For wellness programs · health apps · labs
A personalized blood test, read for the person, not the segment.
A subscription program repeats the same panel every few months and calls the result a personalized blood test. Vitrubo is the reading layer that makes it one: each new result is read against the person's own earlier draws, conditions and medications, with the medical guidance behind every statement. For the programs, health apps and labs that run the test, and for the person who holds the result.
- HIPAA compliant
- GDPR compliant
- Encrypted in transit and at rest
- HL7 · FHIR R4
What makes a blood test personalized: the record it is read against.
LOINC codes
Each value from a blood test lands on one reference standard, whichever lab ran it. A subscription that changes lab partner mid-year keeps one history rather than starting a second one.
Sources behind one view
Up to 30 named publications can sit behind a single view of one marker. The chip beside each statement opens the guidance it rests on, for the person and for the clinician.
Native record
One FHIR R4 record per person, every result coded and dated, the source document kept. HIPAA business associate agreement for US deployments, GDPR Article 28 terms and Standard Contractual Clauses as standard.
Records, on premises
Roughly 100,000 anonymised hospital records read inside a clinical program without leaving the perimeter. Numbers, units, ranges and flags are computed by deterministic code, never written by a model.
Who sells a personalized blood test, and what the reading adds.
A subscription testing program, a health app and a laboratory each put the word personalized on the test. Vitrubo does not run or sell the test; it is the layer that reads the result for one person. In every case the professional signs and Vitrubo drafts.
The second test is the reason to stay.
A member joins for the first panel and stays for what the next one says. Vitrubo reads each repeat test against the earlier ones: which values moved, which came back to normal, which are worth watching before the next draw. The report regenerates each time, so the check-in has something new to say and the subscription has a reason to renew.
A personalized blood test inside your product.
Embed, bundle or resell the reading under your brand. A user uploads a blood test as a PDF or a photo, or your lab partner sends HL7, and the report comes back as JSON or as a white-label page written for that person. Each further test regenerates it, which is what brings the user back.
A drafted comment for the person, not the panel.
For labs that run panels for consumer and membership brands, the comment the client reads is drafted in the lab's own style against that client's earlier results, and approved at the release step. The follow-up test worth adding is raised while the sample is still fresh.
What Vitrubo does with a personalized blood test.
Most personalized panels are personalized to a segment: an age band, a goal, a questionnaire. Vitrubo reads the test against the one record that belongs to this person: their earlier draws, the conditions and medications on file, the note from a consultation last year. A value that is unremarkable for the segment can matter for the person, and the reverse.
A person on a subscription for a few years usually has results from more than one lab, in more than one unit. Every value is normalised to LOINC and UCUM units, so a marker's history reads as one line whichever lab drew the blood. Slow drift that looks normal in any single test becomes visible.
Flagged results are grouped into connected stories and placed in four zones: needs attention, worth watching, back to normal, always normal. A repeat test that came back into range is named as such rather than quietly leaving the list, and a value that has never moved is said to be steady.
The report is not a document written once. Each new blood test regenerates it against the whole history, and previous versions stay as history of their own. What a member reads after the fourth draw is a reading of four draws, not the fourth result alone.
Every statement links to the medical guidance it rests on: EHA, WHO, NICE, KDIGO and BSH among others, up to 30 sources behind one biomarker view. A chip opens the publication. Nothing is asserted without a source a clinician can check.
A plain-language view for the person whose test it is, and a clinician view with the full reasoning, connected conditions and medications, and tests to consider. Vitrudoc, an AI chat grounded only in that person's own record, sits beside either view. The same analysis in two registers, never two analyses.
Fifteen months of repeat tests, read as one person's story.
Ryan Hayes is 30, with no diagnosis on file, and repeated the same panel over 15 months of lifestyle change. The live report reads those draws together: which values moved between tests, which came back to normal, and what is worth re-testing next. Every value now sits in its optimal range, but the report shows the direction of travel rather than the destination alone, with the guidance chip beside each interpretation. This is what a personalized blood test looks like when the reading is personal too.
Connects to the testing program you already run.
Full integration.
Results arrive as HL7 or FHIR R4 from your lab partner, the drafted reading flows back as FHIR or JSON, and the follow-up test is raised from the record. No result from a member's blood test is re-keyed by hand.
PlatformWhite label, no engineering.
A role-based portal and PDFs under your brand, for the program, the clinician and the member. A blood test uploaded as a PDF or a photo is read the same way as an HL7 feed.
API first.
Teams with their own app take the API alone: the structured test and its interpretation as JSON, ready to render for one person. Recognition, normalisation and interpretation together or one at a time.
Standards in, standards out.
LOINC for every value, UCUM for units, the record kept in FHIR R4. Containers on your infrastructure or a private cloud, your keys.
Why not paste the result into a chatbot?
Because a personalized blood test is a record, not a prompt.
A general-purpose model
- Reads the values you paste and nothing else: no previous test, no medication list, no history, so nothing about it is personal.
- Misreads a decimal or a unit in a multi-page lab PDF without noticing, and the wrong number becomes part of the answer.
- No normalisation: a marker measured in two units at two labs stays two numbers.
- Confident prose with no source behind it, so nothing for a clinician to check.
- Answers once. The next test starts the conversation from nothing.
Vitrubo's reading of a personalized blood test
- Numbers, units, ranges and flags computed by deterministic code, never written by a model.
- Every value mapped to LOINC and UCUM, so a test from any lab in any year lands on one scale.
- Several models cross-check the same case; disagreements are resolved and omissions caught before a draft is shown.
- Every statement links the named medical guidance it rests on. One click opens the publication.
- Read against the person's own history, grouped into connected stories in four zones, regenerated with every new test.
Every statement rests on named medical guidance. The chip beside it opens the same publication a clinician would read.
From first call to production, on your own members' tests.
An evaluation runs on the panels your program already sells, in your formats. Six steps, none of them a commitment until the last.
Talk to the team
A short call about your panel, your members or users, and the point in the process where the reading should sit.
Sandbox and API docs
Test access for your team, the API contract and sample payloads for a blood test in PDF, HL7 and JSON.
Run your own panels
Send real results, anonymised. Repeat tests from the same person over a subscription are the most telling.
Validate the output
Your clinician or lab scientist reads the drafts against their own reading of the same tests, alone or together with us.
No commitment at this stage
Configuration of ranges, comment style and branding happens here. An evaluation is an evaluation.
Production when ready
Inside your perimeter or a private cloud, connected to the LIS, EHR or your app, or portal-first.
Questions about the personalized blood test.
In the market the phrase usually means a panel chosen for a person's goals or age group and repeated on a schedule, often as a membership or subscription. The choice of markers is personal; the reading, most of the time, is not. Each result is checked against a population range and coloured green or red. Vitrubo does not perform or sell the test. It is the reading layer a program, app or lab adds so that the result is read against that person's own history, conditions and medications, and the layer a person can use to understand a result they already hold.
Three ways. The test is read against the person's earlier draws, so a value is judged by how it moved as well as by where it sits. It is read with the conditions and medications on file, so a marker a medication is known to shift is read with that in mind. And the guidance it rests on is the guidance that applies to that person's situation, not a general leaflet. The result is a reading that would be different for a different person with the same numbers.
Four steps. Parse: the result arrives as a lab PDF, a photo, an HL7 or FHIR feed or JSON, and numbers, units, ranges and flags are computed by deterministic code. Structure: every value is mapped to LOINC and its unit to UCUM, and the test becomes part of one FHIR R4 record for that person. Understand: values are read together, against previous tests and the record on file, with named guidance cited claim by claim. Deliver: a patient view, a clinician view, a PDF or JSON through the API. Previous versions of the report stay as history.
Because the first test sells the membership and the second test keeps it. If the second result is the same table with different colours, the member has little reason to open it, and less to renew. When each new test regenerates a report that says what moved, what came back to normal and what is worth watching, the check-in has content and the program has a story that continues. The program keeps running the tests it already runs; Vitrubo reads them.
Yes, and for anyone who has been on a subscription for a while this is the usual case. Labs name the same marker differently, report it in different units and publish their own reference ranges. Every value is normalised to one of 17,000+ LOINC codes and to UCUM units, so a test from one lab and a later test from another read as points on one line. Reference ranges are kept per laboratory rather than replaced, because the range that applies to a value is the one the lab that measured it published.
Rather than a table of pass and fail, flagged results are grouped into a few connected stories and placed in four zones: needs attention, worth watching, back to normal, always normal. For each marker the trend across draws sits beside the latest value and its range, with the guidance behind the interpretation one chip away. Where the guidance calls for a repeat, the test to re-run and the interval are named. The Ryan Hayes live report is an example: 15 months of lifestyle change, read as one person's direction of travel.
No. Vitrubo is not a diagnostic device, does not select treatment and does not predict what a marker will do next. It drafts a reading of the values that are there, organises the history and names tests worth considering; a clinician reads and signs where one is involved. For a person reading their own blood test it is educational context to bring to an appointment, not medical advice.
Neither. Vitrubo runs no laboratory, ships no kit and has no consumer subscription of its own. The tests are sold and run by the program, app or lab, and Vitrubo is licensed to them as the reading layer, under their brand. Access is by invitation after an evaluation on your own anonymised panels; there is no self-serve signup and no published price list.
Any marker a laboratory reports as a value, in any panel: metabolic and lipid panels, blood count, inflammation, iron and vitamin status, thyroid and other hormones, kidney and liver markers, and the wider panels wellness and longevity programs sell. Urine and allergy tests arrive as values too. Hospital records, prescriptions and notes are read as context, so the interpretation of a blood test knows what else is on file for that person.
Through the API or a white-label portal. A user photographs a blood test or uploads the PDF, or a lab partner sends HL7, and the app receives the structured FHIR R4 record and the interpretation as JSON, or renders the branded page. Every further test regenerates the report, which is what brings a user back after each draw. Vitrudoc, an AI chat grounded only in the user's own record, can answer questions about the result beside it.
Containers run on your infrastructure or in a private cloud; you own the keys and Vitrubo has no access to the data. Identity never enters the prompt: the models see values and an internal ID. A HIPAA business associate agreement for US deployments, a GDPR Article 28 processing agreement and Standard Contractual Clauses, encryption in transit and at rest. A member's test never leaves your perimeter in identifiable form.
Yes, through a lab, clinic, program or app that offers Vitrubo under its own brand. The person uploads the PDF or a photo of the test and reads the plain-language view: what stands out, why it is grouped that way, and what to discuss with a doctor. Earlier tests uploaded alongside it are read into the same history, so one test is read on its own and two or more are read as a line.
With a call, then sandbox access and API documentation. Your team runs its own anonymised panels, ideally repeat tests from the same people over a subscription, and validates the drafts against its own reading. There is no commitment at that stage; production follows when you are ready, inside your perimeter or a private cloud.
Demo cases
Invented people, real product. Every case is a synthetic record with a stock portrait: no real patient, no real result. The reports are live.
Robert Hensley
58A year of CKD follow-up with full context supplied.
Emily Carter
34Five visits over nine months with no diagnosis supplied.
Ryan Hayes
30Fifteen months of lifestyle change. The direction of travel is read.
See it on your own members' tests.
A walkthrough on the panels your program already sells, sandbox access for your team, and a drafted report on your own anonymised tests. Or open a live report first.