Method
How the Profile is built.
What the picture contains
Psychiatry decides on a very thin sample.
An antidepressant decision is usually made from little data: what the patient remembers at intake and what they report next visit (which might take months).
What would actually inform the decision is different in kind.
Illustrative
Two accounts, both from memory
The Orchard Profile
Four conversations, transcribed and scored in full
The conversation is designed, and a person conducts it.
A patient meets a trained Orchard care manager for a series of semi-structured conversations across the first weeks of treatment or intake. The sessions are non-therapeutic by design: nobody is treating, advising, or intervening. They exist so the patient can describe their experience at length, in their own language.
The sequence is designed so that a few conversations surface the breadth of a person’s health, rather than circling the complaint they led with.
Over 80 domains, each drawn from the literature.
80+
3–7
Every session is read against a set of over eighty health domains drawn from the depression literature. The same frame is applied to every conversation, so each patient gets the same discipline regardless of what they chose to talk about that day.
In addition, the Profile identifies several personalized dimensions that define this person’s illness, in their own words rather than in scale language. Two people with the same diagnosis rarely share the same dimensions, and the best treatment decisions view that in full.
Each dimension carries a severity for every session, anchored to something the patient actually said.
That reading is done by our clinical intelligence engine: frontier AI models directed by methods our clinicians built, and checked by a clinician before anything enters the report.
The Profile tells you what it has not heard.
We analyze the full set of domains every conversation. this means the Profile knows what surfaced and what did not. Both appear in the report.
surfaced across four touchpoints
The remaining forty are printed as not yet assessed.
One demonstration case, not a typical figure
A report that shows only what it found invites the reader to treat silence as absence, which is one of the ordinary ways a symptom gets missed. Marking the unexplored ground points the next conversation at the faint parts of the picture.
Movement, and what moved alongside it.
Because we speak to you several times, each dimension carries a line rather than a point. The Profile shows how each symptom has moved between sessions, and places medication changes and life events on the same timeline, so a shift can be read against whatever preceded it.
From a demonstration Profile
The Profile makes no causal claim about those associations. It sets them side by side, in order, and leaves the interpretation to you and your clinician.
What the current regimen reaches, and what it leaves alone.
The Profile sets the patient’s dimensions against their current medications and describes what the most up-to-date research says each drug targets, does not address, or may aggravate.
Every one of those statements carries a citation, and every citation is visible in the report. A prescriber can read the source rather than take the sentence on trust.
Click a citation to read its source
Sertraline has substantial support for core depressive severity. The literature does not establish an effect on the early-morning waking this patient describes.
Insomnia and fatigue are among the symptoms most often left behind in patients who otherwise respond to an SSRI.
A dose increase may transiently aggravate the anxiety dimension, which is prominent in this patient’s first two sessions.
The Profile does not select a medication, rank options, or predict how someone will respond to a drug. It assembles what the literature already says about this symptom picture and this regimen.
Auditability
Assembled in parts, so it can be checked in parts.
The report is assembled in discrete sections rather than written as one continuous narrative. Some are structured records, some are analysis, and the pharmacological section is entirely cited. Each part can be judged on its own terms, so a reader who disagrees with one is under no obligation to discard the rest.
01
Every analysis must carry its evidence before a report will assemble.
02
"Recommendation" language is never inserted and audited before a report is released.
What the Profile is not.
It is not a diagnosis
The Orchard Profile does not diagnose. Where a diagnosis appears in your materials, it was recorded from the referral and intake paperwork.
It does not choose the medication
The Profile does not select, rank, or predict a medication for anyone. It maps how symptoms present and cites what the literature says; the prescriber decides. That is the design, not a limitation of the current version.
It is not a treatment
Nothing about the Profile is claimed to make anyone better. It is intended to make a decision better informed.
It is not a medical device
The Profile is exempt from FDA device regulation as non-device clinical decision support under the 21st Century Cures Act. This is an exemption, not a clearance. It applies because the clinician remains the decision-maker and can independently review the basis of everything in the report.
It is not for an emergency
The Profile takes two weeks by design. If you are in crisis, call or text 988 in the US.
See the method on a real report.
Thirty minutes with the clinical team, or hold a spot for yourself while your state opens.