1. Why Use Them? By: Marcy Bolek – Alloway
Control Charts
Why Use Them?
By: Marcy Bolek – Alloway

2. Really, Control Charts?
BORING!!!!!

3. Not for me, right?

4. Electronic Code of Federal Regulations

5. 40 CFR 136.7 https://www.ecfr.gov/ Title 40
Part – Quality Assurance and Quality Control
C(1) – “Incorporate the following twelve quality control elements, where applicable, into the laboratory's documented standard operating procedure (SOP) for performing compliance analyses when using an approved part 136 method when the method lacks such QA/QC procedures.”
“ One or more of the twelve QC elements may not apply to a given method and may be omitted if a written rationale is provided indicating why the element(s) is/are inappropriate for a specific method.”
C(1)(viii) – Control Charts or other trend analysis

6. Standard Methods for the Examination of Water and Wastewater
Method 1020 A
Applies mostly to chemical analysis
Method 9020 applies to microbiology Quality Control
Method 1020 B
Section 13 – Control Charts

7. Accuracy & Precision

8. Alloway
Precision
… the degree to which several measurements agree with each other (reproducibility) …
How often can an analyst analyze the same sample and obtain the same result or close to the same result

9. Precision & Range Control Charts
Perform identical analysis on two portions of the same sample (duplicate)
Perform identical analysis on two portions of the same sample that are spiked (matrix spike/matrix spike duplicate)
20 or more data points used to generate chart

10. Evaluation of Precision
Alloway
Evaluation of Precision
Calculate the range (R):
R = │Value 1 – Value 2│ _
Calculate the Average Range (R):
R = (ΣR) / n

11. Evaluation of Precision
Alloway
Evaluation of Precision
Range Charts are used to evaluate precision
Range Charts typically plot
Upper Warning Level (UWL)
Upper Control Limit (UCL)
Average Range of data set (mean)
Confidence Limits
95% Confidence Limit
99% Confidence Limit

12. Precision & Range Control Charts
Calculate range of duplicate determinations in data set
Calculate average range
Calculate UWL
Calculate UCL

13. Precision & Range Control Charts
UCL
Difference
UWL
Mean range
Sample Number

14. Alloway
Range Chart

15. Precision & Range Control Charts
Result 1
Result 2
10.6
11.2
14.7
13.9
10.1
11.0
15.4
17.9
18.3
15.1
21.9
18.7
19.9
18.8
12.5
11.7
14.8
14.6
19.4
23.3
24.4
Range
0.6
0.8
0.9
2.5
3.2
1.1
0.2
0.7
Calculate the range of each duplicate set
Calculate the average range (mean)
mean = 1.37

16. Precision & Range Control Charts
Calculate the upper control limit
Calculate the upper warning limit

17. Precision & Range Control Charts
UCL – 4.48
Difference
UWL – 3.44
Mean range
Sample Number

18. Evaluation of Precision
Alloway
Evaluation of Precision
If there is a wide variety in amounts of analyte, classify precision data into levels
For example:
Low – 1 to 10 mg/L
High – 10 – 100 mg/L
Develop a separate chart for each level

19. Precision Relative Percent Difference (RPD) RPD = Difference * 100
Alloway
Precision
Relative Percent Difference (RPD)
RPD = Difference * 100
Average
Precision of High Level Duplicate
Precision of Low Level Duplicate
Possible discussion of Maximum Advantage QA Software (before Chapter 6):
1. Discussion of System Security
2. Create an analyte header and enter users
3. Create a new analyte with 5 or 6 QC activities
4. Enter data
5. Review Data
6. Evaluate reagent reports
a. Analyte history
b. Exception log
7. Create statistical limits
Also there is a wastewater handout

20. Alloway
Accuracy
… the degree to which the measurement approaches the “true value”…

21. Evaluation of Accuracy
Alloway
Evaluation of Accuracy
Analyze sample containing known concentration of analyte
QC sample run at method specified frequency
20 data points used to make chart

22. Accuracy & Means Control Charts
Analyze sample containing known concentration of analyte
Can be QC standard
Initial Calibration Verification
Continuing Calibration Verification
Laboratory Control Standard
Laboratory Fortified Blank
Matrix spike results are charted using means control chart
20 data points used to make chart

23. Accuracy & Means Control Charts
Confidence Limits
Range within which we can state the probability that the true value lies within this range
68% confidence limit: mean ± 1 SD
95% confidence limit: mean ± 2 SD
99% confidence limit: mean ± 3 SD

24. Accuracy & Means Control Charts
Normal Distribution

25. Accuracy
Percent Recovery

26. Accuracy & Means Control Charts
Calculate the mean
Average of the data set
Calculate the standard deviation
Use sample standard deviation

27. Accuracy

28. Alloway
Means Chart

29. Accuracy & Means Control Charts
Calculate the mean and standard deviation of the LCS results for phosphorus and prepare a means control chart.
Mean = 99.76
SD = 5.102
LCS %R
98.5
103.4
105.6
92.5
93.2
91.8
101.5
108.9
100.8
105.2
97.9
100.7
96.8
99.8

30. Accuracy & Means Control Charts
Calculate the UWL
UWL = mean + (2 x SD)
= (2 x 5.102) = =
Calculate the LWL
LWL = mean – (2 x SD)
= – (2 x 5.102)
= – (10.204)
= 89.56

31. Accuracy & Means Control Charts
Calculate the UCL
UCL = mean + (3 x SD)
= (3 x 5.102) = =
Calculate the LCL
LCL = mean – (3 x SD)
= – (3 x 5.102)
= – (15.306)
= 84.45

32. Accuracy & Means Control Charts
UCL
UWL
Mean 99.76
LWL 89.56
LCL 84.45

33. Control Charts Importance of Control Chart
Alloway
Control Charts
Importance of Control Chart
point out problems as they occur
Detect trends which can lead to out-of-control conditions
documentation that analysis was in control at a specified time
Indication of improving performance when limits are narrowed

34. Control Charts
Trends
5 to 7 consecutive points with an increasing or decreasing slope
Requires corrective action
Corrective action must be documented

35. Trend

36. Trend

37. Control Charts
Shifts
5 to 7 consecutive points above or below the mean
Requires corrective action
Corrective action must be documented

38. Alloway
Shift

39. Shift

40. Spiked Samples Means chart generated similar to accuracy QC
Alloway
Spiked Samples
obtained value – sample value
% Recovery = X 100
spiked amount
Means chart generated similar to accuracy QC
% Recovery example: Obtained value=1.2, Sample value=0.4, spiked amount=1.0
SO, the %recovery=80%

41. Spiked Samples Problems encountered with sample spikes:
Alloway
Spiked Samples
Problems encountered with sample spikes:
High standard deviations caused by two measurements
Mixing problems, especially with two-phase samples
Table 1020:I. Acceptable Limits for Duplicate Samples and Known Additions to Water and Wastewater