2. GENERAL PRINCIPLES OF CONTROL
CHARTS
Control charts are simple graphical displays in
which observed values are plotted versus time
when observations are made
When plotted points fall within control limits this
occurance generally is interpreted to mean that
the method is performing properly
Contol limits are calculated from the mean (×) &
SD(s) obtained from repeated measurements on
known specimens by particular analytical
method that is to be controlled
3. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Test:
T. v alue:
USD:
Method:
Month:
Batch No.:
Current month Previous months
Normal Abnormal Normal Abnormal Normal Abnormal
Mean
SD
0 Technician
1 Instrument
2 Control
3 Reagent Batch
4 Environmental
5 ________
6 ________
7 ________
8 ________
9 ________
PROCEDURE
CHANGE
CODES
Procedure
changes
Department:
+1SD---------
+2SD----------
+3SD---------
Mean---------
-1SD---------
-2SD---------
-3SD---------
+1SD----------
+2SD----------
+3SD---------
Mean---------
-1SD----------
-2SD----------
-3SD----------
Normal
control
Abnormal
control
4. x1= is an individual control observation
n= is the number of observations in time period
being monitored
Initial estimate should be based on
measurements obtained over period of at least
one month when method is working properly
Control charts are used to compare the observed
control values with control limits & provide visual
display that is inspected & reviewed quickly
Observed values are plotted on y-axis versus
time of observation on x-axis
Usually one month data are plotted on chart &
usually only one or two points a day
5.
6. LEVEY–JENNINGS CHART
Consecutive daily measurements of controls are
used to calculate the mean and SD value
These values are plotted on a graph paper
showing horizontal lines of mean 1SD, 2SD
and 3SD.
If precision remains unchanged then 68% fall
within 1SD and 95% values fall within 2SD
range
Permissible limits are one in 20 readings crossing
2SD limit and 3 in 1000 readings crossing 3SD
limit
7. PERFORMANCE CHARACTERISTICS
OF A CONTROL PROCEDURE
Levey-Jennings control chart with control limit set as a
mean ± 2s has high rate of rejections when the method is
actually performing satisfactorily(“false alarms”)
Use of 3s control limits reduces the false alarms to 1% or
less
False rejections are in effect an inherent property of
control procedure
They occur because of control limits that have been
selected not because of any problems with analytical
method
Use of 2s control limits generally is not recommended
With use of 3s control limits false rejections problems are
eliminated
8. WESTGUARD RULES
12s : One control observation exceeds control limit set at
2SD is a warning sign
13s : One control observation exceeds 3SD is a random error
subject to rejection rule
22s :Two consecutive control observations exceed 2SD is a
systematic error subject to rejection rule
R4s : One control observation exceeds the +2SD and the
second control observation exceeds –2SD is a random error
subject to rejection rule
41s : Four consecutive readings crossing 1SD on one side is
a systematic error subject to rejection rule
10x :Ten consecutive control readings on one side of the
mean is a systematic error subject to rejection rule
9. FIVE CYCLE QUALITY PROGRAMME
Designed to facilitate the plotting of daily results
Control runs are repeated 5 times to provide a wide range scale of values
Small numbers to the right assist in identifying correct points for plotting
the result
Horizontal lines are drawn at the mean and 2SD values of reference
material or control
Control is inserted randomly in daily runs and mean is calculated at the
end of the day
I : 32 chance is that 5 results in a row will be either above or below the
mean value --- called upward or downward shift(change in reagent
concentration, equipment calibration or deterioration of reference
material)
If test value keeps on decreasing or increasing on one side of the mean --
- called upward or downward trend(gradual change in calibration or
concentration or apparatus, reagents, reference material or equipment)
10. ROLE OF QUALITY CONTROL IN ERROR
DETECTION
Systematic errors
The occurrence of six or more values on one side of the
mean
Random errors
When any value falls outside + 3 SD, it is termed random
error
Excessive random error
When the control values are within + 2 SD and less than two
third of the values are outside the + 1 SD limits
11. RANDOM ERRORS
Instability of instrument
Variations in temperature
Variations in reagents and calibrators
Variations in handling techniques
Pipetting
Mixing
Timings
Variations in operators
Dissolving of reagent tablets & mixing of samples &
reagents
Lack of stability of temperature baths,time regulation &
photometric & other sensors
12. Impure calibration material
Improper preparation of calibrating solutions
Erroneous set point & assigned values
Unstable calibrating solutions
Contaminated solutions
Inadequate calibration techniques
Inadequate sample blank
Unstable reagent blank
SYSTEMATIC ERRORS
13. QUALITY CONTROL OF
REAGENTS
Date of preparation
Expiry date
Requirment for refrigeration
Radioactivity, flammability
One container should be reserved for one
reagent only
Reagents should be free from dirt and direct
sun light
14. CALIBRATION
Controls and calibrators
Matrix effect
Sufficient for 1-2 years
Verification of new calibrators
Duplicate measurements
QC samples
Patient samples
Average SD test
Significant difference ?
15. INTERNAL QUALITY CONTROL
PROGRAMME
Daily running of same control sera
Calculation of their mean and standard
deviation
Plotting them on control charts
Interpretation of graphs as follows:
Daily checking
Monthly checking
Annual checking
Periodic checking
16. DAILY QC PROGRAMME
Check daily maintenance schedule
Record any change
Frequency of known QC samples (Continuous
?) and calibrators
Frequency of blind QC samples
Run QC samples before unknowns
Normal and abnormal control
17. DAILY QC PROGRAMME
QC of stat analysis
Decision about validity of batch results
Record of QC result and decisions
Error detection and correction
Log book maintenance
Communication with physicians
19. MONTHLY QC REVIEW
Check monthly maintenance
Shifts or trends
Evaluate monthly average against target average
(temporary and final)
Evaluate SD against USD
If difference between target average and monthly
average is more than 1 USD
Decision A: Statistically significant ?
Decision B: Medically significant ?
Decision C: Operationally significant ?
20. MONTHLY QC
Compare monthly SD with USD
If difference is more than half USD
Statistical significance ?
Medical significance ?
Operational significance
Significant change limit
Check against QC goals
Monthly patient average limit
21. USE OF BLIND SAMPLE IN
QUALITY CONTROL
Blind quality control samples
Blind patient samples
Purpose of blind sample system
22. PRECISION
Selection of specimens
Arrangement of
specimens
Effect of concentration
Between-batch CV
Within-batch CV
24. CORRECTION OF THE ERRORS
Errors in accuracy
Stepwise checking/replacement
Control material
Standards
Sampling system
Measuring system like
spectrophotometers
Reagents
Performance of analytical system