are plotted
📌 Once reconstituted, QCM must be immediately used against the control values (converted to control limits) from
last month
Page 1 of 4
(2) Compute for the mean QUALITY CONTROL
ERRORS OBSERVED IN LJ CHARTS
a. Shift – sudden switch of data points to another area of
the control chart away from the mean
(3) Compute for standard  indicative of abrupt changes in the control mean
deviation Causes of shifts
(4) Compute for control 1. Sudden failure or change in the light source
limits: mean±nSD 2. Change in reagent formulation
3. Change of reagent lot number
How to compute for values of increments? 4. Major instrument maintenance
- Subtract mean from -1SD 5. Sudden change in incubation temperature
- Then, divide to number of increments 6. Change in room temperature or humidity
- The quotient is added/subtracted per increment 7. Failure in the sampling system
QC SPECIFICATION VS QC LIMITS 8. Failure in the reagent dispense system
 QC specification – criteria used to decide that a given 9. Inaccurate calibration or recalibration
method meets clinical requirements
 define if the overall performance of the process b. Trends – characterized by an increase or decrease of
meets the quality goals control values for six consecutive days
 QC limits – defined by the process itself based on its  systematic drift in one direction away from
natural variation established mean
 define if a process is in control on a daily basis  gradual loss of reliability in the test system
📌 QC specifications are what you want, QC limits are what Causes of trends
you have 1. Deterioration of the instrument light source
WAYS LJ CHARTS ARE OBSERVED 2. Gradual accumulation of debris in the sample/reagent
📌 Both commonly applied in 22s rule tubing
1) Within run/Across control materials – observing all 3. Gradual accumulation of debris on electrode surfaces
control results obtain for the current analytical run  two 4. Aging reagents
control levels are considered at one time 5. Gradual deterioration of control materials
6. Gradual deterioration of incubation chamber temp.
7. Gradual deterioration of light filter integrity
8. Gradual deterioration of calibration

RANDOM VS SYSTEMATIC ERRORS

 Random error – present in all measurement; due to
chance; subjective
 varies from sample to sample
 basis for varying differences between repeated
measurements
 mainly due to instrument, operator, and
environmental conditions
2) Across run/Within one control material – observing 📌 Mainly affects precision
results of one control level over subsequent runs Causes of random error
1. Variations in line voltage
2. Pipettes/dispensers
3. Contamination
4. Errors in dispensed volume
5. Bubble in lines or reagents
6. Mislabeling of samples
7. Improper mixing of sample and reagent

 Systematic error – error that influences observations

consistently in one direction (constant difference)
 may either be positive or negative
INTERPRETATION OF LJ CHARTS
📌 Mainly affects accuracy
When controls fall within their statistical limits, results are
1. Calibration lot changes
reported are considered valid;
2. Temperature changes in incubator unit
Control results are “in”, the run is accepted and patient
3. Light source deterioration
samples can be assayed or reported
4. Reagent lot changes
If any control in a run is out of control, the results cannot
5. Poorly made standards
be reported with confidence;
6. Instrumentation problems
The run is rejected, the problem is identified, and the new
7. Poorly written procedures
run can begin
8. Inadequate staff training

WESTGARD MULTIRULES
 Multirules – establish a criterion for judging whether an
analytic process is out of control
 Control rules indicate the number of control
observations per analytic run, followed by the control
amount in subscript

Page 2 of 4
 QUALITY CONTROL
RULE DESC. ERROR ILLUSTRATION R/A?
12S One control observation A warning rule that initiates ✅
exceeding the mean ±2s testing of control data by other
rules; Sensitive to random error

13S One control observation High sensitivity to random error 🚫

exceeding the mean ±3s

22S Two control observations High sensitivity to systematic 🚫

consecutively exceeding the error
same +2s or –2s

R4S One control exceeding the +2s Detection of random error 🚫

and another exceeding the –2s

41S Four consecutive control Detection of systematic error ✅

observations exceeding +1s or –
1s

10X Ten consecutive control Detection of systematic error 🚫

observations falling on one side
or the other of the mean (no
requirement for SD size)

31S Three consecutive control Detection of systematic error ✅

observations exceeding +1s or –
1s

Page 3 of 4
 QUALITY CONTROL

OTHER CONTROL CHARTS

Page 4 of 4