QUALITY CONTROL

Quality Management  Because of the fluctuations of electricity, you

 Quality Control will have a background fluctuations (~~~). There
- focuses on the results
is a measure on whether your instruments will
 Quality Assurance
- all methods, parameters, measure that as a signal of analyte or just a
procedures (Standard Operating background noise.
Procedures) that we will do, follow  Referred to as the limit of detection (LoD)
and implement when running a  Measure of the precision of the test, or the
laboratory. minimum amount detectable within a given
Quality Control system
 How reliable is the result? 2. Specificity
 What does this value mean in relation to the  Is the ability of an analytical method to
physiologic health of the patient? measure only the analyte of interest
 Process of ensuring that analytical results are  used to describe the degree of cross-reactivity
correct by testing known samples that and interferences in a test system.
resembles patient samples. Example:
 It is one component of quality assurance Sodium (Na)
system, and is part of the performance  How sure are you that the machine is only
monitoring that occurs after a test has been measuring Na?
established.  There could be interferences on the matrix of
your sample.
Parameters of Quality Control 3. Accuracy
 Things that need to take into consideration  Is an indication of how close the answer
when managing a laboratory in terms of quality obtained lies to the “true” value.
control  Is not defined by statistics, but by comparison
1. Detection Limit of the measured value to the “known” value
2. Specificity obtained from a reference material.
3. Accuracy 4. Precision
4. Precision/Reproducibility 1. Indicates how close the single values are to
5. Practicability one another and is expressed by the standard
6. Reliability deviation.
7. Diagnostic Sensitivity 2. The smaller the standard deviation, the more
8. Diagnostic Specificity precise we are.
3. Ability of an analytical method to give repeated
1. Detection Limit results on the same sample that agree with one
 It is the ability of an analytical method to another.
measure the smallest concentration of the
analyte of interest.
Example:
In urine analysis, we have pH and proteins to analyze.
◉ Sensitivity or analytical sensitivity
- mistaken synonym of detection limit
- The response of the instrument in slight
changes in the concentration of the anayte IMPORTANT:
Example: in Electrochemical system,  For each concentration, a mean and standard
deviation is calculated.
 QUALITY CONTROL
 Coefficient of variation (CV), which is the SD in 8. Diagnostic Specificity
percent of the mean, is a measure of the  Ability of the analytical method to detect the
precision. proportion of individuals without the disease
 Reflects the ability of the method to detect true
negative with very few false positive
 Confirmatory test requires high specificity to be
certain of the diagnosis.

 The smaller the SD, the more precise we are.

There are certain distances from the mean where we
can say that our measurements are close to the mean
value, and they are not scattered around.
SD tells us that 99.73% of the results should not go
beyond 3 sigma. DESCRIPTIVE STATISTICS
5. Practicability (RECAP)
 It is the degree by which a method is easily Measures of Center:
repeated o Mean – average
 AKA replicability/reproducibility o Median – middle
6. Reliability o Mode – higher counts
 It is the ability of an analytical method to 1. Mean
maintain accuracy and precision over an  Is a measure of central tendency.
extended period of time during which  Associated with symmetrical or normal
equipment, reagents and personnel may distribution
change.
 AKA robustness ꜫx
MEAN=
7. Diagnostic Sensitivity n
 It is the ability of the analytical method to
2. Median
detect the proportion of individuals with the
 Value of the observation that divides
disease
the observation into two groups, each
 Indicates the ability of the test to generate
containing equal number of observation
more true-positive results and few false
 Midpoint of distribution
negative
 is the middle of the data after the data
 Screening tests require high sensitivity so that
have been rank ordered
no case is missed
 is the value that divides the data in half
SENSITIVITY=
the number of diseased individuals with a positivetest
Total number of diseased individuals tested
x 100
 QUALITY CONTROL
3. Mode
 Most frequent observation V =SD 2
 Used to describe data with two centers
(bimodal)
Example.
A. In the sample 3, 4, 5, 5, 5, 6, 7.
B. The dataset 3, 4, 5, 5, 5, 6, 7, 8, 9, 9, 9
Example: 5, 4, 6, 5, 3, 7, 5,

Spread:
o Range
o Standard deviation (SD)
o Coefficient of variation (CV)
1. Range
 largest value in the data minus the smallest
value, which represents the extremes of data
one might encounter
Example:
For the sample 5, 4, 6, 5, 3, 7, 5,
ANSWER: 7-3 = 4
2. Standard deviation
 Is a measure of the dispersion of values ANSWER
from the mean. SD = 1.29
 Helps describe the normal curve
 A measure of the distribution range 3. Coefficient of Variation (CV)
 Most frequently used in measure of  is a percentile expression of the mean, an index
variation of precision
 used to determine where errors happened

SD
CV = X 100
Mean
1st: Determine the variance
Variance: SHAPE
 Represents the difference between each value Gaussian distribution
and the average of the data.  Occurs when the data set can be accurately
 A measure of variability described by the SD and the mean.
 Standard deviation squared  Population probability distribution is symmetric
about the mean.
 QUALITY CONTROL
 “bell curve”.  Allows identification of analytical errors within
 Total area under the curve : 1.0 or 100% one week cycle.
2. Interlab QC (External QC)
 Involves proficiency testing programs that
periodically provide samples of unknown
concentrations to participating clinical
laboratories
68-95-99 Rule  Important in maintaining long term accuracy of
the analytical methods
 Used to determine state of the art
interlaboratory performance
 The College of American Pathologist (CAP)
68% of all measurements should fall within area proficiency program is the gold standard of
clinical laboratory external QC program.
Conduct of the External QC
 A series of unknown samples are sent to the
laboratory from the reference laboratory or
authorized program provider
95% of all measurements should fall within area
 Unknown samples must be tested by the
laboratories who regularly perform analysis of the
patient specimen using the same reagents &
equipment for actual patient specimens, and the
results are submitted to the program provider, as
soon as every analysis is done.
99% of all measurements should fall within area  Analysis of the unknown samples should be treated
The “68–95–99 Rule” summarizes the above relationships like atient specimen to determine the true essence
between the area under a Gaussian distribution and the SD. of accuracy
INTERPRETATION of Proficiency testing
Bi-modal Distribution  Each analyte has a define performance criteria
 The “bi” in bimodal distribution refers to “two” (example, +/- 3SD for peer mean), where
and modal refers to the peaks. laboratories using the same method are evaluated
 The two peaks in a bimodal distribution also by comparing within the group
represent two local maximums; these are points  In external QC, difference of greater than 2SD in the
where the data points stop increasing and start results indicates that a laboratory is not in
decreasing. agreement with the rest of the laboratories
Quality Control included in the program
Look at methods so that we produce a  If incase the clinical laboratory failed to identify or
confident inducing result. resolve an error or discrepancy in the test process,
1. Intralab QC (Internal QC) the facility is at risk of continuous operation and
 Involves the analytes of control samples may be recommended for closure.
together with the patient specimen Rationale of the External QC/ Proficiency testing
 Important for daily monitoring of accuracy & Ultimate goal:
precision of analytical mtds.  To ensure our clinicians that patient results are
 Detects both random & systematic errors in accurate
daily basis
 QUALITY CONTROL
 Allows each laboratory to compare and  These are expected values represented by intervals
evaluate test results/outcomes with those of acceptable values with upper and lower limits
laboratories that use the same method.  Expected control values are within the desired
control limits , the clinicians are assured that the test
results are accurate & precise
Objectives of QC:  Control limits: calculated from the mean & SD
1. To check the stability of the machine Linear Regression Analysis
2. To check the quality of reagents  a technique that attempts to link 2 correlated
3. To check technical (operator) errors variables x & y in the form of a mathematical
Gives us confidence of our results. equation, such that given value of one variable the
Control Solutions (QC Materials) other may be predicted.
 General Chemistry assays used 2 levels of control  correlation coefficient (r) : is a measure of the
solutions, while immunoassays uses 3 levels strength of the relationship between the two
 Establish statistical quality control on a new methods.
instrument or a new lot numbers of control
materials, the different levels of control material Comparison of Method
must be analyzed for 20days.  involves measuring patient specimens by both
 For highly précised assays, ( with the CV less than an existing (reference) method and a new (test)
1%) such as blood gases, analysis for 5 days is method
adequate.  The data obtained from these comparisons
consist of two measurements for each patient
specimen.
 It is easiest to visualize and summarize the
paired-method comparison data graphically

 values cover medical decision points

 similar to the test sample
 controls are usually available in high, normal,
and low ranges
Compare unknown sample to the controls
Characteristics of an Ideal QC Material
 Resemble human sample
 Inexpensive & stable for long periods
 No communicable disease
Correlation Coefficient ( r )
 No matrix effects/known matrix effects
 is a measure of the strength of the relationship
 With known analyte concentration (assayed control)
between the two methods.
 Convenient packaging for easy dispensing and
 can have values from -1 to 1, with the sign
storage.
indicating the direction of relationship between
Control limits (Control Values)
the two variables.
 QUALITY CONTROL
 Positive r: both variables increase and decrease
together
 Negative r indicates that as one variable increases,
the other decreases.
 r2=0 indicates no relationship,
 whereas r2 = 1.0 indicates a perfect relationship.