The care plan, the Act, and the return

Crisis plans and advance statements where somebody can find them at three in the morning, detentions counted on the front page, clustering that keeps the clinician’s name on it, and a monthly return that shows what it is missing.

1 The care plan, and the two bits people actually need

A mental health care plan is only useful at three in the morning. commvita puts the two things that matter then at the top of every plan: the crisis plan, and what the person said they wanted while they were well.

Each plan names a coordinator and their profession — a community mental health nurse, a social worker — and a next review date. Underneath sits the crisis plan with the number to ring, and the advance statement: “patient prefers home treatment. Hospital admission only if risk to self or others.”

An advance statement written into a care plan and never seen again is a form. One that appears on the same screen as the crisis number is a document that changes what happens on the night.

2 The Act, counted where people can see it

The records dashboard opens with a sentence, not a chart: two patients are currently detained under the Mental Health Act. Then the counts — active CPA, enhanced CPA, active detentions, high-risk patients, average HoNOS, continuing healthcare pending.

Detention is the most serious thing a health service does to somebody. A service that has to run a report to find out how many people it is currently detaining has the wrong instrument.

Behind the dashboard sit the CPA register, the detention records, HoNOS and risk, patient rights and advocacy, outcome measures and the community team board. Depot and long-acting injectable medication has its own tab with what’s due, and seven-day follow-up after discharge is tracked as a compliance figure, because that window is when the risk is highest.

3 Clustering, and being honest about it

The clustering tool holds the twenty-one HoNOS-based super-classes, with the severity banding and who assigned each one. It’s a mandatory part of the national dataset and it’s judgement, so the record keeps the clinician’s name and the date against every assignment.

Three rows here read “name not linked”. That’s the demonstration data being honest: the cluster exists, the person it belongs to hasn’t been joined up on this instance, and the screen says so instead of inventing a name.

4 The return that shows you what you’re missing

The mental health services dataset goes to NHS Digital monthly, and the submission screen shows the current period as a draft with its record count, cluster count and PERS count — plus, at the top, cluster coverage at 65.4% and PERS completion at 53.6%.

Those two percentages are the useful part. A return can pass validation with zero errors and still be built on half the caseload. Putting coverage next to the error count is what stops a clean submission being mistaken for a complete one.

The page also carries a plain warning that it’s running on seeded demonstration data and not a live feed, which is the honest thing to say about a screen showing 1,842 patients.

Where it lives in commvita