Opening an intervention and finding out whether it helped

An intervention names the concern in a human's words, and names an owner. The owner is required rather than optional, because an intervention nobody owns is a wish — and a list of wishes is what pastoral tracking usually decays into.

It is either immediate or a support plan. The level the student was at when it opened is recorded at that moment, so that any later comparison is against what the school could actually see at the time rather than against a number that has since moved.

• The concern, in the words of the person who noticed it.
• A named owner, always.
• Optionally linked back to the risk review that prompted it.

Actions carry a description, an owner and a due date, and are ticked off as they are done. Adding the first action moves the intervention from open to in progress on its own — the status follows the work rather than needing to be maintained alongside it.

Overdue actions are surfaced across the list. That is the number most worth watching, because an intervention that has quietly stopped moving looks identical to one that is going well until somebody checks the dates.

Contacting the family is recorded as its own fact, with the moment it happened and a note of what was said. Schools are regularly required to show that a family was told, and an inference drawn from a ticked action is not evidence of that.

This is deliberately separate from the actions list. An action can say phone home; only the contact record says the call happened.

Closing asks what the person closing it thinks happened, and records the level at that point. Both go into the record: the account and the measurement are different kinds of evidence and neither replaces the other.

Cancelling exists separately from completing, for an intervention that was opened and then overtaken — a student leaving, or a concern that turned out to be something else. Cancelled is not a failed completion.

The outcome is phrased as what followed rather than what worked. An improvement is reported as what happened next and explicitly not as proof the intervention caused it, because a school that starts reading correlation as proof will keep doing the thing that happened to precede a good term.

Unchanged is treated as a real result rather than a failure — some support holds a situation steady, and that is the intended outcome for plenty of interventions. A student who stopped being recorded produces no comparison at all, and is flagged as something to check rather than counted as either success or failure.

• Disagreement is only flagged in one direction: somebody claiming it helped while the signals worsened or went quiet.
• The reverse is never flagged — a person saying it did not help while the numbers improved may know something the data does not.