UK sickness absence statistics

The two figures, side by side

Sources: ONS, Sickness absence in the UK labour market and CIPD, Health and Wellbeing at Work.

Why do the two figures differ so much?

Three reasons, and none of them is that somebody is wrong:

1. Who is asked. ONS surveys households, so absence is self-reported by workers. CIPD surveys employers, who read it off HR systems. People under-report their own sick days; systems do not.
2. What counts as absence. ONS only counts it where someone actually worked fewer hours than contracted. Make the time up later and, for ONS purposes, it did not happen. An HR system logs it either way.
3. Who is in the sample. CIPD's respondents skew towards larger organisations with formal HR functions - exactly the employers most likely to record every absence properly.

Which should you benchmark against?

If you are comparing your own HR records to a national number, use CIPD's 9.4 days. It is the closer like-for-like, because it comes from the same kind of source your figure does. Benchmarking recorded absence against the ONS household figure will make almost any employer look bad for no reason.

Use ONS 2.0% for economy-wide comparisons, or when quoting a figure that has to match what appears in the news - most coverage uses ONS. The one thing worth avoiding is quoting either without saying which, because that is how two people end up arguing about whether 5% is good.

Work out your own with the absence rate calculator, which shows both benchmarks scaled to whatever period you measure.

How has absence changed?

That is a 62% rise on pre-pandemic levels in employer-recorded absence. CIPD puts mental ill health as the leading cause of long-term absence and the second most common cause of short-term absence.

One caution when reading any 2026 figure: since 6 April 2026 Statutory Sick Pay has been payable from day one, so short absences that once went unpaid and unrecorded now leave a payroll trail. Expect recorded absence to rise again next year partly because it is being counted better, not only because more of it is happening.

What should you actually do with a benchmark?

Less than most people assume. A national average tells you nothing about whether one long-term absence or twenty scattered single days sits behind your number - and those need completely different responses. Read the rate alongside a frequency measure like the Bradford Factor, split short-term from long-term before drawing conclusions, and compare against your own trend last year, which is usually the more honest comparison than any national figure.

And know what absence is costing while you are at it: the absence cost calculator puts a pound figure on it.

When are these figures updated?

Both are annual. ONS publishes its sickness absence article in spring (the 2025 data landed 1 May 2026); CIPD's Health and Wellbeing at Work report has typically appeared in September. We re-check this page against both primary releases each year - if a figure here disagrees with the source, trust the source and tell us.