30-second briefing
The central finding
- Treat unpaid care as an agreed relationship, not a reserve of free capacity. The census comparison makes a narrower point: a smaller overall proportion can coexist with more intensive caring.
- Published evidence and Care Circle interpretation are distinguished below.
- No interviews or unpublished service records underpin this edition.
A single headcount is an inadequate description of unpaid care. The same total can contain people offering occasional help and people whose week is organised around substantial caring commitments. For a provider, that distinction changes the meaning of family support: someone being present does not establish that they have spare capacity.
ONS published its Census 2021 unpaid-care bulletin in January 2023. We revisit that historical evidence in October 2026 because it illustrates a continuing analytical problem. These are observations from 2011 and 2021, not a newly measured picture of carers today.
| Measure | Value |
|---|---|
| 2011 | 4.2% |
| 2021 | 4.7% |
Care Circle calculation: sum ONS 20–49 and 50+ hour categories. Difference 0.5 percentage points; approximately 11.9% relative increase using rounded inputs. Historical observations, not current estimates.
What the headline measures
ONS reports that the age-standardised proportion providing any unpaid care in England and Wales fell from 11.4% in 2011 to 9.0% in 2021. For care of 20–49 hours a week, the proportions moved from 1.5% to 1.9%; for 50 or more hours, from 2.7% to 2.8%. These measures concern usual residents aged five and over.
Age standardisation allows comparison after accounting for different population age structures. It is not the raw share of respondents. We therefore keep the proportions in their published form rather than multiply them by a population total to manufacture a count. Nor do we treat caring hours as equivalent to employment hours: the census records a reported weekly category.
Our calculation, reproduced step by step
Combining the two published categories for 20 or more hours gives 1.5 + 2.7 = 4.2% in 2011 and 1.9 + 2.8 = 4.7% in 2021. The difference is 0.5 percentage points. Relative to the 2011 value, 0.5 ÷ 4.2 × 100 = approximately 11.9%. This is Care Circle’s arithmetic using rounded ONS figures, not a new ONS estimate.
Over the same interval, the any-care measure fell by 2.4 percentage points: 9.0 − 11.4. Those two movements can coexist. They demonstrate why a declining overall proportion should not be presented as evidence that intensive caring has also declined. Rounding limits the precision of the combined estimate; the apparent relative increase should not be reported with additional decimal places.
What the numbers cannot tell us
The census was taken during the pandemic, and ONS identifies changes in question wording and the circumstances of caring as comparability considerations. It does not establish why an individual provided more hours. It cannot show that a particular commissioning policy caused the national change, or that a family in a provider’s caseload has become more exhausted.
This matters for editorial accountability. A historical national pattern can justify a better question, but it cannot replace a current local answer. We have not used the comparison to forecast a 2026 total or to estimate the monetary value of unpaid care. Those would require separate evidence and clearly stated assumptions.
A care plan should not silently allocate the remainder to family
Consider an illustrative home-care arrangement in which scheduled visits cover mornings and evenings. An entry saying family will help between visits can conceal several separate expectations: preparing food, responding to calls, supporting movement, arranging appointments and dealing with an unexpected absence. This is an example of a planning risk, not a case investigated by Care Circle.
A service can make those expectations visible by agreeing what the person wants, what each carer is willing and able to do, and how changes will be communicated. The practical test is whether support remains workable when the carer is ill, at work, asleep or away. The family relationship should not become an unrecorded contingency contract.
Separate willingness from availability
A person may willingly support a relative and still be unable to cover a specific task or time. Recording willingness without the practical conditions can create a fragile arrangement. A useful service conversation distinguishes the tasks agreed, the usual availability and the circumstances in which support would need to change. It should also avoid presenting a carer’s preference as the preference of the person receiving care.
For editorial purposes, a documented pathway should make the same distinctions. An account saying a family declined involvement can sound very different when the reason was an incompatible work schedule or an inaccessible process. Context does not remove the organisation’s responsibilities, but it can correct a misleading interpretation. Where that explanation cannot be verified, report the uncertainty. A family’s absence from a meeting should not be treated as evidence of indifference, and a relative attending every meeting should not be assumed to have unlimited capacity.
The next evidence should come from the local pathway
A useful follow-up would examine how services identify a change in a carer’s circumstances and respond to it. Relevant measures include the interval between a reported change and a plan review, whether a contingency was agreed, and whether the carer says the revised arrangement helped. Published totals alone will not supply those answers.
Any collection of family experience needs voluntary participation, careful handling of sensitive information and enough context to avoid blaming people for withdrawing support. An anonymised example should still explain the service setting and sequence of decisions. A compelling quotation without that context can obscure the very coordination problem an article is meant to reveal.
Questions leaders should ask now
The Care Circle view
Care Circle assessment
Treat unpaid care as an agreed relationship, not a reserve of free capacity. The census comparison makes a narrower point: a smaller overall proportion can coexist with more intensive caring.
For readers, the useful next step is to examine the assumptions in local support arrangements. For this publication, it is to follow those arrangements through a documented service experience rather than repeat the national headline.
Develop the analysis
Read the connected flagship reports.
Workforce & delivery: turning sector improvement into dependable care
October cost controls: turn funding, learning and energy changes into a usable plan
Operational assurance: suppliers, equipment and resident voice
Sources, method & limitations
How to read this analysis
A source-led examination published on 9 October 2026. The linked documents are the evidence; the operational interpretation and proposed questions are Care Circle analysis. This is not original field reporting.
- Historical census observations; no claim about the October 2026 number of carers.
- Question changes, pandemic circumstances and rounded proportions limit interpretation.
- Operational examples are illustrative, not observed cases.