Holistic Diagnosis by Pirxey See a sample case →

Diagnostic support for rare and complex chronic disease — in adults and in children

The patient is still waiting. You’re still looking.

A rare disease case spreads its clues across hundreds of pages — test results from different labs, notes written at odd hours, medications and doses, everything you’ve already ruled out. Holistic Diagnosis digs through all of it with you, so the needle in that haystack doesn’t stay buried.

Not a replacement for your judgment — a fuller view of everything your judgment depends on. The kind of sight that makes a great diagnostician great.

See it on a sample case How long a diagnosis takes ↓
A child lying in bed under a white duvet, a parent sitting beside her
Plate 01 Between flares the child does not return to baseline. The family keeps the notes nobody asks for.

The rare disease diagnostic odyssey

A hospital waiting room: a patient reading, a clinician slumped forward on a chair
Plate 02 Waiting is the default state of a rare disease case — for the family and for the clinician.
4.7 years on average
— in some cases, more than ten

On average, that’s how long a rare disease diagnosis takes today. Some patients wait more than ten years — not because the answers aren’t in the record, but because they’re buried across hundreds of pages no one can read.

Patient surveys across Europe and the US; the figure varies by country and condition.

Behind these files are two years of illness.

One real, anonymised case. Sudden onset in a nine-year-old. Two emergency room visits before anyone put the labs next to each other. Hundreds of values per round of testing, from laboratories in three countries that share neither a format nor a reference range. The family paid for every test out of pocket.

Parents on a sofa with a worried boy hugging himself
Plate 03 Sudden onset in a nine-year-old. Two years, five laboratories, twenty-nine files — and the parents’ notes.
Fig. 01 The shape of the illness — what a timeline has to carry
symptom intensity baseline before onset one marker · five laboratories, two methods sudden onset infection → flare, a day early return to school → relapse track 1 · infection track 2 · anti-inflammatory track 3 · psychiatry antibiotics · antifungals · tonsillectomy 5 days on, 2 off · later as needed taper
  1. Sudden onset “like a light switched on”
  2. Two ER visits ambulance called once
  3. First full work-up 5 labs · 3 countries
  4. Relapse after a return to school
  5. Tonsillectomy chronic focus confirmed
  6. First normal ASO in two years

A timeline here is not a chart. The baseline shifts and symptoms don’t return to it between flares. Triggers come with a lag — sometimes a day, sometimes months. Three treatment tracks run in parallel with different rhythms. And one marker measured in five laboratories is not one curve.

  • 2years from onset to a confirmed cause
  • 2ER visits before the labs were lined up
  • 7flares most of them after an infection
  • 29files PDF, scans, Word, e-mail
  • 691values from five laboratories

The cause was in a report nobody had put next to the rest.

A rheumatologist sees one piece, an immunologist sees another — nobody sees the whole patient.

A physician treating complex chronic illness

You keep looking when the case still makes no sense.

A physician at a desk, writing notes by hand next to a computer
Plate 04 The cross-referencing happens here — in a notebook, after hours.

  1. You line up five laboratories by hand.

    Different names for the same marker, a comma for a decimal point, a reference range that changed in October, a result that says “pending — 11 working days”.

  2. You go back to what the parents wrote.

    The flare started the day before the fever, not after. The dose was missed on Tuesday. Nobody will find that in a lab report.

  3. You check the number against the page.

    Because one value copied wrong turns two years of trend into a story that never happened.

A clinician in scrubs sitting on the floor against a window, head in hand, clipboard on his knee
Plate 05 Persistence has a cost. The record should carry some of it.

Your persistence needs a record you can work with.

Today the cross-referencing lives in a notebook and in memory. The systems around you do records and workflow — billing, e-prescriptions, lab feeds, a trend chart for one parameter. None of them does reasoning. What is missing is a record that holds labs, notes, medications and hypotheses in one place, dated, and traceable to the source — so the clue that cracks the case has somewhere to surface, and nothing gets overlooked.

“We don’t lack the knowledge. We lack a tool that sees across.” — physicians treating these conditions

From records to reasoning

Less time piecing it together. More time deciding.

It’s not a chatbot reading your PDFs. Every value is pulled out, dated, and linked to its page — the evidence a diagnosis stands on, ready to review. The call stays with you; the blind spots don’t.

  1. 01

    Drop in the reports. Check what it read.

    PDF, scan, Word, e-mail. Each value comes out with its unit, reference range, laboratory and sample date, and sits next to the page it was read from. Shaky reads get flagged, so you can fix them right there.

    Fig. 02Timeline · Aug 18Source-linked and reviewable
    • ASO111 IU/mLNormal
    • Anti-DNase B354 U/mLHigh
    • CRP< 1.00 mg/LNormal
    • Ferritinpending11 days
    Lab report · page 5Verified
  2. 02

    Find the needle in the haystack.

    Hundreds of pages of labs, parents’ notes, and meds with doses sit on one timeline — each laboratory as its own series, infections and events on the same axis — so you see what moved together. Click a point and land on the exact page.

    Fig. 03APPatient — anonymizedLongitudinal record · 3 active hypothesesSources linked
    Mar ’25Jun ’25Sep ’25Dec ’25Mar ’26Jun ’26Sep ’26
    Labs
    ASO 111 · first normal in 2 yrs
    Notes
    Meds
    Sertraline 75 → 12.5 mgIbuprofen 5/2 cycle
    Events
    RelapseTonsillectomy
  3. 03

    Keep the reasoning behind every hypothesis.

    Each hypothesis has a status, the evidence behind it and the reason it was dropped — with a date and a source. Suggested next steps always come with a source and an estimated price. You make the call.

    Fig. 04 Reasoning record · one anonymised case 7 entries
    1. Lyme & co-infections ruled out Immunoblot IgM/IgG negative Mar ’25 · p. 2
    2. Autoimmune encephalitis ruled out NMDA/AMPA panel negative · MRI, EEG normal Apr ’25 · p. 1
    3. Immune dysregulation working B cells CD19 11.9 % (13–27) Apr ’25 · p. 3
    4. Tonsils as a chronic strep reservoir confirmed Histopathology: chronic active tonsillitis Mar ’26 · p. 3
    5. Celiac disease ruled out tTG IgA 0.72 AU/mL (< 20) Aug ’26 · p. 12
    6. Thyroid involvement ruled out TSH 1.85 · anti-TPO 0.26 Aug ’26 · p. 3
    7. Derealization — SSRI taper vs. anxiety differential Onset during taper to 12.5 mg Sep ’26 · notes

    Every status has a reason and a source. Nothing is deleted.

Holistic Diagnosis helps the physician reach a diagnosis — surfacing the clues scattered across the record and verifying every source. Every clinical decision stays with the physician.

The next diagnosis you make doesn’t have to take ten years.

One anonymised case, two years of records, every value traceable to its page.

See it on a sample case
A mother holding her daughter close, both smiling, outdoors in soft light
Plate 06 What the record is for.