• Reference Interval is defined as the interval between and including two reference limits • It is designed as the central interval of values bounded by the lower reference limit and the upper reference limit at designated percentiles i.e. • For fasting glucose the central 95th percentile reference interval is 3.1-6.1 mmol/L • 95 % of apparently health population will have a fasting glucose value between 3.1-6.1 mmol/L • They are also known as reference ranges or expected values • Reference Interval usually refer to population – based reference values obtained from a well- defined group of reference individuals • This value (RI) is computed by means of either a parametric or non parametric approach • The parametric approach is based on the assumption that values from healthy subjects follow a Gaussian distribution • Then the X ± 2SD would encompass ̴ 95 % of values of a given population • The non-parametric approach makes no assumption pertaining to the distribution of values • In this approach, one obtains a series o results in a population of healthy subjects • Results are then listed in ↑ or ↓ order and the 2.5 and 97.5 percentile values determined • This values will contain the middle 95 % interval • The parametric approach can be applied to ̴ 50 % of laboratory data • For the remaining 50 % a transformation of the data is necessary to normalize the distribution • The Mean and the SD are then calculated and the reference interval will then be X ± 2SD • Where transformation is not successful in normalizing the distribution, non-parametric approach should be used
FACTORS AFFECTING REFERENCE VALUES
Instrument analysis (reagents, calibration etc) Age sex Body weight – less important though a +ve correlation exists for ALP, CK, ALT (males), AST, ᵞGT and urate Effects of smoking, blood pressure and social class have always been considered small Between day variation larger than the diurnal variation but origin uncertain May represent dietary, climatic changes or attributable to biological rhythms of longer duration • Sodium, potassium and urate may follow food consumption • Calcium, oxalate and phosphate are more dependent on hours of sunshine Genetic Factors - play a role but difficult to demonstrate However, they are important in Bilirubin, glucose, urate and urea RESULT VARIABILITY
• Changes in serial patient results can be due to the
following: Patient getting better Patient getting worse Preanalytical variation - subject preparation - fasting vs. non-fasting - drug regimen - physical activity - stress - time of sample collection - stasis - transport - clotting - separation - storage - preparation for analysis Analytical variation This may result in bias and inherent precision Biological variation • One can minimize preanalytical sources of variation by: Good phlebotomy Standard sample transport Proper handling and storage techniques • To assess whether a patient has improved or not • Change in the result must exceed the inherent variation due to biological and analytical variation • The Reference Change Value (RCV) must exceed Z [ CVₐ + CV₁ ]⁰̇ ̇⁵ Z = std deviate =1.96