QUALITY

ASSURANCE
PROGRAM FOR
FEED TESTING
LABORATORIES

Part 2
ISO 17025 ON QUALITY CONTROL
5.9.1 Quality Control
- Ensure laboratory has quality control
procedures for monitoring validity of tests
and calibration.
- Planned activity, reviewed and includes:
• regular use of CRMs or secondary RMs

• Participation in interlaboratory
proficiency testing program
• Replicates using the same or different
methods
• Retesting or recalibration of retained
items
ISO 17025 ON QUALITY CONTROL
5.9.1 Quality Control
- Planned activity, reviewed and
includes:
• correlation of results for different
characteristics of an item
- Resulting data must be recorded so as
trends are detectable and statistical
techniques must be applied to the
reviewing of the results
 ACCURACY TEST

- RMs and CRMs may be purchased from

NIST and other reputable organizations
- Can be expensive

- Important considerations:
- Matrix
- Certified value should be within your
range of analysis

-- CRM provider should be accredited with

- ISO 34 Standard

Websites
1.http://irmm.jrc.ec.europa.eu/reference_materials
_catalogue/catalogue/Pages/index.aspx
2. www. comar.bam.de
 Link for AAFCO Analyte Code

http://www.aafco.org/Directory/CommitteePages/
CheckSampleProgram.aspx
AAFCO CONTACT
 Victoria S. Siegel Ph.D.
 Chair, AAFCO Collaborative Check Sample
Program
 Office of Indiana State Chemist

 Purdue University

 175 S. University Street

W. Lafayette IN 47907-2063
USA
 T 765-494-1565

 F 765-494-8722

 E-mail: vsiegel@purdue.edu
PAO ACCREDITED LABS
 http://www.dti.gov.ph/dti/index.php?p=543
CONTROL CHART PREPARATION
PROCEDURE

 Identify QC samples or In house checks

- Must be widely used sample, stable over long
periods (at least 6 months)
- Ideally, you should have QC sample for each
critical analysis in the lab
e.g Protein in finished good
Protein in Soya and/or Fish Meal
Fat in Finished good
Calcium in layer feeds
CONTROL CHART PREPARATION
PROCEDURE
 Prepare QC samples or In house checks

- Collect at least 1 kilo of sample

- Grind until uniform particle size is achieved
- Mix whole batch very well
- Transfer into individual bottles
 Analyze QC sample with every batch of analysis
until you gather 10 data points
 Prepare an initial control chart using the first 10
data points
 Check for outliers and exclude data. Gather new
data to replace outliers
 Continue gathering data until you get 20 data
points
CONTROL CHART PREPARATION
PROCEDURE

 Prepare a final control chart using the 20 data

points
 After gathering 60 data points, replace control
chart using the 60 data points, excluding outliers
 Do this again after the 120th data point
CONTROL CHART CALCULATION
 Get the mean of the 20 data points
 Get the standard deviation

 Get the + 1 sd line by adding the std dev to the

mean
 Get the - 1 sd line by subtracting the std dev to
the mean
 Multiply the standard deviation by 2 and:

- add to the mean to get the upper warning limit

- subtract from the mean to get lower warning
limit
CONTROL CHART CALCULATION
 Multiply the standard deviation by 3 and:
- add to the mean to get the upper control limit
- subtract from the mean to get lower control
limit
 CONTROL CHART WORKSHOP
Construct a Control Chart using the following
data:
Date % Protein Date % Protein
Analyzed Analyzed
July 1 20.3 July 17 20.7
July 2 20.6 July 18 20.5
July 3 20.8 July 20 21.0
July 4 19.7 July 23 20.0
July 5 19.5 July 24 20.3
July 7 20.5 July 25 20.5
July 9 20.4 July 26 20.4
July 11 19.8 July 27 20.6
July 13 19.9 July 28 19.8
July 15 20.2 July 29 19.9
STEP BY STEP PROCEDURE FOR
CONSTRUCTING CONTROL CHARTS

 Prepare a Quality Control Sample

 Gather 20 data points by analyzing the Control
sample with every batch of sample you routinely
analyze
 Get the mean and standard deviation

 Get 2 x std dev and 3 std dev

 Plot the mean, 1sd, 2 sd and 3 sd

 Lower control limit = warning limit = 2sd

 Upper control limit = control limit = 3 sd

 CONTROL CHART

% Protein
Upper Control Limit
21.5
Upper Warning Limit
21.1
20.7

20.3

Lower Warning Limit 19.9

19.5
Lower Control Limit
19.1

Analysis Date
 CONTROL CHART WORKSHOP
Using the Control Chart, plot the following data
and comment on the trends:
Date % Protein Date % Protein
Analyzed Analyzed
Aug 1 20.6 Aug 20 20.8
Aug 2 21.7 Aug 21 20.9
Aug 5 20.8 Aug 23 20.8
Aug 7 20.5 Aug 24 20.8
Aug 8 20.2 Aug 25 20.6
Aug 9 19.8 Aug 27 20.5
Aug 11 20.0 Aug 28 20.4
Aug 12 21.1 Aug 29 20.2
Aug 14 21.3 Aug 31 19.9
Aug 15 21.2 Sept 3 19.5
Aug 17 20.2 Sept 4 19.3
Aug 18 20.0 Sept 5 19.0
Aug 19 20.5 Sept 6 19.5
 OUT OF CONTROL SIGNAL
A single point outside the control limits.

Two out of three successive points are on the same

side of the centerline and farther than 2 σ from it.

Four out of five successive points are on the same side

of the centerline and farther than 1 σ from it.

A run of 8 in a row are on the same side of the

centerline or 10 out of 11, 12 out of 14 or 16 out of 20.
 CONTROL CHART WORKSHOP 2
Construct a Control Chart using the following
data:
Date % Protein Date % Protein
Analyzed Analyzed
July 1 48.1 July 17 47.7
July 2 48.5 July 18 48.4
July 3 48.2 July 20 47.6
July 4 48.6 July 23 48.2
July 5 48.8 July 24 47.9
July 7 47.6 July 25 48.2
July 9 47.8 July 26 48.1
July 11 48.1 July 27 48.3
July 13 48.3 July 28 48.2
July 15 48.5 July 29 47.7
 CONTROL CHART

% Protein
Upper Control Limit
49.0
Upper Warning Limit
48.7
48.4

48.1

Lower Warning Limit 47.8

47.5
Lower Control Limit
47.2

Analysis Date
 CONTROL CHART WORKSHOP 2
Using the Control Chart, plot the following data
and comment on the trends:
Date % Protein Date % Protein
Analyzed Analyzed
Aug 1 48.0 Aug 20 48.2
Aug 2 47.8 Aug 21 48.4
Aug 5 47.6 Aug 23 48.3
Aug 7 47.4 Aug 24 48.4
Aug 8 47.3 Aug 25 48.3
Aug 9 47.2 Aug 27 48.2
Aug 11 47.5 Aug 28 48.5
Aug 12 47.0 Aug 29 48.7
Aug 14 47.8 Aug 31 49.1
Aug 15 47.6 Sept 3 48.4
Aug 17 47.7 Sept 4 48.1
Aug 18 47.5 Sept 5 48.2
Aug 19 47.6 Sept 6 47.7
CONTROL CHART
RANGE CHART
 Perform duplicate analysis on Quality Control
sample
 Collect at least 25 duplicate analysis results

 Take the difference between the duplicate analysis

(Range)
 Get the average of the Range

 Calculate the UCL : D4 x Mean of Range

where D4 can be obtained from the table below
 Calculate the LCL for range = D3 x Mean of Range
where D3 can be obtained from the table below
CONTROL CHART
RANGE CHART – CALCULATION OF CONTROL
LIMITS

No. of D3 D4
Observations
20 0.414 1.586
25 0.459 1.541
RANGE CHART – OUT OF CONTROL SIGNALS
 No points outside the control limits (upper or
lower).
 No run of 7 consecutive points above or below the
average line.
 No run of 7 consecutive points upward or
downward.
 No pattern with 2/3 of the points in the middle
1/3 of the control limits.
 No pattern with 2/3 of the points in the outer 2/3
of the control limits.
PREPARE A RANGE CONTROL CHART USING
THE FOLLOWING DATA:
Date of analysis Rep 1 Rep 2 Range
July 1 65.1 64.8
July 3 65.3 64.9
July 4 65.4 65.1
July 5 65.7 65.2
July 6 64.5 65.0
July 7 65.2 65.7
July 8 65.3 64.8
July 9 65.2 64.7
July 10 65.5 64.9
July 12 65.2 64.7
July 13 65.3 64.8
July 14 65.4 65.3
PREPARE A RANGE CONTROL CHART USING
THE FOLLOWING DATA:
Analysis date Rep 1 Rep 2 Ramge
July 15 65.8 65.1
July 16 65.4 65.7
July 17 64.9 65.4
July 18 64.8 65.2
July 19 65.1 65.8
July 20 65.2 65.6
July 21 64.7 64.9
July 22 65.3 65.7
July 23 64.8 64.9
July 24 65.5 65.1
July 25 64.9 64.7
July 26 65.4 65.0
July 27 65.8 65.4
PLOT ON THE RANGE CONTROL CHART THE
FOLLOWING DATA AND OBSERVE TRENDS:
Analysis date Rep 1 Rep 2 Ramge
Aug 1 64.1 64.5
Aug 2 65.2 65.7
Aug 3 64.9 64.4
Aug 4 64.8 65.2
Aug 5 65.1 65.8
Aug 6 65.5 65.8
Aug 7 64.9 65.3
Aug 8 65.1 65.7
Aug 9 65.8 65.6
Aug 10 64.5 64.1
Aug 11 65.9 65.5
Aug 12 65.4 64.2
Aug 13 65.8 65.4
PLOT ON THE RANGE CONTROL CHART THE
FOLLOWING DATA AND OBSERVE TRENDS:
Analysis date Rep 1 Rep 2 Ramge
Aug 14 64.7 65.5
Aug 15 65.2 65.8
Aug 16 64.8 65.4
Aug 17 65.1 65.8
Aug 18 65.2 65.7
Aug 19 65.1 65.8
Aug 20 64.9 65.5
Aug 21 65.1 65.7
Aug 22 65.8 65.1
Aug 23 64.5 64.1
Aug 24 65.9 65.5
Aug 25 65.4 64.2
Aug 26 65.8 65.1
 RECOVERY TEST FOR CALCIUM
TITRIMETRIC METHOD
Procedure:

1. Weigh 2 replicates from the same sample. Ash and digest

samples at the same time.
2. Place digested sample in 100 mL volumetric flasks
3. Prepare Spike Solution : Dilute 4 mL of 1000 ppm Ca with
4 mL distilled water and 2 mL acid (same conc as digestion
acid
4. Add 10mL of the spike solution to one of the volumetric
flasks and to the other, fill up to volume with water.
5. Analyze both samples using the standard method
4. Calculate % Recovery
 RECOVERY TEST CALCULATION:
 Calculate the amount of analyte added :
 multiply the concentration of the standard solution by the
dilution factor

 Concentration of Calcium standard: 1000 ppm

 Volume of standard added to spike solution: 4 mL
 Total volume of spike solution: 10 mL
 Concentration of spike solution: 400 ppm
 Volume of spike solution added to sample matrix: 10 mL
 Concentration of analyte added to the matrix is :
400*10/100 : 40 ppm
RECOVERY TEST CALCULATION:
Result of Actual Analysis:
 Spiked sample = 445 ppm
 Reference sample = 401 ppm

 ppm recovered from analysis: 44 ppm

 Actual amount added: 40 ppm

 % Recovery = amount recovered in analysis x 100

actual amount added

% Recovery = 110%
CRITERIA FOR % RECOVERY
Analyte Concentration vs % Recovery

Concentration Mean Recovery (%)

100% 98 – 102
10% 98 – 102
1% 97 – 103
0.1% 95 – 105
100 ppm 90 – 107
10 ppm 80 – 110
1 ppm 80 – 110
100 ppb 80 – 110
10 ppb 60 – 115
1 ppb 40 – 120
 RECOVERY TEST FOR CALCIUM
AAS METHOD
Procedure:

1. Weigh 2 replicates from the same sample. Ash and digest

samples at the same time.
2. Place digested sample in 100 mL volumetric flasks
3. Take 10 ml aliquot of the solution and transfer to a 100 mL
of volumetric flask
4. Prepare Spike Solution :
Take 4 mL of 1000 ppm Ca standard and dilute to 100 mL
4. Add 10mL of the spike solution to one of the volumetric
flasks and to the other, fill up to volume with water.
5. Analyze both samples using the standard method
4. Calculate % Recovery
 RECOVERY TEST CALCULATION:
 Calculate the amount of analyte added :
 multiply the concentration of the standard solution by the
dilution factor

 Concentration of Calcium standard: 1000 ppm

 Volume of standard added: 4 mL
 Total volume of spike solution: 100 mL
 Concentration of spike solution: 40 ppm
 Volume of spike solution added to sample matrix: 10 mL
 Concentration of analyte added to the matrix is :
40*10/100 : 4 ppm
RECOVERY TEST CALCULATION:
Result of Actual Analysis:
 Spiked sample = 43.8 ppm
 Reference sample = 40.1 ppm

 ppm recovered from analysis: 3.7 ppm

 Actual amount added: 4.0 ppm

 % Recovery = amount recovered in analysis x 100

actual amount added

% Recovery = 92.5%
 RECOVERY TEST WORKSHOP 1:
 Calculate the amount of analyte added :
 multiply the concentration of the standard solution by the
dilution factor (ml standard)/(ml standard + ml sample).

 Example:
 Concentration of Calcium standard: 1000 ppm
 Volume of standard added: 1.5 mL
 Total volume of sample + standard : 25 mL
 Amount added is : 60ppm
RECOVERY TEST CALCULATION:
Result of Actual Analysis:
 Spiked sample = 183.4 ppm
 Reference sample= 122.0 ppm

 ppm recovered from analysis: 61.4

 Actual amount added: 60.0

 % Recovery = amount recovered in analysis x 100

actual amount added

% Recovery = 102%
 RECOVERY TEST CALCULATION WORKSHOP 2:
 Calculate the amount of analyte added :
 multiply the concentration of the standard solution by the
dilution factor (ml standard)/(ml standard + ml sample).

 Example:
 Concentration of Vitamin A standard: 20ppm
 Volume of standard added: 0.2 mL
 Volume of sample matrix: 1.0
 Total volume of sample + standard : 1.2
 Amount added is : 20*(0.2/1.2) or 3.33 ppm
RECOVERY TEST CALCULATION:
Result of Actual Analysis:
 Spiked sample = 13.5 ppm
 Reference sample= 9.8 ppm

 ppm recovered from analysis: 13.5 – 9.8 = 3.7 ppm

 Actual amount added: 3.3 ppm

 % Recovery = amount recovered in analysis x 100

actual amount added

% Recovery = 112%
 RECOVERY TEST WORKSHOP 3:
 Calculate the amount of analyte added :
 multiply the concentration of the standard solution by the
dilution factor (ml standard)/(ml standard + ml sample).

 Example:
 Concentration of Zinc standard: 100 ppm
 Volume of standard added: 0.5 mL
 Total volume of sample + standard : 10 mL
 Amount added is :
RECOVERY TEST CALCULATION:
Result of Actual Analysis:
 Spiked sample = 25.5 ppm
 Reference sample= 21.0 ppm

 ppm recovered from analysis:

 Actual amount added:

 % Recovery = amount recovered in analysis x 100

actual amount added

% Recovery =
CRITERIA FOR % RECOVERY
Analyte Concentration vs % Recovery

Concentration Mean Recovery (%)

100% 98 – 102
10% 98 – 102
1% 97 – 103
0.1% 95 – 105
100 ppm 90 – 107
10 ppm 80 – 110
1 ppm 80 – 110
100 ppb 80 – 110
10 ppb 60 – 115
1 ppb 40 – 120
QA PROGRAM WORKSHOP
 Briefly state your plan for:
- Analyst
- Test Method
- Sample
- Standards and Reagents
- Equipment
- Environment
QA PROGRAM WORKSHOP
 Briefly state your plan for protein and fat
analysis:
- Accuracy Test
- Precision Test
- Control Chart
RECOMMENDED FREQUENCY OF
ANALYSIS OF QA SAMPLES PER ANALYSIS TYPE

Assumption: all analyst training, calibration and

other QC procedures are in place
 Blank analysis – at least once per day unless you
use a different batch of chemicals on the same
day
 Accuracy test – once a day to once a month
depending on cost of standard
 Precision test – do replicate analysis for 1 out of
5 samples (less frequent if results are good, more
frequent if results are not good)
 Control chart – once a day