Interpretation of Laboratory Results
Clinical Pharmacy Graduate Notes
Introduction
Clinical laboratory test results are crucial parameters in diagnosis, monitoring, and screening. Approximately 70-80% of diagnostic decisions are based on laboratory results. With an increasing volume of tests ordered, clinicians must be proficient in test interpretation for optimal patient care and safety.
The laboratory also has a responsibility to provide clinicians with adequate information to assist in correct data interpretation. Rational use of clinical biochemical analyses requires understanding of key concepts outlined in this document.
Key Point: Ordering too many tests uncritically does not necessarily provide more information and may complicate interpretation. The clinician must know how appropriate and reliable a test is for its intended use.
Reference Interval
Interpretation of a laboratory result requires comparison to a relevant reference value, either the patient's earlier results or data from a "normal" population.
Establishing Reference Intervals
- Established by collecting samples from a normal healthy population (at least 100, preferably several hundred individuals)
- Sources: blood donors, hospital staff, or literature data
- IFCC protocols are available for well-defined reference intervals
- Variation arises from natural biological variation and measurement uncertainty
Statistical Basis
For normally distributed variables:
- Approximately 95% of values fall within mean ± 2 standard deviations (SD)
- Approximately 99% of values fall within mean ± 3 SD
- The reference interval including 95% of values is limited by lower (2.5%) and upper (97.5%) reference limits
Important: 5% of a healthy population will have results outside the "normal range" by statistical definition. If two tests are ordered, the probability that both are normal is 0.95² = 0.90. For 10 tests, probability all are normal is only 0.60.
Limitations
- Abnormal results do not always indicate disease
- Normal results do not always indicate absence of disease
- Reference intervals may vary by age, sex, size, or ethnic background
- Some parameters have skewed distributions requiring transformation or non-parametric statistics
| Age Group | Creatinine Reference Interval |
|---|---|
| 0-1 week | 27-62 μmol/L |
| 1 week-1 month | 18-35 μmol/L |
| 1-12 months | 18-66 μmol/L |
| 1-14 years | 50-90 μmol/L |
| >14 years | 60-100 μmol/L |
Intra-individual Biological Variation
Even within a single individual, biochemical parameters fluctuate due to:
- Seasonal variation
- Biological cycles/rhythms
- Food intake
- Exercise
- Time of day
This variation is particularly important when comparing a patient's current results with previous ones (e.g., treatment evaluation).
Critical Difference (CD)
The minimum difference between two results needed to be considered statistically significant:
Where 2.77 = √2 × z-statistic (1.96 for 95% probability, two-tailed).
Example - Glucose: CVbiological ≈ 6%, CVanalytical ≈ 2% for plasma glucose. Critical difference = 2.77 × √(6² + 2²) = 18%. Two glucose values differing by more than 18% are statistically significant.
Accuracy and Precision
Precision (Imprecision)
The degree to which replicate measurements under unchanged conditions show the same results. Measures random analytical errors.
Accuracy (Bias)
The degree of closeness of measurements to the true value. Measures systematic errors.
Total Error (TE)
Combined effect of bias and imprecision:
Analogy: Shooting at a target - Bias is deviation from center; Precision is spread of bullet holes.
Example: C-reactive Protein (CRP)
As a cardiovascular risk predictor, CRP must discriminate between:
- < 1.0 mg/L (low risk)
- 1.0-3.0 mg/L (average risk)
- > 3.0 mg/L (high risk)
High imprecision in lower ranges limits utility for cardiovascular risk assessment.
Diagnostic Sensitivity and Specificity
Definitions
- Analytical Sensitivity: Ability to measure low concentrations of analyte
- Analytical Specificity: Ability to measure analyte without interference
- Diagnostic Sensitivity: Proportion of true positives correctly identified
- Diagnostic Specificity: Proportion of true negatives correctly identified
Where: TP = True Positive, TN = True Negative, FP = False Positive, FN = False Negative
Example: A pregnancy test showing negative when the woman is actually pregnant has low sensitivity. A test showing positive when the woman is not pregnant has low specificity.
Predictive Value
The probability that the disease is present when the test is positive (or absent when negative).
Predictive values depend on sensitivity, specificity, and disease prevalence. When prevalence is low, even tests with high sensitivity/specificity can yield many false positives.
Clinical Application and Brain-to-Brain Cycle
The interpretation-action cycle involves:
Preanalytical Parameters
- Biological variation
- Appropriate timing for sample collection
Analytical Parameters
- Accuracy and imprecision
- Diagnostic sensitivity and specificity
Postanalytical Parameters
- Clinical decision limits
- Failure rates
- Clinical interpretation
Key Finding: Errors in test selection and result interpretation are more common than errors in sample collection, analysis, or reporting. For coagulation disorders, approximately 75% of cases involve some level of test result misinterpretation.
Test Selection Guidelines
- For exclusion of diagnosis: Highly sensitive test required
- For diagnosis of high-risk disease: Highly specific parameter needed
Laboratory information systems should incorporate decision aids and evidence-based tools to support appropriate interpretation and application of laboratory data.
