Reference Change Value Calculator
Reference Change Value Calculator
How much must a result move before the change is real? The reference change value combines analytical imprecision with within-subject biological variation to answer the question every repeat test raises.
Reference Change Value
CVa, CVi, Z → significant changeCreatinine: analytical CV 2.2%, within-subject biological variation 4.4%, two-sided 95%
Formula
Z = 1.96 for two-sided 95% · 2.58 for two-sided 99%
- CVa
- analytical imprecision, as a percentage, from your own internal quality control at a concentration close to the patient’s result. Manufacturer claims are obtained under favourable conditions and usually understate what a routine laboratory achieves
- CVi
- within-subject biological variation: how much the analyte moves within one person around their own set point, independent of any disease. It is tabulated per analyte in the EFLM biological variation database and is often much the larger of the two terms
- √(CVa² + CVi²)
- the two sources of variation combine in quadrature rather than by addition, because they are independent. This is why halving an already-small analytical CV barely moves the RCV when biological variation dominates — a point worth making before buying a more precise analyser
- √2
- there are two measurements, each carrying its own error, so the variance of their difference is twice the variance of one. That factor of two under the square root is where the √2 comes from
- Z
- 1.96 for two-sided 95%, 2.58 for two-sided 99%. Some references halve the tail and use 1.65 when only a change in one direction is of interest, which gives a smaller RCV — say which convention you are using when you quote a figure
Worked example
Creatinine: analytical CV 2.2%, within-subject biological variation 4.4%, two-sided 95%
CVa² + CVi² = 2.2² + 4.4² = 4.84 + 19.36 = 24.2
√24.2 = 4.92
√2 × 1.96 = 2.77
2.77 × 4.92 = 13.6%
So a creatinine of 90 µmol/L must reach about 102 or fall to about 78 before the change is more than noise
A patient whose creatinine moves from 90 to 98 — an 8.9% rise that looks like a deterioration on a screen — has not yet changed by more than the test and the body vary anyway
Roughly what the RCV looks like for familiar analytes
| Analyte | Typical CVi | RCV at 95% (with a 2% CVa) | A change of |
|---|---|---|---|
| Sodium | 0.6% | ≈ 5.8% | 138 → 146 mmol/L |
| Calcium (adjusted) | 1.9% | ≈ 7.7% | 2.35 → 2.53 mmol/L |
| Creatinine | 4.4% | ≈ 13.4% | 90 → 102 µmol/L |
| ALT | 18% | ≈ 50% | 40 → 60 U/L |
| Triglycerides | 20% | ≈ 56% | 1.5 → 2.3 mmol/L |
| PSA | 18% | ≈ 50% | 4.0 → 6.0 µg/L |
What the reference change value is not
| Question | Answer |
|---|---|
| Is it a reference interval? | No. A reference interval compares one result with a population. The RCV compares two results from the same person with each other, which is a different and usually more sensitive question |
| Does exceeding it mean the patient has deteriorated? | No — only that the change is unlikely to be chance. Diet, posture, time of day, a new drug and the sample itself all produce real changes that are not disease |
| Does staying inside it mean nothing has happened? | No. A slow drift well within the RCV at each step can be highly significant over a year, which is why a trend beats a pair |
| Can I use the manufacturer’s CV? | Only as a placeholder. Use your own internal QC at a concentration near the patient’s result; routine imprecision is generally worse than an evaluation study’s |
| Does it work near the limit of detection? | Poorly. Analytical CV rises sharply at low concentrations, so an RCV calculated from mid-range QC understates what is needed down there |
Is this change real, or is it noise?
A patient’s creatinine was 90 µmol/L in March and is 98 µmol/L today. Is that a deterioration? The reference interval cannot answer, because both figures are inside it, and the two results are being compared with each other rather than with a population. What is needed is the size of change that this test on this patient produces when nothing at all has happened. That is the reference change value.
Two things generate that noise. The laboratory contributes analytical imprecision, the CVa, which is measured every day by internal quality control. The patient contributes within-subject biological variation, the CVi — the amount an analyte swings around one person’s own set point with the time of day, the last meal, posture, hydration and ordinary physiology. These two are independent, so they combine in quadrature: the square root of the sum of their squares. Because there are two measurements and each carries its own error, the whole thing is multiplied by the square root of two, and then by 1.96 for a two-sided 95% probability or 2.58 for 99%.
For creatinine with a 2.2% analytical CV and 4.4% biological variation, that comes to about 13.6%. A creatinine of 90 has to reach roughly 102 before the change means anything, so the move to 98 is well within the noise. This is the calculation behind the familiar clinical instinct that small movements in stable patients should be watched rather than acted on — and it also explains the reverse, that a change of the same 8 µmol/L in sodium, where variation is far tighter, would be substantial.
Two implications are worth drawing out. The first is that when biological variation dominates, buying a more precise analyser barely helps: halving a 2% CVa against a 4.4% CVi changes the RCV from 13.6% to about 12.6%. The second is the direction of error people rarely consider — exceeding the RCV does not prove disease, because diet, drugs, posture and the sample itself all cause genuine changes; and staying inside it does not prove stability, because a slow drift that never breaches the threshold between consecutive pairs can still be unmistakable over a year. The RCV tells you whether a difference is bigger than the noise. What caused it remains a clinical question.
Frequently asked questions
What is the reference change value?
The smallest difference between two results from the same person that is unlikely to be explained by analytical imprecision and normal biological variation alone. It is calculated as the square root of two, times Z, times the square root of the sum of the squared CVs.
Why is a reference interval not enough for serial results?
Because it compares a result with a population, not with the same patient’s previous value. Someone whose usual creatinine is 60 and is now 105 has changed substantially while remaining inside most reference intervals throughout. The RCV catches exactly that.
Should I use Z of 1.96 or 2.58?
1.96 gives a two-sided 95% probability, so about one stable patient in twenty exceeds it by chance; 2.58 gives 99% and about one in a hundred. Use 95% for routine monitoring and 99% where acting on a false change would be costly. Say which you used when quoting a figure.
Where do I find within-subject biological variation?
The EFLM Biological Variation Database is the current reference source, built on systematically appraised studies; Fraser’s Biological Variation: From Principles to Practice is the standard text behind it. Figures vary by analyte from under 1% for sodium to over 20% for triglycerides.
Does exceeding the reference change value mean the patient is worse?
No. It means the difference is bigger than the test and the body produce on their own. A new drug, a meal, a change in posture or a difficult sample can all produce a real change that is not a deterioration. The RCV separates signal from noise; interpreting the signal is still clinical.
Related calculators
References
- Fraser CG. Biological Variation: From Principles to Practice. AACC Press; 2001.
- Ricós C, Alvarez V, Cava F, et al. Current databases on biological variation: pros, cons and progress. Scand J Clin Lab Invest. 1999;59(7):491–500.
- Aarsand AK, Fernandez-Calle P, Webster C, et al. The EFLM Biological Variation Database. European Federation of Clinical Chemistry and Laboratory Medicine.
- CLSI EP15-A3. User Verification of Precision and Estimation of Bias; Approved Guideline. 3rd ed. Clinical and Laboratory Standards Institute; 2014.
Medical Disclaimer: The tools and content provided here are for educational and reference purposes only. They are not intended to substitute for professional medical advice, diagnosis, or treatment. Clinical decisions should always be based on the comprehensive assessment of a qualified healthcare professional.