Messages

From the contact form. The row is the record and the email is only a notification, so a message is here whether or not the email about it arrived. Relay says which: “no” means the message reached us and the notification did not.

Message

People

Newest activity first, and counts only: how many tools each person has favourited, opened and saved. Never which ones. What is inside a saved calculation is not shown here at all. For which tools get used across everybody, read the Usage tab.

This is where Pro is granted. Select a name, and the panel below carries the control. A Pro user opens any tool marked 'access' => 3 in config/tools.php, which is the level to use for a tool you want a chosen group to see before it is published. An administrator already opens those without the flag, so granting Pro to one is refused rather than recorded.

The selected account

Delete this account

This deletes the account and everything in it, at once and permanently. It is also the only way to end somebody’s access: the allowlist decides who may open an account, not who may sign in, so an account keeps working after its address leaves the list.

Specialities

Accounts

What is stored

A “series” is one patient on one tool for one clinician. The same patient seen by two clinicians counts twice and cannot be matched across them, because each account derives its keys from its own salt. That is the design working, not a counting error.

Most favourited

Most opened

One row per person per tool, not one per visit: opening the same tool twice in a day does not count twice. This is “how many people use it”, not “how busy it is”.

What these numbers are, and what they are not

Publication view. Hide any count below five. Turn this on before you copy, screenshot, or put any of this in a document: a rare conclusion on a single day in a small setting is a thin sliver of a real patient, and this screen is the only place the exact figure belongs.

What each column means

Every counting tool, hardest to finish first

The tools people open and cannot finish sit at the top. Each rate is printed with the number it is a rate of, because a percentage with an invisible denominator is the easiest figure on this page to misread.

How far each tool gets

Opened, then finished, then copied into the record. The gap between the first two bars is the tool asking for something the reader could not give it; the gap between the last two is an answer that never left the screen.

What each tool concluded

The bands the engine reached, and the refusals. A band that never fires in a year is either a branch that was removed or a threshold that is wrong, and both are worth opening the module for.

What the visits were like

The same few questions asked of every tool, and none of them is about a patient. How long the answer took, what it was read on, how the reader arrived, and whether the answer followed a refusal.

Week by week

The share of openings that reached an answer, one point per week. A tool that was changed should move here, and a tool that quietly stopped working falls off a cliff. Drawn only where there are at least three weeks to draw.

What could not be supplied

Required answers that were still empty when the page was closed, by field name and never by content. A field at the top of this list on a tool built for a district hospital is the tool asking for something the setting does not have.

The catalogue

Counted from the registry, not from the database. Restricted means a tool carries an access level below 4, so it is off the public hub and opens only for the levels named against it. Unlisted is the editorial decision, and no level brings one back.

Due for review

Every module is re-checked against its sources at least every 18 months. Listed tools only: a tool still behind a gate is not in front of anybody. Never reviewed comes first, then the longest wait. Stamp one with node tools/review.js <slug> from cdss-dev, and never by hand.

Light and dark

Ticked means the reader of that section can choose Light, Dark or follow their machine, from the gear in the bar. Unticked takes the choice away: the control disappears and every page is light, including for somebody whose browser had already been set to dark. Nothing else about the section changes. It takes effect on the next page a reader loads, and on a page already downloaded for offline use it can take a few hours.

Sign-in email, last 24 hours

This is the part worth checking after a quiet complaint that “the site never emailed me”. Two daily ceilings can stop a code: one on how many are delivered across the whole site, and one on how many go unredeemed to a single address. Neither says anything to the person it stops. A clinician who gets no code sees a form that appeared to work, so this card is the only place either becomes visible.