Record the clock time it starts, what happened in the two hours before it, and what the previous night's sleep was like. Adjectives describe a bad day; times describe a pattern. Two weeks of start times in a column will show whether it is drifting earlier, which is information a clinician can act on and memory cannot supply.
It has a name, which surprises people the first time they hear it. The late-afternoon and early-evening restlessness, confusion and agitation that many people with dementia experience is common enough to be recognised and discussed — and if you have been living with it, learning that it is a known pattern rather than a personal failure is worth something on its own.
What is less well known is how much the recording of it matters. Sundowning described in words is a complaint. Sundowning recorded with times is data, and the difference decides what a clinician can do with it.
This guide covers what to write down, why the clock is the most important column, and what a fortnight of entries can show that a month of remembering cannot.
"Difficult in the afternoons" is a description. It is true, it is useful to say once, and it does not change over time in any way you can see.
A start time does. Write down when it began — 4:40, 4:15, 3:50 — and after two weeks you have a column. A column drifts, or it does not. If it is creeping earlier, that is a change; if it is stable, that is also information; if it is scattered with no pattern, that tells you to look at what else varies.
None of that is visible day to day. Living through it, every afternoon feels like the same afternoon, and the drift of forty minutes over three weeks is precisely the kind of change that a person in the middle of it cannot perceive and a column of times makes obvious.
This is the general principle behind keeping any log, and sundowning is where it pays off most clearly: you are not writing to remember, you are writing to be able to compare.
What happened in the two hours before. Not everything — the notable things. A visitor, a bath, a change of room, a television left on, an appointment, an unfamiliar person, a nap. Triggers are often invisible until they are written next to a time three days running.
What the previous night's sleep was like. Sleep and afternoon agitation interact, and the interaction runs both ways. A broken night frequently produces a harder afternoon, and a harder afternoon frequently produces a broken night. Recording both is how you find out which way it runs for the person you care for.
What helped, and what did not. This is the column caregivers most often skip and most often need. Dimming the lights, closing the curtains before dusk rather than after, a familiar piece of music, a walk, food, quiet, company, being left alone. Some of these will work and some will not, and which ones work is individual enough that no book can tell you — only your own record can.
Three columns and a time. That is the whole of it, and it fits in the space of a few lines at the end of the day.
The aim is a record another person can act on, so favour the concrete over the interpretive.
| Write this | Rather than this | Because |
|---|---|---|
| Started 4:40, settled about 6:15 | Bad afternoon | A time can be compared; an adjective cannot |
| Woke three times, up from 4:30 | Slept badly | Three wakings and one waking are different nights |
| Paced, kept asking to go home | Agitated | Names the behaviour so a clinician can recognise it |
| Curtains closed at 4, lights on — calmer by 5 | Tried to help | Records what worked, so it can be repeated |
| Visitor 2–3pm; grandson, unfamiliar | Had company | Turns a possible trigger into a checkable one |
| Ate about half of lunch | Not eating well | A proportion tracks; a judgement does not |
Four things, none of which is available any other way.
Whether the start time is moving. Earlier over weeks is a change worth reporting; stable is worth knowing too, because it means something else is driving the variation in how hard the afternoons feel.
Whether it tracks sleep. If the worst afternoons follow the worst nights with any consistency, the conversation with a clinician becomes about sleep rather than about afternoons — which is a different and often more tractable problem.
Whether a trigger repeats. Three difficult afternoons that each follow an afternoon visitor is a hypothesis you can test in a week, and you would never have seen it without the dates.
And whether what you have been trying is working. Caregivers frequently keep doing something that helped once. The record tells you whether it is still helping, which is a question that is genuinely hard to answer from memory.
Bring the pages, not a summary. A summary is your reading of the pattern; the pages let the clinician read it themselves and notice things you did not.
Lead with the times. "It starts between four and five, and over the last fortnight it has moved about half an hour earlier. The worst days follow the nights with two or more wakings." That is fifteen seconds and it is more clinically useful than any amount of description.
Say what you have tried and what happened. It saves being told to try something you have already tested, and it shows the pattern of what does and does not help this particular person.
And mark the date if anything changes as a result — a new medication, a changed routine, a different evening schedule. The fortnight either side of that date is what tells you whether the change worked, and it only works as evidence if you know exactly when it started.
One boundary worth stating plainly: a caregiver's log is a record of observations, not a medical record, and not a diagnosis. What it is for is making a short appointment productive. Decisions about medication and treatment belong to the clinical team, and their current instructions win over anything written in a log book.
The clock time it started and the time it settled, what notable thing happened in the two hours before, what the previous night's sleep was like, and what you tried and whether it helped. Four short entries at the end of the day.
Because times can be compared across days and adjectives cannot. A column of start times shows drift, stability or scatter — none of which is visible to someone living through it one afternoon at a time.
Two weeks is enough to show a pattern and short enough to read in a fifteen-minute appointment. If you have not started, start now rather than deciding it is too late for the next visit.
Yes, and they are more useful than they feel. A good afternoon following a good night is exactly the evidence that establishes the link — a record of only the bad days cannot show a pattern, only a list of complaints.
Not reliably, which is the whole argument for recording it. What settles one person can unsettle the next. Your own log is the only source that can tell you what works for the person you care for.
