Investigations

The Investigations tab holds your patient's lab and blood test results. It exists for one reason: to be a faithful, unaltered record of what the lab actually reported — nothing smoothed over, nothing recalculated, nothing decided on the patient's behalf.

What you'll find here

Results are grouped into result sets — for example a blood panel or a set of ECG findings — each dated and, where relevant, tagged with the specimen type. Open a set and you see every analyte as a row: the value, the unit, the reference range, and any abnormal marker.

Stored verbatim, always

Every field on a result — the value, the unit, the reference range, and the source marker (the H, L, or * a lab prints next to an abnormal reading) — is stored exactly as it appeared on the source. MIA does not compute whether a value is high, low, or out of range, and it never assigns a severity, colour, or risk band to a lab result.

If a printed marker is present, you'll see it next to the value, shown as transcribed. If the source didn't mark it, nothing is inferred in its place. What you see in Investigations is what the lab actually said, not MIA's opinion of it.

Append-only, not editable

Once a result set is filed, it's locked. There is no button to edit a value in place, because a lab result is a historical fact — what was reported on that date, from that specimen. If a result needs correcting (a lab reissues a report, or a transcription turns out to be wrong), the correction is filed as a new result set. The original stays exactly as it was.

This means the Investigations tab can show more than one entry for the same test date if a correction was made later. That's expected — it's the audit trail working as intended. You're never looking at a value that's been quietly changed underneath you.

AI interpretation (optional, off by default)

Some clinics have AI interpretation switched on for Investigations. Where it's enabled, you'll see a separate panel under a result set offering to generate a plain-language read of the values — one observation per analyte, in ordinary language.

This is deliberately narrow. The interpretation:

• Never assigns a severity, risk band, or colour. There is no "this is concerning" or "this needs urgent review" — MIA doesn't make that call.
• Never recommends an action. No suggested referral, no suggested test, no next step.
• Sticks to neutral observations about what a value is and, where useful, what it might relate to in plain terms — not a diagnosis, not a judgement.

If your clinic has this switched on, generating an interpretation is itself a permission a clinician needs to be granted — it doesn't appear for everyone by default, even where the feature is active for the clinic.

Verifying an interpretation

An AI interpretation sits in an unverified state until a clinician acts on it. You have three options for each one:

• Accept — the observations are correct as generated.
• Accept with edits — mostly right, but you're recording that you changed something.
• Reject — the interpretation doesn't hold up; note why if it's useful for the record.

You can add a short verification note with any of the three. Until one of these actions is taken, the interpretation is marked as not clinically validated and carries no weight in the record — it's a draft, not a finding.

Why it works this way

Investigations is deliberately conservative. A lab value is a clinical fact with real consequences if it's misread, so MIA's job here is narrow: capture it accurately, keep it unaltered, and offer a second pair of eyes only as an optional, clearly-labelled aid — never as a substitute for reading the actual result.