This document discusses general principles of control charts used to monitor analytical testing methods. It provides information on how control charts are constructed using mean and standard deviation values from repeated measurements on known specimens. Control limits are calculated and plotted, with points falling within the limits indicating proper method performance. The document also describes procedures for identifying random and systematic errors using rules for interpreting control chart results.

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