Notes that hold up months later
Most notes are written for the next shift. The ones that matter are read years later by someone who was not there — and the habits that serve them cost almost nothing at the time.
Most clinical notes are written for the next shift. The ones that matter are read years later, by someone who was not there, looking for something specific, and often in circumstances nobody anticipated.
Writing for that reader costs very little at the time and changes the value of the record entirely.
Who reads a note, and when
A psychiatric record has a longer and stranger life than most clinical writing.
- The next clinician, often within hours, forming an impression before they meet the person.
- The on-call team at 3am, who have never met the person and have four minutes.
- A tribunal, where the note is evidence and the writer may be asked to explain a phrase they no longer remember choosing.
- A serious incident review, reading with the outcome already known.
- A coroner, asking whether the risk was considered and what was done.
- The person themselves, who has a right of access and increasingly exercises it.
Write for the on-call clinician at 3am and you will usually satisfy the tribunal too. Both want the same things: what was found, what was decided, and why.
Attribute every account
The single most common failure in psychiatric notes is merging sources. Three streams feed most assessments — what you observed, what the person told you, and what someone else reported — and they carry different evidential weight.
He assaulted a member of the public earlier today and has been behaving erratically in the street.
Police report that he punched a cyclist earlier today. He denies this, and says he was training. His sister reports that he has not slept for three days.
The first sentence states an allegation as established fact. If it later turns out to be wrong, it will already have been copied forward into a dozen entries. The second version costs eight extra words and is accurate.
Specific beats comprehensive
A long note is not necessarily a good one. Three precise sentences are worth more than a page of general impressions, because only the precise ones can be checked, contradicted, or acted on.
Patient presents as unwell with evidence of thought disorder and disorganisation. Mood appears abnormal. Risk noted. Plan: review.
Thought form showed flight of ideas; his account was difficult to follow at several points. He rated his mood 8 out of 10 and said he had slept three hours in three days. He denied thoughts of self-harm. Discussed with the duty consultant; medication restarted; review tomorrow morning.
Quote more than feels natural
A verbatim phrase is the most durable thing in a clinical record. It does not depend on your interpretation, cannot be disputed as a paraphrase, and often conveys more than a paragraph.
Quote in particular: the person's description of their own mood; their explanation of an unusual belief; their reason for stopping medication; and what they say when asked what keeps them going.
He said the police "have been targeting me because of my race". Asked about his diagnosis, he said: "I know they think I am ill, but I have never felt better in my life."
Paraphrasing the first as "expressed persecutory ideas about the police" is accurate and loses almost everything: what he believes, why, and the fact that the allegation is one a reader might reasonably want to weigh on its own terms.
Say what you did not do
A silent gap reads as a normal finding. This is the mechanism behind a surprising proportion of notes that look negligent in hindsight but were not.
If cognition was not tested, the note should say so. If the person left before the interview finished, the note should say when and what had been covered. If collateral was not available, say who you tried.
Not assessing something is a defensible clinical decision made under time pressure. Appearing to have assessed it and missed something is not. One sentence separates them.
Record the negatives
Related, and just as often skipped. If you asked about suicidal thoughts and they were denied, write that down. If you asked about hallucinations in every modality and found none, write that down.
A blank in the risk section does not read as "no concerns". It reads as "not asked", and in a review after an adverse outcome it will be read that way by someone whose job is to establish whether the question was put.
No thoughts of self-harm, suicide or harm towards others were disclosed on direct questioning. Hallucinations in all modalities were denied and he was not observed responding to unseen stimuli.
Two sentences. Fifteen seconds. They convert an absence into a documented enquiry.
Write it sooner
Most clinicians write up after the interview rather than during it, which is usually better for rapport. The cost is memory, and it decays faster than people expect — quotations first, then the sequence of events, then the detail of what was observed.
Two habits help. Note key phrases verbatim at the time, even mid-conversation; most people accept "can I write that down exactly as you said it?" as a mark of being taken seriously. And write the rest up before the next patient rather than at the end of the clinic, when six assessments have begun to merge.
The person can read it
Under the Data Protection Act 2018 and UK GDPR, people have a right of access to their health records, and many services now provide direct access as standard.
This is not a reason to soften findings. It is a reason to make sure every sentence describes something rather than characterising someone — which, conveniently, is also what makes a note useful clinically.
If a sentence would be difficult to justify to the person's face, it is usually because it contains an inference presented as an observation. Removing the inference improves the note.
A test before you save
Six questions, perhaps twenty seconds.
- Could someone who was not there picture this person from what I have written?
- Can a reader tell which parts I observed, which the person reported, and which came from elsewhere?
- Have I recorded the questions I asked that produced negative answers?
- Have I said what I did not assess?
- Is there at least one quotation?
- Does the note say what was decided, by whom, and when it will be reviewed?
Most notes that fail badly under later scrutiny fail on questions two, three and six. None of the three takes more than a sentence to fix at the time, and none can be fixed afterwards.
Common questions
How detailed should a psychiatric note be?
Specific rather than long. Three precise sentences are more useful than a page of general impressions, because only precise statements can be checked, contradicted or acted on. The test is whether someone who was not present could picture the person and understand what was decided.
Should I record that a patient denied symptoms?
Yes. A blank does not read as 'no concerns', it reads as 'not asked'. Recording that suicidal thoughts were denied on direct questioning, or that hallucinations in all modalities were denied, converts an absence into a documented enquiry. It takes one sentence.
Is it acceptable to write notes after the interview rather than during it?
It is normal practice and usually better for rapport. The cost is memory, which decays quickly, particularly for quotations and the sequence of events. Note key phrases verbatim at the time and write up before the next patient rather than at the end of a clinic.
Can patients request to see what I have written about them?
Yes, under the Data Protection Act 2018 and UK GDPR, and many services now offer direct record access as standard. This is not a reason to soften clinical findings, but it is a good reason to ensure each sentence describes an event rather than characterising a person.
What is the most common failure in psychiatric documentation?
Merging sources. What the clinician observed, what the person reported and what a third party alleged all appear in one narrative as though equally established. Attributing each takes a few extra words and prevents an allegation being copied forward as fact.