Introduction to Clinical Chemistry and Quality Assessment/Quality Control

Screening

Detection of subclinical disease with regular periodic testing and no symptom presentation

Diagnosis

Confirmation or ruling out of a disease with symptom presentation

Monitoring

Checking for treatment or disease progression

Prognosis

Long-term outcome of disease state (recovery or relapse)

Specimen factors

• Age, sex, ethnicity, pregnancy, posture, exercise, stress, nutrition, time of day, drugs

• needs to be in appropriate containers, labels, and transport conditions

Issues with specimen collection

• hemolysis (pink serum/plasma)

• delayed centrifugation (false low blood sugar because of glycolysis)

• incorrect anticoagulant

• temperature requirement deviation

Organic biochemical markers

• carbohydrates

• lipids

• proteins

• nucleic acids

Inorganic chemical markers

• ions (Na+ and K+)

• minerals

• dissolved gases (arterial blood gas)

Exogenous chemical markers

Taken in:

• therapeutic drugs (chemotherapy)

• harmful substances (poisons/toxins)

Quality assessment/assurance

System of creating and following procedures and policies providing most reliable patient laboratory results (includes QC)

Quality control

System designed to ensure individual patient sample measurements are within clinically acceptable limits

• monitors quality of lab testing and accuracy and precision of results

• routine collection and analysis of data

• allows for immediate corrective action

Qualitative testing

Determines presence or absence of substance or analyte, requires + and = control in each run of patient samples

Quantitative testing

• measures amount of substance present

• at least two levels of controls should be included in each day of patient testing

  • normal and abnormal ranges determine likelihood of error

Pre-analytical errors (QC)

Problems with specimen (collection, handling, anticoagulant, contamination) and patient identification

Analytical (QC)

• implementation of QC processes

• selection and managing of control materials

• analysis and monitoring QC data

• staff training, documentation, review

How to choose QC program

• choose high quality QC materials with known analyte concentrations

• 2-3 levels of control

• values cover medical decision levels

• similar to test specimen

• available in large amounts

Assayed control material

Mean is calculated by manufacturer and is commercially produced, needs to be verified by lab

Unassayed control materials

Less expensive than assayed (made “in house” with pooled sera or specimen)

• data analysis is performed

• mean is determined by lab

How to establish control ranges

• 20 data points over 2 weeks or 10 working days

• not over 4 weeks or 20 working days

• represent procedural variation (different operators and time of day)

• determine amount of allowable error

Median

Value at center (midpoint) of observation points

Mean

Calculated average of valuesM

Mode

Value that occurs with the greatest frequency

Precision

Degree of fluctuation in measurements indicates precision

Accuracy

Closeness of measurements to true value is accuracy

Between run precision

Reproducibility - closeness of individual results with same method and test material but different conditions, operator, and instruments

Within run precision

Repeatability - closeness of individual results with same method, test material, conditions, operator, instrument, and time intervals

Measures of dispersion of variability

Range, variance, standard deviation, coefficient of variation

Standard deviation

• Gaussian distribution with +/= 1, 2, 3 SD

• result in values falling within range 68.2, 95.4, and 99.7% of the time, respectively

Standard deviation criteria for reference ranges

• usually +/= 2 SD are used to set criteria for acceptable limit (95.4% confidence)

• values fall outside of criteria 4.5% of time most likely due to error

Coefficient of variation (CV)

• SD expressed as percentage of mean (index of precision), ideally less than 5%

• CV% = SD/mean x 100

Levey-Jennings Chart

Graphical method of displaying control results and evaluating if a procedure is in or out of control

• plotted vs. time

Random error

Error that occurs unpredictably due to poor precision

Systemic error

Error that occurs predictably once a pattern of recognition is established

Shift

Sudden and sustained change in quality control results above or below mean

Trend

Gradual and steady change in QC results moving up or down away from mean

westgard rule 1_2S

• one control point of two fall outside of ±2 SD

• warning rule - alert to potential problems

• do NOT reject run

westgard rule 1_3S

• one control point of two fall outside of ±3 SD

• detects random error

• reject run

westgard rule 2_2S

• two consecutive control points for the same control level falls outside of ±2 SD in the same direction

• the same control point in both control levels in the same run exceeds ± 2 SD

• detects systematic error and requires corrective action

• reject run

westgard rule R_4S

• one control exceeds the mean by -2 SD and the other control exceeds the mean by +2 SD

• the range between the two results exceed 4 SD

• random error has occured

• reject run

westgard rule 4_1S

• four consecutive control points for one or more levels of controls are outside ±1 SD on the same side of the mean

• this is a shift that indicates systematic error

• reject run

westgard rule 10_x

• ten consecutive control points for one level of control is on one side of the mean

• shift indicating systematic error

• reagent or calibration problem

• reject run

westgard 7_T

• seven control points (not including first point in graph) trend in the same direction

• can cross mean

• reject run

imprecision

large amount of scatter around the mean

• usually caused by errors in technique

inaccuracy

may be seen as a trend or shift

• usually caused by change in testing process or deterioration of reagents

random error

no pattern

• caused by poor technique or malfunctioning equipment

post analytical factors

• reference ranges

• medical decision values

• critical values