General Laboratory Practice Solo 2 Lecture 3

Reference Values

Results of a certain type of quantity obtained from a single individual or group of individuals corresponding to a stated description

The value range for a test or analyte that most of the random population using the same testing method included

Recommended for use by the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC)

Normal Values

Used in the past, but the term has become obsolete. Lacks conformity (no stated description) with different connotations

Health-Associated

Values associated with non-diseased state

• Distinctions help prevent the misunderstanding that reference values are associated only with health

• Diabetic patient, hospitalized diabetic patient, ambulatory diabetic patient

Subject-Based

Previous values from the same individual obtained when the individual was in a defined state of health

Population-Based

Obtained from a group of systematically defined reference individuals. The type of values referred to when the term reference values is used without any qualifying words

Establishment of Reference Values

Data collected during the medical interview, clinical examination, and supplementary investigations must be interpreted by comparison with reference data

• Positive diagnosis: Condition of the patient resembles what is considered typical of a particular disease

• Diagnosis by exclusion: Observed symptoms and signs do not fit the pattern characterized by a set of alternative diseases

• A patient’s lab result is not medically useful if appropriate data for comparison are lacking

Guidelines

Reference Individuals should be clearly defined

Patient should resemble reference individuals

Conditions in which the sample was obtained and processed for analysis should be known

All quantities compared should be of the same type

All lab results should be produced with the use of adequately standardized methods under sufficient analytical quality control

Stages in the pathogenesis of the disease that are the objectives for the diagnosis should be stated

Diagnostic sensitivity and specificity, prevalence, and clinical costs of misclassification should be known for all tests used

Reference Individual

An individual selected as the basis for comparison with individuals under clinical investigation through the use of defined criteria

• Statements describing the source population

• Specification of criteria or health

• Disease of interest

Random sample…

… of individuals in the parent population that meet the selection criteria

• Best reference sample obtained after all practical consideration

• Data interpreted with caution because of possible bias

Selection of Reference Individuals

Must be a random sample from the parent population who fulfill the selection criteria

The defined set of criteria includes:

• statement source of population

• specification of criteria for health or disease of interest

Essential Components required for a set of reference values are

Analytical method: Should include information on equipment, reagents, calibrators, type of raw data, and calculation method

Quality control

Reliability Criteria

Sample Collection

Pre-analytical standardization to reduce bias and variation

• Preparation of the individual

• Sample collection

• Handling of specimen before analysis

Analytical Procedure and QC

Specifications carefully established and described to be reproducible

Method: Equipment, reagents, calibrators, raw data, calculations

Quality control

Reliability criteria

Exclusion Criteria

Used in establishing Health-Associated reference values

Diseases: Cancer, Lupus

Risk Factors: Obesity, occupation, genetics, hypertension

Medication: Drug treatment for disease, drug abuse, oral contraceptives, alcohol, tobacco

Physiological State: Pregnancy, Stress, Excessive exercise

Evaluation of reference values

After the analysis of reference specimens is performed, the reference values are subject to

Partitioning of the reference values into appropriate groups

Inspection of distribution of each group

Identification and handling of outliers

Determination of reference limits

Partitioniong Criteria

Use of specific criteria in the sub classification of reference groups to reduce the biological variation in each group

Age

Sex

Genetic Factor: Ethnic origin, ABO group, Histocompatibility antigens (HLA) typing, genes

Physiological Factor: Stage in menstrual cycle, stage in pregnancy, physical conditions

Other Factors: Socioeconomic, Environmental, Chronobiological (Periodic phenomena)

Inspection of Distribution

Display the reference distribution graphically (histogram) and subsequently examined for

• Highly deviating values (Outliers) that may represent Erroneous Values

• Bimodal or polymodal distributions indicating nonhomogeneous results

  • Reevaluate reference individual selection criteria: Partitioning of values

• The shape of distribution may be skewed than the Gaussian distribution

• Initial estimates of the location of reference limits are useful checks on the validity of computations

Normal

Bimodal

Positive (Right) Skew

Negative (Left) Skew

Identification of Outliers

Erroneous values that deviate significantly from the proper reference values

Values near the furthest point on the long tail of a skewed distribution may be misinterpreted as outliers

Deviating values identified as possible outliers should not be discarded automatically

• Include or exclude rationally after checking records of suspect values and correcting errors

Determination of Reference Limits

Reference Interval

• A set of values usually defined by an upper and lower reference limit, representing a specified proportion of the values frequently the central 95%

Reference Limits

• Describe the reference distribution and provide information about the observed variation of values in the selected set of reference individuals

• Observed patient’s values are compared with the corresponding reference interval, which is bounded by a pair of reference limits

Clinical Decision Limits

• Provide optimal separation among clinical categories; usually based on analysis of reference values from several groups and thus used for differential diagnosis

Nonparametric Analysis

Statistical approach to reference value analysis that requires no assumptions about the distribution; Gaussian or Non-Gaussian

Preferred method to establish reference intervals

Requires 120 reference values per partition

Parametric Analysis

Statistical approach to reference value analysis that requires the distribution of values to be Gaussian (or mathematically manipulated to become Gaussian)

Reference Values are used for interpretation of patient values

Appropriate reference interval presented with the patient value

Abnormal results often indicated

• Low

  • Yellow (IM); down arrow (EPIC)

• High

  • Purple (IM); up arrow (EPIC)

• Critical

  • Red (IM); exclamation mark (EPIC)

Subject-Based vs Population-Based Reference Values

Transferrability

Prerequisite

• Comparable populations with no major ethnic, social, or environmental differences

• Standardization protocols for qualifying reference individuals, preparing individuals for specimen collection and performing specimen collection

Other sources

• Manufacturers’ (package inserts)

• Peer-reviewed literature

• Multicenter trials

Verify appropriateness of values

Clinical Sensitivity

The fraction of those subjects with a specific disease that the assay correctly detects

Clinical Specificity

The fraction of those individuals without the disease that the assay correctly detects

Changing the decision limit of an assay affects clinical sensitivity and specificity