A biological age report arrives designed to be read in one direction. The headline number sits at the top in the largest type, coloured green or amber, next to your chronological age. Everything that would help you judge that number sits further down, in smaller type, if it appears at all.
Reading it usefully means going in roughly the opposite order.
The Verdict
What each section of the report is worth
| Section | What it shows | What to check | How much weight it deserves |
|---|---|---|---|
| The headline age | A single number, usually compared against your chronological age | Which clock produced it, and whether it is a first- or second-generation clock | Almost none on its own. It is the least informative number on the page. |
| Pace of ageing | A rate, where 1.0 means one biological year per calendar year | Whether the provider reports one at all | High. A rate is directly comparable between two of your own tests. |
| Organ or system scores | Separate ages for heart, brain, liver, immune system and others | How many people the underlying model was trained on, and whether it is peer-reviewed | Moderate. Newer and less validated than whole-body clocks. |
| Confidence interval or error range | A plus-or-minus figure around the headline age | Whether it appears at all, and whether it reflects assay noise or model error | Very high. Its absence is itself a finding. |
| Telomere length | Average length of chromosome end caps | Whether it is presented as a primary result | Low. High measurement variability and weak individual-level prediction. |
| Recommendations | Lifestyle or supplement suggestions generated from the result | Whether any suggested product is sold by the same company | Depends entirely on the answer to that question. |
Start with the error range, not the age
Every measurement has a margin of error. Biological age assays have a large one, and reports vary widely in how openly they show it.
Two different quantities can appear here, and providers rarely distinguish them. The first is model error, meaning how closely the clock predicts age across the population it was trained on. The second is assay noise, meaning how much two duplicate samples from the same blood draw differ when run through the same pipeline. The second is the one that determines whether your change between tests is real, and it is the one less often disclosed.
If your report shows no error range at all, the working assumption should be that the assay's own variation is comparable to the differences you are trying to detect. That is not a reason to disregard the test. It is a reason to stop reading single results as precise.
How to interpret each result pattern
| What the report says | What it usually means | Reasonable next step |
|---|---|---|
| Biological age below chronological age | Often reassuring, and frequently just regression to the mean on a noisy assay. | Nothing. Do not change a plan on the strength of a single favourable number. |
| Biological age above chronological age by 1 to 3 years | Within the measurement error of many assays. | Check the confidence interval before treating this as a finding. |
| Biological age above chronological age by 5+ years | Worth taking seriously as a prompt, not as a diagnosis. | Review the standard risk markers that do have clinical thresholds. |
| Pace of ageing above 1.0 | The most interpretable adverse result these tests produce. | A reasonable trigger for a conversation about cardiometabolic risk. |
| One organ score far out of line with the rest | Least reliable part of most reports. | Do not chase it in isolation. Confirm with an established marker for that system. |
| A large change between two tests | Usually assay noise, timing, or a different lab. | Compare only same-assay, same-lab results, and look for a trend across three. |
Confounders that explain most surprising results
Before treating an unexpected result as a signal, rule these out. Nearly every one of them shifts the composition of white blood cells in the sample, which is what most clocks are reading.
| Confounder | Why it moves the result |
|---|---|
| Recent infection or vaccination | Shifts immune cell composition, which several clocks read directly |
| Time of day and fasting state | Affects white cell distribution in the sample |
| A hard training block in the preceding week | Transient inflammatory and immune cell changes |
| Changing labs or providers between tests | Different assay, different normalisation, non-comparable numbers |
| Smoking status | GrimAge in particular incorporates smoking-related methylation directly |
| Acute illness at the time of the draw | The single most common reason a result looks abnormal |
The practical control is to standardise your own testing conditions. Same provider, same assay, same approximate time of day, same season, and not within a few weeks of an illness or vaccination. Most of the variation people attribute to their interventions is explained by failing to do this.
Turning a report into something actionable
- Record which clock produced the number. A result from a first-generation clock and one from GrimAge are not comparable measurements, and providers do change their underlying models between versions.
- Write down the conditions of the draw: date, time, fasting state, recent illness, recent training load. This is what makes your next result interpretable.
- Ignore the headline gap unless it is large. Anything inside a few years is compatible with measurement noise on most assays.
- Cross-check any flagged system against a marker with a real threshold. A flagged cardiovascular score means look at ApoB and Lp(a). A flagged metabolic score means look at fasting insulin and HOMA-IR.
- Check who profits from the recommendations. If the suggested action is a supplement the testing company sells, treat it as advertising until shown otherwise.
- Plan the next test before acting on this one. Three results on a consistent protocol produce a trend. One result produces a number.
What a good report contains
Across the providers we compare, the reports worth paying for share four features: the clock is named and cited, an error range is stated, a pace-of-ageing measure is included alongside the headline age, and the recommendations are not a catalogue. Reports that hide the underlying clock, present a single precise-looking number with no interval, and end on a discount code for the vendor's own supplement are optimised for retention rather than for interpretation.
That distinction is visible before you buy, on the provider's sample report. Asking for one is the cheapest quality check available in this category.
Where to read the research
- Higgins-Chen et al. 2022, Nature Aging: the paper on epigenetic clock test-retest noise, and the principal-component clocks built to reduce it.
- Belsky et al. 2022, eLife: DunedinPACE, the pace-of-ageing measure worth looking for on a report.
- Lu et al. 2019, Aging: GrimAge, which incorporates smoking-related methylation directly.
- Horvath and Raj on epigenetic clock methodology: useful background on what these models are trained to predict.
Frequently Asked Questions
What does it mean if my biological age is higher than my real age?
It means the model placed your sample among people older than you, which is a prompt rather than a diagnosis. Before treating it as a finding, check two things: the confidence interval on the result, and whether you were unwell, recently vaccinated, or training hard around the blood draw. A gap of one to three years falls inside the measurement noise of many assays. A gap above five years, or a pace of ageing meaningfully above 1.0, is worth discussing alongside standard risk markers such as ApoB, blood pressure and fasting insulin.
Which number on the report matters most?
The pace of ageing, if the provider reports one. A pace is a rate rather than a snapshot, which makes it directly comparable between two of your own tests and less dependent on how well the model guessed your absolute age. The headline biological age is the number given the most visual weight and the least interpretive value.
Why did my biological age change so much between tests?
Most large swings between two tests are measurement variation rather than biological change. Some methylation clocks show test-retest differences of a year or more on duplicate samples from a single blood draw. Add a different lab, a different time of year, or an infection in the weeks before the draw, and a multi-year swing needs no biological explanation at all. A trend across three or more results on the same assay is the only pattern worth interpreting.
Should I act on organ-specific ages?
Not in isolation. Organ and system-specific epigenetic scores are the newest part of these reports and rest on smaller training datasets than whole-body clocks. If a report flags accelerated ageing in one system, the reasonable response is to check the established marker for that system rather than to act on the score itself. For a flagged cardiovascular score that means ApoB, Lp(a) and blood pressure. For a metabolic score it means fasting insulin and HbA1c.
Are the recommendations in these reports trustworthy?
Check whether the recommended products are sold by the company that produced the report. Several providers in this category sell both the test and a supplement line, and an algorithm that recommends the same company's own products after every result is a marketing funnel with a laboratory attached. Recommendations pointing to sleep, resistance training, protein intake and cardiovascular risk factors are generic but defensible. Recommendations pointing to a specific proprietary formula deserve more scepticism.
Should I show this report to my doctor?
You can, though expect limited engagement, and that is a reasonable clinical position rather than dismissiveness. No major guideline incorporates biological age testing into screening or treatment decisions, so a clinician has no protocol to apply to the number. The parts of the report a physician can act on are the conventional markers if the provider measured any, along with anything genuinely out of range. Bring those rather than the headline age.