Our scoring methodology
Every score on Compare Health Test is calculated from a published formula over the facts we hold about each test and provider. This page documents exactly how that formula works.
Principles
Our methodology is built on three principles. First, scores should be reproducible: the same inputs always produce the same score, and this page tells you exactly how. Second, scores should be honest about their own confidence: every score says how much of the underlying data the provider actually publishes. Third, scores should be useful: they should help you choose between tests that measure similar things, not just rank the whole catalogue abstractly.
Scores are calculated, not hand-picked. We hold a structured record for every test (what it measures, how the sample is collected, how long results take, who reviews them, what it costs) and a set of facts about each provider (customer ratings, laboratory accreditation, regulatory registration, how long they have traded). A published formula turns those records into six criterion scores and one overall score. Nobody at Compare Health Test, and no provider, can nudge an individual score up or down.
Evaluation criteria
Each test is scored from 0 to 10 on six criteria. The overall score is the weighted average of the six, shown out of 10. The weights are:
Clinical depth
25%How much the test measures relative to the tests it competes with. Where a test belongs to a comparison group of three or more like-for-like tests, we rank its biomarker count within that group: the group median scores about 7.5, the most comprehensive test in the group about 10, the least about 5.5. A single-marker test in a single-marker group scores a flat 7.0, because depth is not what you are buying. Tests with no comparison group are ranked against every test in their category instead, on a wider 4 to 10 scale. A test whose provider publishes the full list of markers earns an extra 0.5. If the provider does not say how many markers are measured, the test scores 3.5.
Clarity and documentation
15%How completely the provider describes the product before you buy. Every test starts at 2, then earns 2 for a detailed description (1 for a short one), 2 for publishing the full list of biomarkers (1 for a partial list), 1 for a stated turnaround time, 1 for explaining the collection options, 1 for stating the sample type and 1 for describing the report you receive.
Value for money
20%Price against the closest alternatives. Inside a comparison group we compare each test's price with the group median: a test at the median scores 6.5, one 20% cheaper scores about 7.9, one 20% dearer about 5.1. Tests without a group are compared with the median price of similarly sized tests in their category (between half and double the biomarker count, at least five of them) on a gentler scale, so that a test 20% cheaper than that median scores about 7.5. Where there are fewer than five comparable tests, value is scored a neutral 6.0. Value scores never drop below 2 and never exceed 10. Tests without a UK price are not scored at all.
Collection convenience
15%How easy the test is to do. Every test starts at 3. A sample you take at home without a needle (a cheek swab, saliva, urine, stool or breath sample) earns 4; a finger-prick blood sample you take at home earns 3, plus 1 if a venous or clinic option is also available; a home nurse visit earns 1. Speed is judged against tests of the same kind, because a DNA panel takes longer to sequence than a lipid panel: results in half the category's typical turnaround or better earn 2, results at or faster than typical earn 1, slower earns nothing.
Trust signals
15%Facts about the provider rather than the individual test, so every test from the same provider shares this score. Every provider starts at 3. Customer ratings add 3 for a Trustpilot score of 4.5 or higher from at least 500 reviews, 2 for 4.0 or higher, 1 for 3.5 or higher, and 1.5 where we have no rating on record. A stated laboratory accreditation such as UKAS ISO 15189 adds 2, Care Quality Commission registration adds 1, doctor review of results as standard adds 1, and ten or more years of trading adds 0.5. We record the sources and check dates for these facts and refresh them periodically.
Results and follow-up
10%Who helps you understand the results, and what you get. Every test starts at 3, then earns 4 if a doctor reviews and comments on your results as standard, or 3 if a one-to-one consultation with a practitioner is included in the price (a doctor-reviewed test that also includes a consultation earns 1 more). Where the provider describes the report you receive, the test earns 1. We record an included consultation only when the provider's own listing says it is part of the price; an optional paid add-on does not count.
Each criterion score is multiplied by its weight and the six are added together. A test scoring 8 on everything gets an overall 8.0; a test that is cheap but shallow can still score well on value and poorly on depth, and the page shows both so you can weigh them yourself.
Confidence scores
Every overall score carries a confidence figure from 1 to 10. It measures how much of the input data the provider actually publishes, not how good the test is. Six inputs count: the number of biomarkers, the named biomarker list, a stated turnaround time, a clear collection method, whether results are doctor-reviewed, and a UK price. A test with all six known scores 10; each missing input costs one point; being part of a like-for-like comparison group adds half a point because the relative measures are then more meaningful.
- 8–10
- The provider publishes almost everything we need. Treat the score as a fair like-for-like comparison.
- 5–7
- One or two inputs are missing, usually the biomarker list or the turnaround time. The score leans on defaults for those criteria.
- 1–4
- Limited information available. The score should be treated as indicative only, and it is worth reading the provider's page before buying.
We do not currently buy tests to evaluate them first-hand, and no score on this site claims that we have. If that changes, the confidence scale will be updated to say so.
Written reviews
The verdict, summary, pros and cons and the descriptive text on each test page are written from the same structured record that produces the score. First drafts are produced with AI assistance and are marked for editorial review; an editor checks them against the provider's own page before they are signed off. Written content never changes a score. If you spot a factual error in a review, our corrections process applies.
Ranking logic
Within a category or provider listing, tests with a UK price appear first, ordered by overall score from highest to lowest, then by price from lowest to highest, then alphabetically. Unscored tests appear after scored ones. Comparison tables can be sorted by any column. We do not apply manual rank adjustments: if a test appears above another, it is because it scored higher or, on a tie, costs less.
Pricing data
Prices are sourced directly from provider listings and updated on a rolling basis. We do not guarantee prices are current at time of viewing — always verify on the provider's website before purchasing. Where a sale price is detected, we label it as such.
Score reviews and corrections
Scores are recalculated whenever our record of a test or provider changes, which happens each time we refresh a provider's catalogue and prices. The date of the last calculation is shown on every scorecard. If you believe a score is materially incorrect, please contact us with the specific concern and any supporting evidence. We will review within 10 working days.