Gaussian Distribution,

Mean, SD, CV and Range

Labs for Life Project

QC Training
12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 2
12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 3
 What is Gaussian distribution in the
analytical phase in a lab?

Normal distribution is exhibited by well-preserved

biological material on repeated examinations.
This property is used in the statistical quality
controls.
Statistical Quality Controls (SQCs) are used to verify
the on-going performance of analytical systems

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 4

Gaussian (Normal)
Distribution
• Bell shaped curve
F • Symmetrical
r • Points cluster around
e the mean
q
u
• More points lie closer
e to the mean
n • Farther away from the
c mean, the fewer data
y points exist
• Measures of central
Measurement tendency are the same
number
12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 6
Standard Deviation and
Probability
99.7%
95%
68%
2.5% 13.5% 34% 34% 13.5% 2.5%

When data exhibit Gaussian (Normal) distribution, we can

predict the likelihood of data falling within specific
percentages of the mean on the Gaussian curve.
12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 7
 Stable
Monitoring Potentially Unstable Analytical System
System

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 8

Using the
Monitoring
System
Gaussian is the Key

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 9

Gaussian Distribution is the Key to
Statistical Quality Control

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 10

 Thus in SQCs, there is a central point (Mean) around
which the data points arrange themselves with some
amount of variability (SDs).

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 11

Demonstrates accuracy as its mean is located in the center
of the curve and is stable
Demonstrates precision as all the data is grouped together
as represented by ‘well rounded bell curve’ and the
dispersion or SD is stable
12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 12
 Accuracy maintained but precision affected
as data is more spread-out, resulting in a
‘flatter, wider Gaussian’.
IMPRECISION!!!
OLD

NEW

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 13

 NEW

OLD

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 14

Calculating and quantifying
these errors.
1. To understand the location of central point and any deviations from
it we need to calculate and track the mean.
2. To understand the degree of dispersion and thus quantify random
errors we need to calculate and track the Standard Deviation (SD)
3. To understand the SD in a uniform manner we need to convert it
into a percentage called Coefficient of Variation (%CV)
4. To understand the acceptability of variations and shifts, we need to
calculate the ranges

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 15

Recap: Mean ( x )

◦ Average of a set of values

◦ Primary indicator of accuracy
◦ Measure of systematic error (error in a given
direction)

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 16

Recap: To Calculate the Mean
x = xi

n Xi = individual value
n = number of individual values

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 17

Recap: Example

For the data set containing 7 numbers

{2, 5, 9, 3, 5, 7, 4},

2+5+9+3+5+7+4
= 35/7
=5 is the mean

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 18

Let’s understand SD and %CV
2 statistical
measures of
Variation

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 19

Standard Deviation (SD)

Standard Deviation (SD) = is a measure of how much the

data VARIES around the MEAN
◦ Used to measure dispersion/scattering of a group of
values around a mean
◦ Primary indicator of precision
◦ Measure of random error ( ±)

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 20

 Statistical  x  x 
2

Formulas SD 
n  1 

n = the number of observations (how many numerical values )

Σ = the sum of … in this case, the sum of

x = the mean value

X = the value of each individual observation

The Standard Deviation is an expression of dispersion … the greater the

SD, the more spread out the observations are

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 21

  x  x 
2

SD 
SD Example n  1 
For the data set containing 7 numbers
2, 5, 9, 3, 5, 7, 4: Find the mean: (35/7= 5)
Subtract mean from all the values:
Square the differences
Add the differences
Divide by 6 (n = 7, n-1=6)
Find the square root

SD = 2.4

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 22

 Coefficient of Variation (CV)
CV is SD expressed as a proportion of the mean
CV (%) = (SD / Mean) x 100

Utilizing CV allows you to quantify the SD as

percentage of the MEAN value

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 24

Why CV%
Percentages are not influenced by the
denominator (Mean)
Hence can be used in comparisons
◦ Different methods
◦ The same method at different clinical decision
levels
◦ Precision of methods over time
◦ Precision of different Analytes.

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 25

 Why CV%
Different methods

PERFORMANCE SPECIFICATIONS OF 2 TESTS. WHICH WILL YOU

CHOOSE?
Analyte Method CV% L1 CV% L2

Glucose Hexokinase 1.5 1.1

Glucose Oxidase 3 2.4

Hb SLS 1 1.1

Cyan 0.7 0.6

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 26

Why CV%
The same method at different clinical decision
levels

CV IS GREAT STATISTICAL EQUALIZER

QC Level SD Mean CV%
Level-1 2 100 2
Level-2 3 200 1.5

IT IS INDEPENDENT OF DENOMINATOR

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 27

Why CV%?
Trend Analysis (Different Analytes, Performance
over time)

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 28

RECAP-Precision

Coefficient of Variation - The great statistical equalizer

◦ Measure of relative variability
◦ Expresses the SD in percent
◦ Comparison of overall precision, since standard deviation
typically increases as the concentration of the analyte
increases
◦ Estimates the SD at the concentration level of interest
◦ Provides an estimate of method precision

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 29

Range (of Control Chart)
Once the mean and the Standard Deviation are
understood, the range of acceptability can be assigned.

Using the SD, a chart can be developed to monitor the

daily control values.

Usually the range of acceptability is ±3SD. Statistically

this covers 99.7% of the expected values.

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 30

EXAMPLE, Mean =190.5 , SD = 2
The range of acceptability is ± 3 SD

Standard Ranges % of data points

Deviations
±1 SD 188.5 - 192.5 68%

±2 SD 186.5 - 194.5 95%

± 3 SD 184.5 - 196.5 99.7%

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 31

What does that mean??
It means: That if your Analytical
System is Stable,
68% of your data-points will be
between 188.8-192.5
95% (another 27% in addition to
the 68%) of your data points will
be between 186.5 and 194.5
99.7% (another 4.7% in addition
to the 95%) will fall between 184-
196.5

Any data point going below 184 or above 196.5 is an outlier

Any shift in the above percentages is an outlier

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 32

 Label the Data Points

Control Limit - the

defined limits or ranges -3SD +3SD
expected due to random Contro Contr
l Limit ol
variation of the data
Limit
points -- beyond those
limits, some course of
action involving
investigation and
troubleshooting are
taken.

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 33

Recap: Data points within specified Standard
Deviations in a Gaussian Distribution.

99.7%
95%
68%
2.5% 13.5% 34% 34% 13.5% 2.5%

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 34

 THANK YOU

12/10/24 QC TRAINING - LABS FOR LIFE PROJECT 35