Health, Sleep and the Limits of Self-Data
Where time data meets health data, what the one measurement with clear evidence is, and where self-tracking stops being useful and needs a professional.
General orientation, not medical advice. Persistent problems with sleep, energy or mood are a reason to see a doctor rather than to collect more data. In a workplace setting, this meeting workflow can make time and computer activity visible, but the boundary in “Health, Sleep and the Limits of Self-Data” still applies: purpose, notice, access, retention and a correction route must be agreed before collection starts.
Time data touches health at one clear point and several unclear ones. The distinction is worth keeping.
The clear one
Sleep duration and timing.
The effects of insufficient sleep on attention, mood and judgement are well established, and the measurement is simple and interpretable.
If you track one thing about yourself, this is the one with the best evidence behind it, and it explains more variation in how weeks go than any allocation figure.
Bedtime and wake time, written down. Nothing more is needed.
The unclear ones
Energy levels, which are real, poorly measured by self-report, and heavily influenced by what you expect to find.
Productivity by time of day, where the popular chronotype categories have weaker support than their confidence suggests.
Step counts and rings, which measure movement rather than health and are sold as the latter.
And any device figure presented with two decimal places about a thing biology does not hold that steady.
Wearable sleep figures
Consumer devices estimate sleep stages from movement and heart rate, and the stage breakdown is not reliable.
Total duration is reasonably good. The rest is a plausible fiction.
Which matters because people make decisions from the fiction, and because a poor score can itself produce a worse night — a documented effect.
Use the total, ignore the breakdown.
When tracking harms
Sleep data producing anxiety about sleep is common enough to have a name in the clinical literature.
If checking the score is making the night worse, stop checking. This is the clearest case in this collection where the measurement is the problem.
Similarly for food, exercise and mood logging where a history of disordered patterns exists — in that situation self-tracking is a matter to raise with a professional rather than a method to adopt.
What time data can honestly tell you about health
Whether the week contained any rest.
Whether sleep is being displaced, and by what.
Whether a bad stretch coincides with something identifiable.
And whether something has changed, which is the strongest use: a trend over months is more informative than any absolute figure.
Where to stop and ask somebody
Persistent tiredness that sleep does not fix.
Mood changes lasting weeks.
Anything that worsens when you pay attention to it.
And the situation where you have a great deal of data and no improvement, which is a common and quietly distressing place to end up.
None of these is a tracking problem, and continuing to measure delays the thing that would help.
The honest boundary
Time records show how a week was arranged.
They do not diagnose anything, and the confident inferences people draw from personal data about their own health are mostly unsupported.
Use them to notice and to describe, then take the description to somebody qualified if it warrants that.
What to check
Do you record sleep at all — the two simple numbers?
Are you making decisions from a stage breakdown that is not reliable?
Is any measurement making the thing it measures worse?
And is there something you have been tracking instead of asking about?
The two numbers
Bedtime and wake time, written down.
Nothing else about yourself has comparable evidence behind it, and it explains more variation in how weeks go than any allocation figure.
Worth keeping in mind
Use the sleep total and ignore the stage breakdown. The stages are a plausible fiction, and people make decisions from the fiction — including decisions that make the next night worse.
In short
Consumer devices estimate sleep stages from movement and heart rate, and the breakdown is not reliable.
Total duration is reasonably good; the rest is a plausible fiction that people nonetheless make decisions from.
The point
Bedtime and wake time.
That is the whole recommendation, and it has better evidence behind it than anything else in this collection. For a complementary perspective on work, measurement or planning, consult the Information Commissioner's Office.