post mid 3.8 - 5 both slides and manual Flashcards
Westgard rules are rules that define
specific limits for control values
If a control value violates
a rule by exceeding the limits, an error in X has been detected and the results on
patient samples analyzed concurrently cannot X
measurement
be released to the physician
once problem is fixed what is the last step
Patient and controls would be rerun to ensure accurate and precise results
To interpret the Westgard shorthand
the first number is the number of :
The next number refers to the:
-quality control results that violate the rule
- s level that is being violated
1-2s
what is it
type of violation
what if there are other issues
-One control value exceeds the mean by more than 2s but less than 3s.
-Warning
-If there is an error or violation detected the technologist must repeat the QC for that
level or run and perform troubleshooting corrective actions
1-3s
One control value exceeds the mean by more than 3s in either the upward or downward direction.
(Random Violation)
2-2s
what is it
2 appications
two consecutive control values exceed the mean by more than 2s. These two controls must be consecutive and lie in the same direction from the mean (upward or downward).
(Systematic Violation)
2 applications to this rule:
Within run (same day) and across run (day to day
R-4s
The total difference between two levels of controls within a single run is 4s or greater.
Applied only with the current run between two control levels.
(Random Violation)
4-1s:
what is it
type of violation
application 2
Four consecutive control values exceed the mean by more than 1s.
(Systematic Violation)
These 4 controls must be consecutive and lie in the same direction from the mean.
2 applications:
* Across control material (day-to-day)
* Within control material (same day with 2 or more levels)
4 consecutive control values outside 1SD on the same side of the mean
10x:
Ten consecutive control values exceed the mean in the same direction.
(Systematic Violation)
2 applications:
* Across one level of control material (day-to-day)
* Within control materials (Level 1, 2, and 3 control results in combination
1-2s rule is a flag that indicates a possible change in relative
accuracy and precision
The justification for using the 1-2s rule as a flag and not a violation is based on the
95% confidence limits
1 in every 20 assay values will be
outside of our 95% confidence limit. Chances are, the next time we run our control, its
value will be within the 95% confidence limit.
T or F
It is common lab practise to repeat the control when there is no data from the previous day to rule out other violations.
T
The violation of the 1-3s rule is always reason for rejection bc
As mean +/ 3s
takes in 99.7% of
all data, we can be highly certain that an error has occurred in measurement.
2 advantages of westguard rules
-Distinguishes between random and systematic error
-Westgard rules can indicate the direction that the investigation of a systematic error
should take
random error is one that is (1)
in LJ vs WG
indeterminable
LJ-a random error is seen as a control value which is significantly different from the other values
-1-3s and the R-4s rules indicate a random error
A systematic error is an error that is (2)
in LJ vs WG
predictable and identifiable
LJ=trend or a shift
WG= 2-2s, 4-1s and the 10x r
Cumulative Charts (Cusum) sensitive means to detect:
systematic error
Corrective Actions
Any violation of the Westgard rules requires further investigation
Generalized Flow Chart for Corrective Actions using Multirules
for 1-2s
step 1,
2 random error:
3 Systematic error
4 final step
5 if nothing works
-for 1-2s if you have prior data to analyze you can just do that, otherwise a repeat of QC is done
step1-is a rejection occurred analyzed the control sample that was out, for human errors like the correct bottle, if only one QC is out then it is not the bottle
step 2-for random repeat QC and accept QC if repeat is ok.
step 3-for systemic Troubleshoot and identify the source of error
step 4: Repeat the QC after troubleshooting
Step 5: If the control value is still out of range, another laboratory professional with more technical
experience may be able to help. The last step in this series of corrective actions should be to obtain
help from an expert, usually a manufacturer’s technical representative. Most manufacturer’s have a
hotline number
if it is a systematic error what 3 things do we look at
- Problem with QC vial (end of bottle, expired, new lot, wrong vial) – repeat using fresh vial of QC
- Instrument problem (humidity, debris, tubing, probe, maintenance) – correct the instrument problem
- Reagent problem (not mixed properly, expired, bubbles in reagent, new lot) – correct the
reagent problem and calibrate the reagent.
Mean of Normals (Application of Bull’s Algorithm)
what is it
what should remain constant
type of errors detected
when is it most often used
-In this particular QC tool, the average of all the patients’ results (at least 250) that fall within the normal reference range is calculated.
-The mean target value should remain constant
-detects only systematic erroors
-This QC method is often applied
in the hematology laboratory, in particular, to the red cell indices, MCV, MCH and MCHC
Maintenance refers to
the activities associated with keeping laboratory instrumentation in good working order.
Preventative maintenance
-objective
-incorporates those same activities with perhaps a few more specific ones related to an instrument.
-object is to decrease the chances of that instrument breaking down and having to call in a service
-Preventative maintenance should be part of the daily routine of any laboratory personnel
Some of the current models of instrumentation are programmed to do self-checks. These may
include
electrical checks, flushing of valves, priming of reagents, analyzing a control, etc.
Material management
the acquisition and utilization of supplies to ensure both availability and cost effectiveness
Inventory Control involves:
(4)
what are standing orders
- Ordering supplies in sufficient quantities
- rotating stock
- checking of stock levels: want max amounts
standing orders: constant used, or lower shelf life items that are ordered based on an an established scheduled - minimize the cost, not to much or too little ordered
When lot numbers of control material
change, it must be analyzed
to obtain new target values (means) for the constituents (all the tests or assays being run using that control) before incorporating it in the daily routine for any method
effective inventory control maintains X and minimizes Y
quality assurance
one of the analytical variables
An important part of quality assurance involves the analysis of the new method to ensure its X as well as for its ability to X
accuracy and precision
correctly diagnose disease
non-statistical test for linearity and several statistical tests
- tests for significance: precision and accuracy or reliability
- correlation and
- diagnostic method evaluation
Linearity checks
- preliminary checks on the accuracy of the new method
-involves assaying a series of increasing concentrations of standards with the new method
-include concentrations at the high and the low end of the manufacturer’s listed range of linearity
-If the method is linear, it should hold true when specimens are diluted
-Do the data points produce a line with a 45 degrees
Tests for Significance 3
Probability
Precision: F-Test
Accuracy: Student t-Test
Probability
Probability is a means of expressing the prediction of outcome, or the chances that something
will happen
Tests for significance are used to compare
the accuracy and precision of two methods
If we are considering bringing a new method into the lab, we need to evaluate its reliability
against X or Y
he old method or against a reference method
We compare the4 things of the 2 methods
linearity, accuracy, precision and correlation coefficient of the two methods
X % zone is known as the significance level
5
which allows us to make a decision as to
whether or not the two sets of data are the same
We decide on the similarity of the two methods
by using a calculation involving the probability that
the two normal distribution curves representing the data are the same
Null Hypothesis:
there is no significant difference
between two sets of data. The 2 sets of data are close enough to satisfy
our 95% (P0.05) confidence limits.
Alternate Hypothesis
here is a significant difference between two sets of data and
their similarities are not within the allowable 95% (P0.05) limits.
P value:
probability of being wrong when a test statistic proves the null hypothesis.
Our goal is to prove the X hypothesis
null
cut-off point for proving our null
hypothesis which is selected as the degrees of freedom
(n-1)
F-Test
-used to compare the precision of a new method to the old (or reference) method
compares the variance of two methods and judges whether the two variances are essentially the same.
Student t-Test
used to compare the accuracy of a new method or instrument to a reference method or to the old method
This is done by comparing the means of the two methods
Correlation Coefficient
compare the results of one method to those obtained by a second method.
It determines the relationship between any two items, provided the items have a numerical value.
formula for the correlation coefficient is based on the formula for linear regression by least
squares
The closer the “r” value is to 1, the better the correlation.
A positive “r” value indicates that the line slopes upward to the right.
A negative “r” value indicates that the line slopes downward.
Acceptable correlation produces a 45 angle
and a slope of ±0.92 – 1.08
Diagnostic Method Evaluation
new method also be evaluated for its ability to correctly diagnose disease
true positive, TN, FP, FN
Sensitivity of a Test
If we can calculate 100% sensitivity then we have 0 false negatives.
Sensitivity is usually expressed to one decimal place.
positive controls
detects low amounts
Specificity of a Tess
e. If we have 100% specificity, we have 0 false positives. Specificity is usually
expressed to one decimal place.
Predictive Value of a Positive Test
ability of a test to correctly diagnose disease in the population being tested.
It asks the question, what is the frequency of diseased patients out of positive tests?
compares the prevalence of the disease with the sensitivity and specificity of the test
If a test has 100% predictive value, we have zero false positive results
Post-Analytical Checks
verifying quality in those variables that can have an impact on test results after the testing of patients’ samples is complete
These post analytical checks are the final
review before results are released outside the lab and serve as an important aid to the other
internal quality assurance methods used
Clinical Correlation
simple check of the patient’s test results to the clinical diagnosis
Correlation with Other Laboratory Tests
e, errors can be identified and corrected before the report is
released to the patient’s file
Intralab Duplicates monitor for which errors?
dividing the same specimen into two aliquots and having each aliquot analyzed for all the same tests ordered. The differences between the duplicates can
be plotted on control charts to aid in monitoring.
Duplicates obtained from more than one method monitor both random and systematic
error.
Delta Checks
patient has had the same lab tests done on more than one occasion, the delta checks or
variations between the previous test results can be compared to see if they fall within the limit
allowed for normal changes.
Limit Checks and Critical Values
checked beyond the reference ranges (impossible or
absurd results) or incompatible with life (critical results)
caused by clerical errors, short sampling or clot formation, improper identification of fluid being analyzed or improper specimen collection
Mean of Normals
Variation demonstrates what error
mean of the normal patients’ values can
be calculated
A variation in this mean will demonstrate a systematic error but will not show a random error
not a substitute
Reporting Mechanisms
reporting of results to physicians has a direct outcome on the care of the patient. Areas
where problems can be found in reporting that can have an effect on patient outcomes
Record Keeping
refers to the maintenance of records related to patient results
importance of good record keeping is to be able to refer back to previous results and so
check the patient’s history upon re-admittance or transfer to another healthcare facility
Test Completion Checks
This check is to verify that all tests ordered are completed before releasing the results
Reflex Testing Protocols
refers to the automatic “reflex” of one (1) result
triggering another test to be performed.
external QC
External Agencies and Programs
To ensure more complete objectivity, laboratories will register in external QA programs.
This may also be a requirement of the accreditation process
results are submitted to the agency which will provide an evaluation based on the target results
Company Sponsored QC Programs
supply laboratories with daily control material which they analyze and send statistical or raw data on the daily control values to the company at the end of each month
company provides the laboratory with information on the amount of
long term inaccuracy and imprecision that may be occurring
information on the closeness of its results on the control pools to that of peer laboratories using
the same control materials
Some examples of company sponsored QC programs are
Beckman Coulter’s EQAP program for
hematology controls and Biorad’s UNITY program
Proficiency Testing
most common type of external QC involves testing biological samples submitted to a
laboratory by an outside agency.
unknown samples “blind”
results are sent back to the agency which compares the results from the laboratory with the known value or interpretation of the samples and informs the lab of the closeness of the test result to the true value or interpretation
requirement for accreditation
usually must be purchased in addition to daily controls
The standard deviation for proficiency testing is referred to as
standard deviation interval or SDI
to determine the amount a lab’s result may differ
from the average of the results from a group of peers
CAP Surveys
most developed and sophisticated proficiency testing program is the one by the College of
American Pathologists or CAP Survey
participants receive a number of specimens for analysis and return their
results to the CAP Computer Center
Any reported results that differ exclusively from the mean value
are flagged to alert the participant of possible analytical error.
quarterly and annual reports to participants to provide other means of providing an overview of
the laboratory QC.
Accreditation
Accreditation is not government run but is an independent group of experts that examines every
medical laboratory to ensure organizations demonstrate competence to manage and perform the
activities defined by Quality Management Systems ISO9001. Accreditation provides the medical
community and public with confidence that the highest quality standards are being met.
Lab Licensing
-provincial government set of rules
-dministered by the respective Ministry of
Health that regulate the laboratory facility itself and grant permission to that facility to engage in
the business of testing patient samples.
-The provincial College of Physicians & Surgeons
is usually the group given the authority to administer the accreditation process which establishes
the standards and QA policies.
Provincial Programs
Some provinces may have specific regional QA programs in place for monitoring laboratories
Professional Regulation
-protect the public from incompetence and
misconduct
These provincial regulations can be in two forms. The “right to title”
a certain
qualification based on an established training program acts as a recognized standard and is assigned to a certain occupational group.
example the CSMLS certification process serves as
the minimum recognized standard in most provinces in Canada
“No person other than a
member of the SSMLT shall use the title medical laboratory technologist or registered
medical laboratory technologist or any word, title or designation, abbreviated or otherwise, to
imply that the person is a member.”
Another form of professional licensing is the “right to practice”
in which a certain occupational
group is given, not only all the benefits of the “right to title”, but also the authority to restrict
others from performing functions that are deemed exclusive to their practice. There are only a
few occupational groups that have this right—doctors and lawyers for example