Managing Sickness Absence: Are You Measuring the Right Things?
UK sickness absence is at its highest level in 15 years. According to the CIPD and Simplyhealth's 2025 Health and Wellbeing at Work report, employees took an average of 9.4 sick days over the past 12 months, up from 7.8 days in 2023 and 5.8 days before the pandemic.
Look at the official figures from the Office for National Statistics, though, and the picture barely moves. ONS data for 2025 puts the sickness absence rate at 2.0%, or 4.4 days lost per worker, statistically flat on 2024 and only 0.1 percentage points above 2019.
Two credible, well-resourced sources. Two very different stories about the same workforce.
That gap is the real story of this article. "How much absence do we have" depends entirely on what you count, who you ask, and when you look, and most organisations have never stopped to check whether their own answer to that question is actually telling them anything useful.
Why absence is a lagging indicator, and why that matters
Sickness absence is a lagging indicator. It tells you what has already happened, not what is about to happen. By the time someone is off sick, work has already been missed, cover has already been arranged, and whatever built up to that point, workload, unresolved stress, a manager who didn't spot the signs, has already done its damage.
That doesn't make absence data useless. It's still one of the clearest, most quantifiable signals an organisation has about the state of its workforce. But treating it as the whole picture, rather than one output among several, is where most measurement goes wrong.
What absence data can tell you
Done well, absence data is genuinely valuable. It can show you:
Which teams, sites or roles carry a disproportionate share of absence, and whether that's shifting over time
The balance between short-term and long-term absence, which usually point to different root causes
Seasonal or cyclical patterns worth planning around
Whether policy changes (a new return-to-work process, a change to sick pay) are having any measurable effect
Used this way, absence data is a useful diagnostic. It tells you where to look.
What it cannot tell you on its own
What absence data can't do is tell you why. A rising absence rate in one team could mean burnout, a bullying manager, a genuine flu outbreak, an ageing workforce with more chronic conditions, or simply better recording since a new HR system went in. The number looks identical in every case. The response should not be.
This is exactly where the CIPD and ONS figures diverge. The CIPD's number comes from employer perception, HR professionals reporting what they believe absence looks like across their organisation. The ONS number comes from the Labour Force Survey, a much larger, self-reported dataset that measures actual hours not worked due to sickness. Neither is "wrong". They're measuring different things, through different lenses, and a single top-line absence figure, from either source, will always flatten that nuance away.
The hidden drivers behind rising absence
The causes behind both sets of figures are consistent, even if the totals differ. CIPD's 2025 report found mental health conditions such as depression and anxiety account for 41% of long-term absence, with musculoskeletal problems close behind at 31%. For short-term absence, minor illness still dominates, but stress and childcare pressures each show up in around a quarter of cases. ONS figures similarly point to minor illness, musculoskeletal conditions and mental health as the leading causes nationally.
Underneath those categories sit the conditions that actually drive them: sustained workload without recovery, low psychological safety, weak line management, and change fatigue from one restructure too many. These rarely show up as their own line in a HR report. They show up later, as a sick day, a resignation, or a disengagement score, once the cost has already landed.
The business cost of getting it wrong
Absence itself is expensive. The ONS puts total UK working days lost to sickness at 148.8 million in 2025. But the direct cost of the missed day is often the smallest part of the bill. There's the cost of cover, the cost of the work that doesn't get done, the strain on colleagues picking up the slack, and, if the underlying driver goes unaddressed, the risk that the next outcome isn't a sick day but a resignation.
Organisations that only measure absence after it happens are, by definition, always managing the fallout rather than the conditions that caused it.
Why rear-view-mirror data leads to reactive decision-making
When absence is the first and only signal an organisation tracks, the response is structurally reactive. A team's absence rate climbs, someone notices, an intervention gets scoped, usually weeks or months after the pressure that caused it first appeared. By then the cheapest, easiest point to have acted has already passed.
This is the core argument for treating absence as one lagging indicator among several, not the dashboard. Upstream conditions, workload, belonging, recovery, manager capability, tend to deteriorate quietly before they ever show up in an absence figure. An organisation that's only watching the lagging number has no way of seeing that decline until it's already cost something.
What better early-warning measurement looks like
Moving from purely lagging to genuinely predictive measurement doesn't mean abandoning absence data. It means adding leading indicators alongside it: workload manageability, psychological safety, recovery and rest, and manager capability, tracked regularly enough to catch a shift before it becomes an absence statistic.
In practice, that means:
Identifying the risk zones. Not every team needs the same level of scrutiny. Focus where retention, capacity or performance is most exposed.
Choosing a small number of leading indicators that are genuinely upstream of absence, not just another lagging metric with a different name.
Making sure a signal actually triggers something. A leading indicator that nobody acts on is just another number. It needs to connect to a real response, manager support, workload redesign, a conversation, not sit in a report.
Checking that it's working. Track whether acting on early signals is actually reducing downstream absence and attrition, not just assume it will.
What does good sickness absence management actually look like?
Good sickness absence management treats the absence figure as a symptom worth investigating, not a target to manage down in isolation. In practice, that looks like:
Absence data that's segmented by team, role and length, not just reported as one company-wide average
Return-to-work conversations that ask what's actually going on, not just when someone's fit to return
Line managers equipped and trusted to spot pressure building before it becomes a sick day
Leading indicators sitting alongside absence data, not replacing it
A genuine willingness to look at what the data can't explain, rather than stopping at the number
How Kamwell approaches absence measurement
We start from the same premise this article has been making: a single absence number, however accurately recorded, is not the same as understanding what's driving it. Our work is about separating what's happening from why it's happening, root cause over symptom, and it's grounded in recognised research, including the Oxford Wellbeing Research Group's Workplace Wellbeing Playbook.
We also don't believe averages tell the whole story. Workforce experience differs meaningfully by role, level, geography and more, and an organisation-wide absence rate can hide real inequalities between groups that a single dashboard number will never surface. That's why segmentation, understanding where pressure and risk actually sit, matters as much as the headline figure itself.
None of this replaces expert judgement with a formula. It's methodology, context and interpretation working alongside the data, not instead of it, so leaders aren't left alone with a spreadsheet and no sense of what to do next.
FAQ: your sickness absence questions answered
What is the average sickness absence rate in the UK? It depends which source you use. The ONS puts it at 4.4 days lost per worker in 2025 (a 2.0% absence rate), based on the Labour Force Survey. The CIPD and Simplyhealth's 2025 employer survey puts perceived absence considerably higher, at 9.4 days per employee, the highest figure in 15 years of their reporting.
Why do sickness absence figures vary so much between sources? Different methodologies measure different things. Official statistics like the ONS figures capture actual hours not worked. Employer surveys like CIPD's capture what HR professionals believe is happening across their organisation, which can be influenced by recall, self-report and which absences get formally recorded.
What are the main causes of sickness absence? Mental health conditions and musculoskeletal problems are the leading causes of long-term absence. Minor illness remains the most common cause of short-term absence, with stress and caring responsibilities also significant factors.
How can organisations reduce sickness absence? By treating absence as a lagging indicator rather than the whole picture: segmenting data properly, understanding root causes rather than just counting days, and adding leading indicators, workload, psychological safety, manager capability, that can flag risk before it shows up as absence.
Is a rising absence rate always a bad sign? Not necessarily on its own. It could reflect genuine deterioration, but it can also reflect better recording, a demographic shift, or a change in policy. That ambiguity is exactly why absence figures need context, not just tracking.
From counting days lost to understanding risk before people fall out
The CIPD and ONS figures won't agree any time soon, and that's fine. The real takeaway is that any organisation relying on a single absence figure, from any source, to tell them how their workforce is doing has already limited what they can see. Absence data can tell you where to look. It can't tell you why, and it can't tell you what's coming next.
Getting ahead of that means treating the sick day as the last signal in a chain, not the first one worth measuring, and building the kind of visibility that catches pressure while it's still just pressure, before it becomes a statistic.