114 terms

Quality Assurance and Performance Improvement

What is a comprehensive set of policies and procedures, and practices necessary to make sure that the overall operation of the health care facility provides the yes possible management of the patient?
Quality assurance.
What is a set of laboratory procedures designed to ensure that the test method is working properly and that results meet diagnostic needs of the physician?
Quality control.
Control samples
Charting results
Analyzing results statistically
Laboratories Role in QA
Ensure that results are reliable
Provide ranges called reference ranges that indicate what is "normal"
Have ongoing QA process for monitoring analytic results
Ensure reliability (accuracy and reproducibility) of results
Centers of Medicare and Medicaid Services (CMS)
Replaced Health Care Financing Administration (HCFA)
Established regulations to implement the Clinical Laboratory Improvement Act of 1988 (CLIA)
Any facility performing quantitative, qualitative, or screening tests on materials derived from the human body are regulated by CLIA
Regulations depend on complexity of testing performed
Complexity determined by
Risk of harm to patient
Likelihood of erroneous results
Type of testing method
Degree of independent judgment required
Interpretation and availability of home use for the test
Public Service Act
Lab receiving Medicare or Medicaid must be licensed under this act
Responsibility of CMS
Laboratory Standards
Set by Clinical and Laboratory Standards Institute (CLSI)
Formerly called NCCLS
Components of QA
Facilities and resources
Technical competence
QA procedures in the laboratory
Specimen procurement and labeling
Specimen transportation and processing
Quality control
Procedure manuals
Problem solving mechanisms
Instrumentation, turn-around times, accuracy of results, documentation, & record keeping
Facilities and Resources
Physical lab site must be conducive to good performance
Must be safe
Adequate, properly maintained equipment and supplies
Technical Competence
Well-trained, competent personnel
Able to perform tests, QC, maintain instruments, keep accurate & systematic records
Thoroughly oriented on lab policies and procedures
Continuing education for upgrading skills
Competency evaluated regularly
Supervisors monitor all analytic reports and evaluate for errors
Quality Assurance Procedures
Specimen procurement and labeling
Specimen transportation and processing
Quality control
Procedure manuals
Problem solving mechanisms
Instrumentation, turn-around times, accuracy of results, documentation, & record keeping
Test Requesting Procedures
Must include patient identification data, time and date of collection, source of specimen, test requested
Complete form must accompany specimen
Specimen Procurement and Labeling
Information on specimen must be matched with laboratory test(s) request form
Accessible handbook of specimen requirements (must be current)
Emphasis on universal precautions
All specimens must be labeled by person collecting specimen
Analytic result is only as good as specimen received
Specimen Transportation & Processing
Specimens must be transported to lab in safe, timely manner
Specimen arrival must be documented and status in the laboratory determined at any time
Quality Control
Includes monitoring the performance of lab instruments, reagents, products, and equipment
Must be carefully documented
Deviation from normal results
Problems or failures in instrument function
Corrective action taken in response to problems
Includes preventive maintenance records, temperature charts, QC charts for procedures
Reagents and products must be checked before use in patient testing
Use of QC samples, proficiency testing and standards
Procedure Manuals
Must be complete for all analytic procedures performed
Should contain patient preparation, specimen requirements, test request information, criteria for performing test, procedural information including reagents and controls used, instrument calibration, maintenance checks performed, QC data, reference values, how to report results
Procedure Manual
Must be reviewed and updated regularly by the supervisor
Clinical and Laboratory Standards Institute (CLSI)
Formerly National Committee for Clinical Laboratory Standards (NCCLS)
Sets guidelines for writing procedure manuals
Problem Solving Mechanisms and Performance Improvement
Any problem or situation that might affect the outcome of a test must be recorded and reported
Corrective action must be taken to address problems
Corrective action must be documented

Any problem in the overall operation of the laboratory must be addressed, corrected, and documented to improve performance of the laboratory
Other QA Procedures
Determining turnaround times
Evaluating accuracy of final result
Complete, systematic documentation of all procedures involved in obtaining final result
Information processing
Laboratory Information System (LIS)
Result reporting
Record keeping
Reliability Descriptors
Common terms
How close the result is to the true value
Can be improved by
Use of properly standardized procedures
Statistically valid comparisons of new procedure with established reference methods
Use of samples with known values (controls and standards)
Participation in proficiency testing
Ability to reproduce a result
Can be improved by:
Proper inclusion of standards and controls
Statistically valid replicate determinations of single sample
Duplicate determinations of sufficient numbers of unknown samples
Sources of Error
Impossible to obtain exactly same result each time a sample is run
Due to variance or error inherent in any procedure
Limitations related to sampling
Specimen is only 5 - 6 mL of total blood volume of 5 - 6 L (1/1000 of total)
Time of day sample was taken
Patient's position and state of activity
Interval since last meal (fasting or not)
Time interval and storage conditions between obtaining specimen and processing by lab
Limitations related to procedural factors
Age of reagents
Personal bias or limited experience of technologist
Variations in standards, reagents, environment, or methods
Reference Values
"Normal" range
Value of an analyte for a healthy individual
May vary according to age, sex, race, geographic location, ethnic, cultural, and economic characteristics
Reference Interval (Normal Range)
Range of values statistically derived from data obtained from a reference population
Interval represents limits of normal or 95% of a healthy population
Reference Range or Interval (Normal Range)
Commonly used when referring to the range of values for a particular blood constituent in healthy individuals
Determined usually by calculating mean and standard deviation of values from reference population for analytes that follow Gaussian distribution
Laboratory Reference Intervals
Each lab determines its own reference intervals for its instrumentation and the population it serves
Reference intervals published by manufacturers of reagents and instrument must be confirmed by the laboratory
Establishing Reference Intervals
Sample collection, processing, and analysis should be closely monitored
Specimens should not be treated in a "special manner"
Do not use samples usually considered unacceptable (hemolyzed, lipemic, etc.)
Establish ranges when new method is instituted
Establishing Reference Intervals
Specimens should be from individuals representative of the population to which the ranges will be applied
Number of individuals required depends on method of reference interval calculation
Confidence Interval
Establishes allowable limits for quality control material
Also known as "control range"
Calculation for values that follow Gaussian distribution
Standards (Calibrators)
Material with accurately determined concentrations
Used to confirm a method's accuracy or estimate amount of analytical error
Also used to set instrument when performing ratiometric testing (calibration)
Used to make standard curve in manual methods
Characteristics of Standards
Should have same physical and chemical properties as patient samples and react in same manner
Values are assigned after thorough testing by reference methods
Cannot be used as calibrator and QC material in the same procedure
Quality Control Materials
Controls are analyzed with patient samples during testing
Material must closely resemble patient specimens in appearance and physical and chemical characteristics
Use of QC Materials
Must be used in entire determination and be tested in exactly same manner as patient samples
Normal and abnormal QC samples are usually run
Characteristics of Ideal QC Material
Values determined by variety of analytic methods performed in several independent laboratories
Closely controlled manufacturing from lot to lot
Composition of QC Materials
Lyophilized Controls
Protein based (usually human serum)
Requires reconstitution with water or diluent
Preservative and antimicrobial agents added
Stable for limited time after reconstitution
Problems with Lyophilized QC
Some serum properties altered
Lipoproteins are denatured
Increases turbidity
Preservatives change viscosity
Improper reconstitution causes changes in concentrations of analytes
Liquid Serum Controls
Serum diluted with ethylene glycol which lowers freezing point
Remains liquid at -20C
No defrosting required
No problems with turbidity
Eliminates errors in reconstitution
Stable for several days to weeks if refrigerated
Aqueous-Based Controls
Chemical/water mixtures
Less expensive
No vial-to-vial variation
Stable for long periods of time
Different physical matrix than patient samples
Lower viscosity, lighter color, less turbidity
Limitations of Liquid QC
Ethylene glycol increases viscosity
Interferes with freezing point osmolality
More expensive
Types of QC Materials
In-house prepared
Assayed QC Material
Commercially prepared
Values determined by many independent laboratories
Unassayed QC Material
Commercially prepared
Values determined by the laboratory and pool of other laboratories using same lot of material
In-House Prepared QC
Not recommended
Deterioration, contamination, and loss of potency may occur during storage
Possibly infectious
Choosing Type of Controls
Chosen by laboratory according to workload and expense
Lab with large workload usually uses unassayed
Lab with small workload usually uses assayed
Internal Quality Control
Routine monitoring of test results using controls within acceptable ranges
When controls fall outside acceptable confidence interval, decision to accept or reject patient values must be made
Includes instrument verification and preventive maintenance
QC Program
Purchase controls in batch large enough for at least 1 year
Use more than one level of control
Analyze QC for minimum of 20 consecutive working days for each method
Evaluate QC results by CUSUM or Westgard multi-rule
Calculate statistics
Analyze at frequency that is sensitive to method's degree of imprecision
Commercial QC Programs
Convenient record keeping
Perform all necessary statistical tests
Can provide information useful in selecting new methods
Regional lots provided for 1 to 2 years
Lab sends data monthly
Report includes monthly and cumulative mean, s, & CV
Your lab's result compared to other labs using SDI or Youden plot
Levey-Jennings plot also provided
Reasons for Monitoring QC
Monitor variations in results due to procedural errors, failure of materials (systematic error) or random error
Internal QC
Based on premise that random variations can be measured, calculated, and predicted by statistical process
Insures that test performance is maintained with established allowable limits for variation
Process of Internal QC
Establish allowable limits of variation
Evaluate all data based on the criteria
Take remedial action when indicated
Establishing Confidence Interval
Analyze normal & abnormal controls for 20 or more days
Calculate mean, s, CV
Calculate confidence interval
Must be within allowable limits of error established for method in use
Establishing Confidence Interval
Lab must establish its own C.I.s even when using assayed controls
Manufacturer's ranges may be too lenient to obtain accuracy desired by your lab
Between-lab variation in methods, equipment, personnel and instruments may result in values within stated range but not at the mean
Manufacturer's range may be required for rarely performed tests
Statistics: confidence interval
Statistical estimate of where 95% of control values are expected to fall
Statistics: mean
Average value of set of data points
Statistics: Standard deviation
Statistics: % CV
Evaluation of Control Results
Frequency of control analysis varies according to method
Controls should be placed randomly in the patient workload
Evaluation tools
Levey-Jennings plots
Westgard multi-rule analysis
Levey-Jennings Charts
Plot controls on chart which indicates mean, 1s, 2s, 3s, etc.
Levey-Jennings Plot
Easily detects trends & shifts
Imprecision detected
Wide scatter of data points
Corrective action must be taken when trend or shift occurs before patient results are reported
When control value ↑ or ↓ for 6 consecutive days or when both controls ↑ or ↓ for 3 consecutive days
A gradual change in the mean
Usually indicates a systematic proportional error
Abrupt change in the mean
Occurs when
One control is beyond 1s for 4 consecutive days or on one side of the mean for 6 consecutive days
When both controls are beyond 1s for 2 days or on one side of the mean for 3 days
Usually indicates constant systematic error
Multi-Rule Analysis (Westgard)
Evaluation of paired control runs
Involves 6 rules and indicates when run is out-of-control or gives warning of possible error
1(2s) Rule
A control value is plotted outside a 2s limit
1(3s) Rule
Possible random error
One value is plotted outside 3s limit
Out of control
2(2s) rule
2 consecutive values are outside same 2s limit (either -2s or +2s)
May be within same run involving both levels of control or 2 consecutive analyses of the same control
R(4s) Rule
2 consecutive values are more than 4s apart involving both levels of controls
1 control beyond +2s limit and one beyond the -2s limit
4(1s) Rule
4 consecutive values on same side of 1s range within and across both control levels
Indicates shift or trend
10x Rule
10 consecutive values fall on same side of mean either within same control or across both control levels
Westgard Algorithm
Cumulative Summary (CuSum)
Evaluation of single levels of control
Based on random distribution of control values about the mean
Involves keeping sum of differences between each value and the mean
Sum of differences should be zero
If sum exceeds +2.7s or -2.7s, the method is out-of-control
Shift or trend will cause CUSUM to move away from zero & exceed out-of-control limit
Out-of-Control Values
Reporting patient results must stop immediately when run is declared out-of-control
Do not resume reporting until problem is found and corrected
Steps in Investigation
Visually inspect for obvious problems
- Check instrument for proper performance
- If nothing visual, reanalyze new aliquot of control
If problem persists, inspect control material for problems
- Reconstitute new control or use assayed control
Steps in Investigation
Other options
- Recalibrate
- Systematically replace parts
- Contact manufacturer
Corrective action
- Document all actions taken
Patient Values as Controls
Duplicate or replicate analysis of patient samples
Mixed samples
Delta checks
Patient Values as Controls
Daily average
Numbers plus
Anion gap
Duplicate or Replicate Analysis
Patient samples repeated to check precision of manual & automated test methods
Duplicate or Replicate Analysis
Average differences & standard deviation of differences are plotted on chart
Mixed Samples
Evaluates precision
Mix equal parts of 2 randomly chosen patient samples
Calculate average of 2 results to determine target value
Analyze mixture
Compare result to calculated average
Mixed Samples
Difference between target and obtained values are plotted
Control limits are 2X the average difference of results collected over 40 day period
Delta Checks
Compare patient result with result from previous sample for same analyte
Delta Checks
Difference greater than expected run-to-run precision, day-to-day precision or a designated clinically significant value may indicate an analytical error
Daily Average
Averaging daily results for constituents with a narrow physiological range can be followed
Deviations from established average may indicate error
Daily Average
Moving averages are statistically smoothed by computer to reduce effect of individual values on the average
Numbers Plus
Results of 500 consecutive samples are plotted on frequency histogram
Mode is determined
Numbers Plus
Percentage of values above the mode will remain the same regardless of number of patients analyzed daily
Anion Gap
Used to estimate validity of electrolyte results
Na+ - (Cl- + HCO3-)
Reference range = 5 to 14 mmol/L
Checking Absurd Values
Investigate results beyond reference range or that are not consistent with life (critical value)
When detected, repeat analysis by same or alternative method
Checking Absurd Values
Values may be result of random error or instrument malfunction, or sampling error (short sampling or clot)
If no error is found, notify physician immediately
Pattern Recognition
Detects unlikely combinations of results from related tests
Pattern Recognition: Examples
BUN and creatinine
Sodium, potassium, and CO2
Instrument Verification and Preventive Maintenance
Regular monitoring of instrument functions, adjustment, care, and cleaning of equipment required
Instrument Verification and Preventive Maintenance
Necessary to ensure accuracy and precision
Necessary to provide sustained, uninterrupted service
Elements Essential for Quality Instrument Performance
Performance verification
Function verification
Instrument verification
Documentation of maintenance performed
Performance verification
Evaluation of linearity, wavelength, photometric performance, etc
Function verification
Checking reagents, calibration, linearity, sensitivity, run adjustments, monitoring QC
Instrument verification
Cleaning, oiling, replacing tubes & worn parts
External QC
Comparison of lab's performance with outside agency or group of other labs
Monitors accuracy of lab results
External QC
Checks lab's internal QC
Detects between laboratory error
External QC
Useful in evaluating and selecting new instrumentation
Proficiency Testing
Participation in national survey such as CAP
Analyzed too infrequently to substitute for internal QC
Proficiency Testing
Agency sends unknowns to participating labs
Unknowns are analyzed with routine workload
Proficiency Testing
Results returned to agency
Lab's results are compared to results of all participants
Proficiency Testing: Standard Deviation Index
Same as Z-score
Proficiency Testing: Youden Plot
Plots normal control on X-axis and abnormal on Y-axis for lab results and group means
Proficiency Testing
2s box drawn and lab's results are evaluated
Good QA Program
Many activities at bench & supervisory levels
Ensure good quality sample
Good selection of method or instrument
Good QA Program
Define good reference intervals
Effective use and evaluation of QC
Participation in proficiency surveys
Good QA Program
Adequate maintenance and monitoring of instrument performance