F(3.2): Quality Assurance Flashcards
- Assumes population specimens produce bell-shaped curves (histograms).
- Mean is at the center; mean, median, and mode coincide.
- Dispersion about the mean is identical in both directions
Gaussian (Normal) Distribution
- Biologic frequency distributions often have a “tail” on the high end.
Example: Reticulocyte percentage RI in adults adjusted from 0.5%-1.5% to 0.5%-2% due to high-end values.
Log-Normal Distribution
- QC managers may use log-normal distribution or transform data using semilog or log-log graphs.
- Transformation decisions may become national practice standards.
Transforming Data
- Mean (X) is computed by dividing the sum of observed values by the number of data points (n).
- SD is calculated using a specific formula.
- Typical RI is computed as mean ±2 SDs, assuming normal distribution.
Calculating Mean and Standard Deviation (SD)
- Limits at mean ±2 SDs encompass 95.46% of results from healthy individuals (95.5% confidence interval).
- 4.54% of results from healthy individuals fall outside the interval by chance.
Confidence Interval
- SD from a non-Gaussian distribution may be too narrow, generating excess false positives.
- Assays with high coefficient of variation (CV%) have high random error (broad curve).
- Low CV% assays with tight dispersion have smaller random error (narrow curve).
Non-Gaussian Distribution
- Occurs when an abrupt change is followed by six or more consecutive QC results on one side of the mean.
- These results typically fall within the 95% range, clustering around a new mean.
- Important to note because shifts indicate a problem, leading to rejection of results
Shift in QC Results
- Occurs when values gradually and continually move in one direction over six or more analytical runs.
- Values may display across the mean or only on one side of the mean.
- Important to note because, like shifts, trends lead to the rejection of results
Trend in QC Results
Reject the results when there is a trend of seven consecutive control measurements moving in one direction.
7T Rule Violation
control of tests
Selection of test should be based on four factors:
- result
- procedure
- value
- reports
The assessment factors for method evaluation and selection
- precision
- accuracy
- sensitivity
- specificity
FAMILIARIZE:
MUST KNOWS IN SPECIMEN COLLECTION
Proper way to collect a specimen
Correct details concerning name of tests
Preparations of patients
Containers ana labels
Preservative or anticoagulant
Handling or storage of specimen before delivery to lab
Special instruction and other special specimen
Requirements
FAMILIARIZE
Guidelines in the use of reagent kits
It should give consistent result form day to day and in patient samples with low, normal, and elevated values
Manufacturer’s data must be complete
Statement of principle, description or test performance and calculation, data on precision, accuracy and stability of reagents
Cheaper than manual methods
Every employee should be competent to use the kit
FAMILIARIZE
RELIABILITY OF REAGENTS
Check for contamination or deterioration
Frequent assessment of reagents’ quality
May employ a one-step method involving running QC materials (e.g. Clinical chemistry and hematology)
FAMILIARIZE
suggestion in the control of equipment
Use National Bureau of Standards guidelines for monitoring water baths, heating blocks etc.
Verify reliability of analytical balances and weights
Check speed of centrifuge using a tachometer
Check all water used in testing for its pH, purity and any foreign substance
Perform method validation for new analyzers/equipment or as deemed necessary for the old ones
FAMILIARIZE
control of staff
Maintain an atmosphere of professional challenge
Offer fair wages
Provide a means for obtaining continuing professional education and career advancement
- Program where participating laboratories are given unknown samples for analysis
- Samples are to be treated as ordinary human specimens for the usual processing and examination
- Administered by the different National Reference Laboratories (NRL) – almost all NRLs send one set of samples per year
- Certificate of participation is required for renewal of the laboratory’s license from the DOH-HFSRB
external quality assessment program (eqap)
NRL
Clinical chemistry
Lung Center of the Philippines (LCP|)
