Transcript Clinical Chemistry Chapter 3

MLAB 2401: Clinical
Chemistry
Quality Control, Quality
Assessment and Statistics
1
Quality Assurance/Assessment (QA)
 An
all inclusive / comprehensive system
monitoring the accuracy of test results where
all steps before, during and after the testing
process are considered. Includes pre-analytic,
analytic and post analytic factors
 Essentials include commitment to quality,
facilities, resources, competent staff, and
reliable procedures, methods and
instrumentation
 Provides a structure for achieving lab and
hospital quality goals
2
Quality Control (QC)
 QC
systems monitor the analytical process; detect
and minimize errors during the analysis and prevent
reporting of erroneous test results.
 It uses statistical analysis of test system data
 Requires following published rules

Westgard Rules
3
Types of QC
Internal
 Daily
 Establishment of
reference ranges
 Validation of a new
reagent lot and/or
shipment
 Following instrument
repair
External
 Proficiency testing





Determination of laboratory
testing performance by means
of intralaboratory comparisons
CAP, CLIA, The Joint
Commission requirement
Must be integrated within
routine workload and analyzed
by personnel who are running
the tests.
Ongoing evaluation of results
to correct for unacceptable
results
Used to access employee
4
competency
Pre-Analytical & Analytical Causes of Error
5
Post- Analytical Causes of Error
Incorrect reference values
 Physician not notified of a panic or critical
value
 Incorrect interpretation of lab results by
physician
 Incorrect data entry of lab result

6
Introduction to Statistical Analysis

When evaluating laboratory results, how do we determine what is
normal or acceptable? In other words: What is “normal” or
“OK”?

When does a laboratory test result become “weird” or “abnormal” ?
When do we become uncomfortable with a result?

At some point we have to draw a “line in the sand” … on this side
of the line you’re normal … on the other side of the line you’re
abnormal. Where and how do we “draw the line” ?

Answer: Statistics are used to determine the lines of ‘normal’
and ‘acceptable’.
7

Statistical Concepts

Statistics is a (science of )branch of mathematics that collects,
analyzes, summarizes and presents information about
“observations.”

In the clinical lab, these “observations” are usually numerical test
results

A statistical analysis of lab test data can help us to define
 Reference ranges for patient’s (normal and abnormal)
 Acceptable ranges for control specimens ( “in” and “out” of
control)
8
Measures of Central Tendency



Mean (x̄) - the mathematical average of a group of numbers,
determined by adding a group of numbers (events) and dividing the
result by the number of events
Median - determined as the ‘middle’ of a group of numbers that
have been arranged in sequential order. That is to say, there are an
equal number of numbers on either side of the ‘middle’ number. In
an odd # of observations, it is the middle observation. In an even #
of observations, average the two middle values.
Mode - the number that appears most frequently in a group of
numbers. There can be more than mode, or none at all.
9
Gaussian/Normal Distribution
• All values are
symmetrically
distributed around the
mean
• Characteristic “bellshaped” curve
• Assumed for all quality
control statistics
10
11
Accuracy and Precision

The degree of fluctuation in the measurements
is indicative of the precision of the assay.
 Precision-refers
to the ability to get the same (but not
necessarily ‘true’) result time after time.

The closeness of measurements to the true
value is indicative of the accuracy of the assay.
 Accuracy
- An accurate result is one that is the ‘true’
result.
12
Precise and Accurate
•
•
Systematic
Error
Random Error
Systematic error
Systematic change in the test system
resulting in a displacement of the mean
from the original value
 Systematic error of an analytic system is
predictable and causes shifts or trends on
control charts that are consistently low or
high

15
Causes of Systematic Error









Change in reagent or calibrator lot numbers
Wrong calibrator values
Improperly prepared reagents
Deterioration of reagents or calibrators
Inappropriate storage of reagents or calibrators
Variation in sample or reagent volumes due to pipettor
misalignments
Variation in temperature or reaction chambers
Deterioration of photometric light source
Variation in procedure between technologists
16
Random Error

Imprecision of the test system causing a
scatter or spread of control values around
the mean
17
Causes of Random Error






Air bubbles in reagent
Improperly mixed reagents
Reagent lines, sampling, or reagent syringes
Improperly fitting pipette tips
Clogged or imprecise pipetter
Fluctuations in power supply
18
Bias

Bias – the amount by which an analysis
varies from the correct result.
 Example,
If the Expected Value is 50 units,
and the result of an analysis is 47, the bias is
3 units.
19
Statistical Formulas
 Standard

Deviation (SD)
Is a mathematical expression of the dispersion of a
group of data around a mean.
SD 
  x  x
2
 n  1
20
Standard Deviation :
SD 
  x  x
2
 n  1
n
=
the number of observations (how many numerical values )
Σ
=
the sum of … in this case, the sum of
x
=
X
= the value of each individual observation
x
 x
2
the mean value
The Standard Deviation is an expression of dispersion … the greater the
SD, the more spread out the observations are
21
Standard Deviation and Probability

For a set of data with a
normal distribution, a
value will fall within a
range of:
 +/- 1 SD 68.2 % of the
time
 +/- 2 SD 95.5% of the
time
 +/- 3 SD 99.7% of the
time
22
Statistical Formulas

Coefficient of Variation (CV)



Indicates what percentage of the mean is represented by the
standard deviation
Reliable means for comparing the precision or SD at different
units or concentration levels
Expressed as a percentage

CV% =
Standard deviation X 100
mean
23
Coefficient of Variation (CV) %
Analyte:
FSH Concentration
SD
CV
1
0.09
9.0
5
0.25
5.0
10
0.40
4.0
25
1.20
4.8
100
3.80
3.8
•The smaller the CV, the more reproducible the
results: more values are closer to the mean.
•Useful in comparing 2 or more analytical
methods
•Ideally should be less than 5 %
24
Establishment of a QC System

Two or three levels of control material used
A
control is a material or preparation used to monitor
the stability of the test system within predetermined
limits
 Measure of precision and reproducibility
 Purpose:
verify the analytic measurement
range of instrument for a specific analyte
25
Establishment of a QC System

Control material matrix should resemble actual
specimens tested
 Lyophilized/liquid
 Assayed
 Mean calculated by the manufacturer
 Must verify in the laboratory
 Unassayed
 Less expensive
 Must perform data analysis in house
26
Establishment of a QC system

Collecting data
 Run
assay on control sample & manually
enter control results on chart

One chart for each analyte and for each level of
control

27
Establishment of a QC system

Collecting data
 Many
modern chemistry analyzers have
computer program that maintains the QC log.

i.e Dade Dimension
28
Collecting Data for QC

Charting techniques
 Levey
Jennings chart is a graph that plots QC
values in terms of how many standard
deviations each value is from the mean

29
Use of Standard Deviation

Once you have determined the standard
deviation, must use the information to
evaluate current/ future analysis.

Most labs make use of ± 2 SD or 95%
confidence limit. To put this into a
workable form, you must establish the
range of the ± 2 SDs
30
So, how do we determine the range of
acceptable results ?

Scenario
 Mean
of group of control values = 104 mg/dL
 Standard Deviation = ± 5 mg/dL
 Determine the Range of ± 2SD; (which will allow you
to evaluate acceptability of performance of the control
on subsequent days.)

Is a control value of 100 mg/dL acceptable?
31
Shifts and Trends

Shift
 QC
data results are distributed on one side of
the mean for 6-7 consecutive days

Trend
 Consistent increase or decrease of QC data
points over a period of 6-7 days
32
But what if your control specimen is “out of
control?”

“Out of control” means that there is too much dispersion in
your result compared with the rest of the results

This suggests that something is wrong with the process
that generated that observation

Patient test results cannot be reported to physicians when
there is something wrong with the testing process that is
generating inaccurate reports

Remember … No information is better than wrong
information
33
Westgard System
Is 1
Is 1
Are 2
Is SD
Are 4
Are 10
control
control
consecutive
consecutive
controls
difference
> 2 SD? Yes > 3 SD? No > 2 SD No betwe en No
controls No
controls No





on same
> +/- 1 SD?
on same 
any 2
12 S
13 S
side of
side of
controls
mean?
4
mean?
> 4?
2S
No
Yes
22 S
10X
R4 S
Yes
Reject
Accept
Yes
Yes
run
Yes
Reject
run
run
Test
Reject
Reject
Reject
Report
rem aining
run
Test
run
run
rules
Results
rem aining
Test
rules
Test
rem aining
rem aining
rules
rules
Violation
indicates
random
error
Violation
indicates
system atic
error
Violation
indicates
random
error
Violation
indicates
system atic
error
Accept
run
Report
Results
Violation
indicates
system atic
error
34
But what if your control specimen is “out of control?”

Corrective methods
Things that can go Wrong
Corrective Action
Instrument malfunction
Identify malfunction and fix
Reagents: preparation,
contamination, volume
New reagents
Tech error
Identify error and repeat
test
Control specimen is old or
prepared improperly
Use new control
35
QC terms


AMR= Analytical Measurement Range
 Range of analyte values that a method can directly measure on
the specimen without any dilution, concentration or other
pretreatment
CRR= Clinical Reportable Range
 Range of analyte values that a method can report as a
quantitative result, allowing for specimen dilution, concentration,
or other pretreatment used to expand the direct AMR.
36
System Flags

Delta check
 Comparison of individual patient results throughout
the day or week with computer detection of changes
from earlier individual patient results
 Helpful to identify pre-analytical errors
Test
Change
Time Frame, hours
Sodium, adult
7%
24
Creatinine
50%
72
Hemoglobin
3.0 g/dl
48
Other
Transfusion/ Bleeding?
37
Establishment of Reference Ranges

Reference ranges – the ‘normals’
 The normal or expected value for patients.
 Are defined as being within +2 Standard
Deviations from the mean
 A large sampling of clinical normal
representatives.

Each lab must establish its own reference
ranges based on local population
38
Establishment of Reference Ranges
Factors affecting reference ranges:











Age
Sex
Diet
Medications
Physical activity
Pregnancy
Personal habits (smoking, alcohol)
Geographic location (altitude)
Body weight
Laboratory instrumentation (methodologies)
Laboratory reagents
39
Test results

Critical values and read back of results
 Values that indicate a life-threatening situation for the
patient
 Require notification of the value to nurse or physician
 Nurse or physician must “read back” the results to the
technician
Test
Results
Decreased
Significancelow results
Results
Increased
Significance- high
results
Glucose, adult
< 50 mg/dL
Brain damage
>500 mg/dL
Diabetic coma
Sodium
<120 mEQ/L
Paralysis,
arrhythmias
>160 mEQ/L
Dehydration, heart
failure
40
References



Astles, J. R., Stang, H., & Alspach, T. (2013, September).
CLIA requirements for proficiency testing: the basics for
laboratory professionals. MLO, 45(9), 8-15.
Bishop, M., Fody, E., & Schoeff, l. (2010). Clinical Chemistry:
Techniques, principles, Correlations. Baltimore: Wolters
Kluwer Lippincott Williams & Wilkins.
Sunheimer, R., & Graves, L. (2010). Clinical Laboratory
Chemistry. Upper Saddle River: Pearson .
41