Chapter 4: Quality Management

quality control

quality control

system of techniques that ensure the laboratory result obtained from each series of analysis is true and correct

intralab/internal QC

• detects changes in the performance between the present operation and the ‘stable’ operation

• involves the analyses of control samples together with the patient specimens

interlab/external QC

• important in maintaining long-term accuracy of the analytical methods

• involves Proficiency Testing programs that periodically provide samples with unknown concentration of analytes to participating laboratories

variations

errors in quality control; fundamental basis of any statistical analysis

random error

non-recurring/unpredictable error; imprecision of the test system

systematic error

recurring/predictable error; causes shifts or trends

delta check

• checks pre-analytical methods

• most commonly used patient based-QC techniques where the difference between consecutive laboratory data is calculated and compared to previously established limits

numerical and percentage

two ways delta difference is calculated

analytic sensitivity/detection limit

the smallest concentration that can be measured accurately

analytic specificity

a method’s ability to measure only the analyte of interest

reliability

ability to maintain accuracy and precision over an extended period of time

accuracy

refers to how close the measured value is to the true/actual value

precision

refers to the reproducibility of a measurement; the smaller the SD the better

mean/average

• summation of data divided by the total number of data

• measure of central tendency

• evaluates a trend

normal distribution/gaussian distribution

characteristic “bell-shaped” curve drawn when all the values are symmetrically distributed around the mean

non-gaussian distribution/Log-Normal distribution

• distribution of values on both ends of the peak is not equal

• mode value is used instead of mean

standard deviation

measure of the distribution of values around the mean

± 2 SD

limits of the acceptable range for a test

68.7%

How many of the values lie within mean ± 1 SD?

95.5%

How many of the values lie within mean ± 2 SD?

99.9%

How many of the values lie within mean ± 3 SD?

coefficient of variation

• the standard deviation expressed as a percentage of the mean

• ideally should be less than 5%

variance

a statement of variability and measures the significant differences between groups of data; square of the standard deviation

reference range

• usual range of values for the healthy population

• consists of the central 95% of the population of interest

20

One person in how many will fall outside the reference range?

control solution/materials

• monitors the stability of the test system

• results compared with the manufacturer’s range of values

human and animal (bovine)

two sources from where control reagents are taken

normal and abnormal

two kinds of control reagents based on levels of analyte/constituents

commercial and non-commercial

two kinds of control reagents based on preparation

gaussian curve

occurs when data samples are centered around the mean

cumulative sum graph (cusum)

• gives earliest indication of a trend for systematic errors

• identifies consistent bias problems

youden plot

used for intralaboratory quality control programs

levey-jennings chart/ shewhart charts

most commonly used histogram, a gaussian curve on its side

in-control

all values fall within the confidence limit of ±2SD

out of control

values fall outside the confidence limit

trend

• 6 consecutive days passing the mean

• indicates gradual loss of reliability, usually subtle

shift

• 6 consecutive days distributed on one side of the mean

• abrupt changes

• sudden and dramatic positive or negative change

westgard multiple chart

• used in conjunction with LJ and is used intraquality control testing

• 2 levels of control material analyzed per run

• error detection rates can be increased without increasing the false rejection rate

1_2s

• one of the two control observations exceeding the mean ±2SD control limits

• warning rule

1_3s

either one of the two control observations exceeding the mean ±3SD control limits; run must be rejected

2_2s within

both controls in the same run exceed ±2SD

2_2s across

two consecutive control values for the same level fall outside of ±2SD in the same direction

R_4s

• one control exceeding the +2SD and another exceeding

  the -2SD

• allows detection of random error

4_1s

• four consecutive control observations are on the same side of the mean and exceeding ±1SD control limit

• allows for detection of systematic error

10_mean

ten consecutive control observations falling on one side of the mean

• 1_3s

• 2_2s

• R_4s

• 4_1s

• 10_mean

errors that warrant a run rejection

quality assurance

a comprehensive system that monitors the entire process to ensure the integrity of test results where all steps before, during, and after the testing process are considered

pre-analytical phase

• test ordering

• patient preparation/identification

• specimen collection/transport/processing

analytical phase

• test analysis

• quality control

• reagents

• calibration

• preventative maintenance

post-analytical phase

• verification of calculations and reference ranges

• review of results

• procedures for notification of critical values

• result

• result reporting

• test interpretation by physician

• follow-up patient care

method evaluation

ensuring a method remains valid over time