For cliniciansThe platform behind an ADHD service

The admin, done before the appointment starts.

ADHDconnect collects a patient's history, questionnaires and collateral before you meet them, and hands you the questions still worth asking. It does not diagnose anyone.

What it actually does What it refuses to do

The idea

A clinician should open a patient file already knowing what that patient has said — and should never have to wonder whether the software has quietly made a decision on their behalf.

Four questions a clinic asks every day

None of them had an answer anywhere in the system before. Each one is now a page.

Who am I ringing today?

A triage worklist of everyone whose history and questionnaires are in and who has not been called yet — longest wait first, with anything flagged in their referral pulled to the top. A separate list shows who is not ready, and names the one thing that is missing.

Who has waited longest?

A live cancellation list ranked by the real gap between a patient’s first and second appointment, plus the patients with nothing booked at all — which is a different problem and shown separately. It is current because it is read from the record, not exported from it.

What is this appointment for?

A structured ADHD interview starts from zero on every criterion, twice, which is most of why it takes ninety minutes. Much of the adult half is already in what the patient wrote. The worklist shows which criteria are covered and which are not, so the hour goes on the gaps.

What have we not done yet?

One queue for every patient request, safety flag, script request and unmatched referral — urgent first, and items leave only when somebody marks them handled. An inbox lets an urgent message be buried. A queue makes silence visible.

Before you meet them

The history, the rating scales, the collateral from a partner or a parent, the supporting documents and the consent are all in before the first appointment — and the patient is never sent back to their GP to chase paperwork.

What is waiting on your desk when you open a file

What they said, in their words

Every questionnaire answer and every history field, verbatim, on the same screen as the call guide — because a patient on the phone rarely remembers what they wrote a fortnight ago.

The follow-ups worth asking

Their answers turned into the specific questions that would settle them, with a place to record what was said. Where two conditions sit on top of each other, the question that separates them.

What nobody has read yet

Correspondence is marked read for risk by a person, or shown as unread. A referral nobody has opened is never displayed as a referral with nothing in it.

The rules for their state

Whether a GP may initiate or continue only, which real-time monitoring system applies, and who issues the approval — each with its source and the date it was last checked. After six months the panel says on its face that it is out of date.

What it will not do

This is the part we would most like you to read. The limits are enforced in the software, not promised in a disclaimer — several of them fail the build if somebody tries to cross them.

It does not diagnose

No diagnosis, no likelihood, no ranked list of what a patient might have. It produces the questions still worth asking and somewhere to record the answers.

It never declares a criterion met

The DSM symptom count required for the patient's age is worked out and then deliberately shown nowhere — not on screen, not in an export. Software that prints a patient's symptom count against a required number has decided the diagnosis, whatever the wording underneath says.

It does not score severity

There is no low, moderate or high. Operational urgency and clinical severity are kept apart on purpose, so a busy queue can never be read as a sick patient.

It never says whether to prescribe

It shows the rule for the patient's state and names the source it came from. The moment software answers “yes, you may prescribe”, it is practising medicine — and it will be wrong the week a regulation changes.

It blocks nothing on clinical grounds

A patient with flags across their record can still book. “On hold” is a label with a review date attached, never a gate, and a hold cannot be recorded without that date.

There is no AI in any clinical logic

Every rule is written down, fixed and auditable. The one AI feature — a draft write-up for a clinician to rewrite — is off unless it is deliberately switched on, and when it is off nothing leaves the country.

In one line

The patient provides. The platform connects. The clinician interprets.

Nothing looks like something

A questionnaire nobody has answered used to render as a row of zeros, which reads as a patient who scored nothing. The platform now distinguishes six kinds of absence — never asked, asked and skipped, sent and not returned, answered as no, nobody available to ask, and not applicable — and shows them differently, because two of them are opposites.

For the same reason a patient who has not completed their history has no tier at all. It reads as not assessed, in grey. A green band over an empty record is the most dangerous thing a screen like this could do.

Built to be audited

Every open, not just every edit

An append-only, hash-chained record of who did what and when, including who looked at a file. Every change to an answer is kept, not only the final value.

Two-factor, enforced properly

Mandatory, and re-checked on every request rather than only at the login page — including the routes that hand out patient PDFs. Sign-in attempts are rate limited and failures are written to the audit chain.

Australian, and clinic-controlled

The application runs in Sydney and the database is in Sydney, in an organisation the clinic holds itself. Patient exports are never emailed.

What we are not claiming. ADHDconnect is not a diagnostic device and is not a validated instrument — the questionnaires inside it are validated; the software around them is not, and does not borrow the word. It does not yet connect to Best Practice. And it cannot show that treatment helped: there is no pre-treatment measurement on the same instrument, so it can show you where a patient is today and we will not pretend it shows more than that.

Questions

Is this a diagnostic tool?

No, and it is built so that it cannot quietly become one. The platform collects what a patient reports, organises it, and puts it in front of a clinician with the open questions attached. Interpretation, assessment, diagnosis and treatment stay entirely with the clinician. Automated checks fail the build if a screen starts declaring a criterion met.

Which questionnaires does it use?

Published, validated instruments, scored against their own published cut points — never an invented percentile. Each has a reference page written for the person reading a result rather than the person who chose the scale: what it measures, how it scores, what the evidence supports and what it cannot tell you. Clinic-authored questions are styled differently from validated ones, so our own wording is never mistaken for theirs.

How does it handle a disclosure of self-harm?

It raises a work item for a person, including when the disclosure is sitting in a health history the patient has not submitted yet. It does not rate the risk, and it does not email the content anywhere. If the notification could not be sent, the queue item says so in red and stands in as the notification.

What happens to the data?

It stays in Australia. The application runs in Sydney and the database is in Sydney, in an organisation the clinic controls. Patient exports are never emailed. Two-factor authentication is mandatory and is re-checked on every request rather than only at the login page.

Can I see who opened a record?

Yes. The audit trail is append-only and hash-chained, and it records every time a file was opened, not only when it was changed. Alongside it is a staff activity view that pairs sign-ins with sign-outs, and an alert if one account opens twenty-five or more different patient files within six hours.

Does it connect to Best Practice?

Not yet. The connection point exists in the code and the integration is being discussed, but nothing is connected today and appointments booked in Best Practice do not appear in the platform. Reports come out as PDFs built to be saved straight into a practice management system.

Can my practice use it?

It is running in one clinic. Whether it is offered to others, and on what terms, is not something we have decided — so there is nothing to sign up to here. If it looks like it would help your practice, we would genuinely like to hear what you would need from it.

Referring a patient rather than running the platform? The page for referrers has what we need and what we do not.

Come and have a look at it.

We would rather show you the real thing than send a brochure.

Start a conversation