Quality control

1. GENERAL PRINCIPLES OF
CONTROL CHARTS

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

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

18. WEEKLY QC REVIEW
 Short term changes
 Weekly maintenance checks

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

23. Interferences Calibration Nonlinearity
Contamination Uncorrected
blank
Bias
Instrumental
shifts
Lack of
Precision
Operator
Bias
Matrix
effects
Reagent
Instability

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