Scribd
Upload
76 views30 pages

IQC - An Introduction

- Internal quality control (QC) is an important tool for ensuring quality of laboratory results and includes internal and external quality assessment - Control charts (e.g., Levey-Jennings charts) plot test results over time and establish control limits to detect errors in testing processes - The Westgard method uses multiple rules applied to control charts, including rules to detect random errors (e.g., 13s), systematic errors (e.g., 22s), and errors between control materials (e.g., R4s), improving on single-rule methods

Uploaded by

Bhageshwar Chaudhary
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPSX, PDF, TXT or read online on Scribd
76 views30 pages

IQC - An Introduction

- Internal quality control (QC) is an important tool for ensuring quality of laboratory results and includes internal and external quality assessment - Control charts (e.g., Levey-Jennings charts) plot test results over time and establish control limits to detect errors in testing processes - The Westgard method uses multiple rules applied to control charts, including rules to detect random errors (e.g., 13s), systematic errors (e.g., 22s), and errors between control materials (e.g., R4s), improving on single-rule methods

Uploaded by

Bhageshwar Chaudhary
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPSX, PDF, TXT or read online on Scribd
You are on page 1/ 30

Internal Quality Control (QC) for Medical

Laboratories: An introduction

Dr. Otto Panagiotakis and Dr. Alexander Haliassos

ESEAP – Greek Proficiency Testing Scheme

for Clinical Laboratories

http://www.eseap.gr [email protected]
• Analytical Quality Control •

The most important tool for ensuring the quality

of laboratory results through the identification and

reduction of errors

It includes two parallel mechanisms:


• Internal (intra-laboratory) quality control

• External quality control, or External quality assessment

(EQA) or Proficiency Testing (PT)


What means a value of total cholesterol 245mg/dL

reported for the analysis of an QC control?

Correct result: There is not a


value, but a range from
repeated measurements
We need a tool to compare the
reported with the expected
235 245 255 265 result.
240 250 260

It is the control chart Levey-Jennings


Control charts Levey-Jennings:

+3s
+2s
+1s
mean
-1s
-2s
-3s

Central horizontal line: expected mean value


Dotted horizontal lines: control limits (mean ± nSD)
Control charts Levey-Jennings:
Their design is based on the assumption that:
• The values resulting from previous measurements are
subject to random variation
• This variation follows a uniform (normal) distribution
How we draw the control charts?

• We select a parameter (p.ex. Cholesterol)


• we measure this parameter in a control material for
20 days using the method (assay) and the instrument
(analyzer) that we evaluate
• from those values we calculate the mean and the SD
p.ex. mean = 200 mg/dL and SD = 4 mg/dL

• subsequently, the control limits at the level of 2s, 3s


mean ± 2s: 200 ± 2(4)=200 ± 8  from 192 to 208

mean ± 3s: 200 ± 3(4)=200 ± 12  from 188 to 212


Cholesterol (mg/dL), Lot No: xxx, January 2015
216
x+3s
212
x+2s
208

mean
200

192 x-2s
188
x-3s
184
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Interpretation of control charts (1)

• The control charts reflect the distribution resulting from


the initial measurements of the control materials
• In the charts we draw the values of each measurement
• Evaluation: depending on their position in the chart
1) If the analytical procedure is correct, the new measurements
will have the same distribution with the originals:
• Extremely rare (0,3%) a value> mean ± 3s
• Unlikely (5%) a value> mean ± 2s
• Very likely (32%) a value> mean ± 1s (limits without a value)
Control Rules
Rules to decide whether a series of measurments is
under control or out of control
Control rules control limits

12s mean ± 2s

13s mean ± 3s

Single rule methods


Multiple rules methods
Interpretation of control charts (2)

2) If the analytical procedure has a problem, it increases

the probability of a value exceeding the control limits

This can happen :

• Either with the appearance of a constant error (bias)


(shifting of the mean of the distribution to or values)

• or by increasing the random error

(enlargement of the distribution)

Situation out of control / Unacceptable results


Situation Systematic Distribution
under control error (bias) enlargement
(normal)

•
The problem of erroneous rejects
The QC methods are detection systems
The detection systems have some characteristics:
The frequency of true warnings, true alarms
The frequency of erroneous warnings, false alarms
In the QC methods they are called respectively :
• Error detection probability (Ρ)
• Probability of false rejection (Ρο)
The ideal on a single rule QC method would be:
Ρ=1,0 (100%) and Ρο=0 (0%)
However, for each control rule: Ρ < 1 and Ρο > 0
A realistic goal is : Ρ=0,90 and Ρο=0,05
Rule 12s: high Ρ and high Ρο

For Ν=1 Ρο=5%


For Ν=2 Ρο=9%
For Ν=3 Ρο=14%

• The single rule QC method using the 12s rule

should only be used with Ν=1.


• If Ν=2, almost a false rejection in 10
• This is not a problem of the analytical procedure.
• This is an intrinsic problem of the QC method
and related to the selected control threshold.
Rule 13s: low Ρο and low Ρ

resulting in non detected large errors

As the control limits are extended erroneous rejections

(Po) decrease, but also P decreases

If a rule with high P is selected will have also high Ρο

Single rule QC methods have serious drawbacks

Need to find other QC methods


Multiple rules methods

They do not use a single control rule


but a combination of rules (at least 2)
Advantage: Low Po and simultaneously high P
The most well known Westgard method:
• 6 control rules
• 2 control sera (N=2) resulting to
• 2 Levey-Jennings control charts L-J, one for each
serum (control material)
• Control limits at three levels(±1s, ±2s, ±3s)
Cholesterol (mg/dL), Lot No: xxx, January 2015
216
x+3s
212
x+2s
208
x+1s
204

200

196
x-1s
192 x-2s
188
x-3s
184
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
The 6 control rules according to Westgard

12S

13S

22S

R4S

41S

10x
Rule 12S
One value (measurement) > 2s limit

+3s
+2s
+1s
mean
- 1s
- 2s
- 3s

1 2 3 4 5 6 7 8 9 10

Not a rejection but warning for a potential problem


Further control is required based on the other criteria
Rule 13S
One value (measurement) > 3s limit

+3s
+2s
+1s
mean
- 1s
- 2s
- 3s

1 2 3 4 5 6 7 8 9 10

Applied in a series for each of the two sera


Sensitive to random errors
Rule 22S
2 consequent values > the same limit of 2s

+3s
+2s
+1s
mean
- 1s
- 2s
- 3s
1 2 3 4 5 6 7 8 9 10

In the last 2 series for each serum


In the same series for both
+3s
Rule 22S
+2s
2 consequent values > the same limit of 2s
+1s
mean
+3s - 1s
+2s - 2s
+1s - 3s
mean
1 2 3 4 5 6 7 8 9 10
- 1s
- 2s
+3s
- 3s
+2s
1 2 3 4 5 6 7 8 9 10 +1s
mean
In the last 2 series for each serum- 1s
In the same series for both - 2s
- 3s
Sensitive to systematic errors
1 2 3 4 5 6 7 8 9 10
Rule R4S
The difference in value of the two sera> 4s

+3s
+2s
+1s
mean
- 1s
- 2s
- 3s
1 2 3 4 5 6 7 8 9 10

In the last 2 series for each serum


In the same series for both
+3s
Rule R4S +2s
The difference in value of the two sera> 4s +1s
mean
+3s - 1s
+2s - 2s
+1s - 3s
mean
1 2 3 4 5 6 7 8 9 10
- 1s
- 2s
+3s
- 3s
+2s
1 2 3 4 5 6 7 8 9 10 +1s
mean
In the last 2 series for each serum- 1s
In the same series for both - 2s
- 3s
Sensitive to systematic error
1 2 3 4 5 6 7 8 9 10
Rule 41S
4 consecutive values > the same limit of 1s

+3s
+2s
+1s
mean
- 1s
- 2s
- 3s
1 2 3 4 5 6 7 8 9 10

In the last 4 series for each serum


In the last 2 series for both
+3s
Rule 41S +2s
4 consecutive values > the same limit of 1s +1s
mean
+3s - 1s
+2s - 2s
+1s - 3s
mean
1 2 3 4 5 6 7 8 9 10
- 1s
- 2s
+3s
- 3s
+2s
1 2 3 4 5 6 7 8 9 10 +1s
mean
In the last 4 series for each serum- 1s
In the last 2 series for both - 2s
- 3s
Sensitive to systematic errors
1 2 3 4 5 6 7 8 9 10
Rule 10Χ
10 consecutive values at the same
side of the mean

+3s
+2s
+1s
mean
- 1s
- 2s
- 3s
1 2 3 4 5 6 7 8 9 10

In the last 10 series for each serum


In the last 5 series for both
Rule 10Χ +3s
10 consecutive values at the same
+2s
side of the mean +1s
mean
+3s - 1s
+2s - 2s
+1s - 3s
mean
1 2 3 4 5 6 7 8 9 10
- 1s
- 2s
+3s
- 3s
+2s
1 2 3 4 5 6 7 8 9 10 +1s
mean
In the last 10 series for each serum
- 1s
In the last 5 series for both - 2s
- 3s
Sensitive to systematic errors
1 2 3 4 5 6 7 8 9 10
Flowchart according to Westgard

values

control

no
12s Situation under control - Series Accepted

yes no

no no no no
13s 22s R4s 41s 10x

yes yes yes yes yes

Situation out of control - Series Rejected


Differences of Internal and External QC
Internal QC
• performed daily in the laboratory
• uses samples (control materials) of known concentration
• it is always required
External QC
• performed periodically (weekly, monthly …)
• uses samples (control materials) of unknown concentration
• useful in conjunction with the internal Q C
External QC does not replace the Internal QC
• the inter-laboratory comparisons are infrequent
• the results are reported deferred (not in real time) and

therefore it is not possible the immediate intervention


with corrective measures
• even if the performance is satisfactory, it assures the
proper functioning of the laboratory only on the day of
the inspection (participation)
These programs do not lead to quality improvement of
the laboratory if it is not performed daily the internal QC.

You might also like