Audit Guidance
How to run the Hidden Work Audit in an adult social care team · AmiHC
The audit measures how much registered practitioner time goes on work that does not need a registered practitioner. It takes about twenty minutes per person, and the whole thing can be done inside one team meeting.
This guidance covers which method to use, how to score, what usually goes wrong, and how to present the result so it survives contact with a finance business partner.
Before you start
- Send the team briefing round first. People answer very differently depending on whether they think this is about capacity or about performance. The briefing exists to settle that before anyone opens the tool.
- Decide who is in scope. Everyone in the team, not just the qualified social workers. Occupational therapists, wellbeing assessors and coordinators, community care officers, reviewing and assessment officers, support workers, duty officers and team managers all pick this work up.
- Agree who will hold the result. One named person, usually the team manager or a service lead. The tool saves to whichever browser it is opened in, so there is no shared copy unless you make one.
- Do not wait for salary figures. The tool works without them and its defaults are national averages, so you can run the whole session and add money later. If you do want the cost figures exact, ask finance or HR for the midpoint salary for the practitioner grade and the support worker grade, matching the contracted hours you enter, plus your council's on-costs multiplier.
Choose your method
There are three ways to run it. Pick one before the meeting, because they are not interchangeable.
| Method | How it works | Use it when | Watch out for |
|---|---|---|---|
| Team consensus | Fill it in once, together, on a screen. Agree each answer as a group for a typical practitioner. | Fastest. Best for a first pass, or a team of fewer than about eight where people do broadly similar work. | The loudest voice sets the answer. Ask the quietest person first on each of the big tasks. |
| Individual then pooled | Everyone fills in their own copy on paper using the capture sheet, then one person enters the median answer for each task. | Most robust. Best for teams of eight or more, or where roles differ a lot. | Takes two sittings. Use the median, not the mean, or one outlier will distort everything. |
| By sub-team | Run it separately for each sub-team and compare. | Best when you suspect the problem is structural rather than individual, for example one team has a support worker and another does not. | Only worth it if you can act at sub-team level. Otherwise it produces an interesting chart and no decision. |
If in doubt, use individual then pooled. It takes longer, it produces a number you can defend, and it surfaces the distribution. In the original audit the median practitioner was carrying about 24 hours a month and the highest was carrying 405. A consensus session would never have found that.
Scoring the tasks
The three questions
For every task that is happening, three answers. How often a typical practitioner does it, how long it takes each time, and where it came from.
| How often | How long each time |
|---|---|
| Never | A few minutes |
| Once or twice per month | 15 to 30 minutes |
| About once per week | 30 minutes to an hour |
| Two to three times per week | 2 to 3 hours |
| Daily | About half a day |
| More than once per day | A day or more |
The bands come from the original survey and are deliberately wide. Note there is no band between one hour and two hours, so a ninety-minute task has to be rounded one way or the other. People tend to round up, so if a task genuinely sits in that gap, round down and note it.
What counts
- Door to door. Travel, waiting and writing it up all count. If you drove someone to the dentist, the whole afternoon counts, not the ten minutes in the chair.
- Failed attempts count. Forty minutes on hold that achieved nothing still consumed forty minutes.
- Chasing counts. The three phone calls to find out whether the bulky waste collection happened are part of the task.
- Do not double count. If a single visit covered a joint visit and a prescription pick-up, put it under one task, not both.
Where the work comes from, which is the question that matters
Every task that is happening also asks where it came from: a hospital or secondary care, another NHS service, housing, benefits, a utility or bank, a pharmacy or GP practice, your own systems, or no service exists for it.
This is the only question in the audit with somebody else's name on the answer. How many hours your team spends on this work tells your team something it already knows. Where those hours came from tells the sending organisation something it has no other way of finding out.
It also separates three problems that look identical on a timesheet and need completely different answers.
| What the origin says | What kind of problem it is | What to do about it |
|---|---|---|
| A named organisation | Work transferred from somewhere that stopped doing it | Take the number to them. This is the only category where anybody could be asked to pay. |
| Our own systems | Failure demand, made by a process rather than a person | Fix the process. Asking for money for this would pay to keep it. It is also the part you can act on without anybody's permission. |
| No service exists | A need with nobody attached to it | Nobody dropped this, nobody was ever asked. It needs a decision about who owns it, and it is the hardest of the three to shift. |
Answer from what you saw, not from what should have happened. If a hospital asked you to chase a result, the origin is the hospital, whether or not they were right to ask. You are not being asked to judge another organisation's internal process, and you could not reliably do so even if you were. Where you genuinely do not know, leave it as not sure. An honest gap is worth more than a guess, and the tool will tell you how much of your total is unattributed.
Who should be doing it
Each task also asks who should be doing it: a registered practitioner, a team support worker, or another agency or service.
Every task starts assigned to a registered practitioner. That is deliberate. The tool does not pre-judge which work is beneath a qualified role, because that judgement is not the tool's to make and a tool that made it would deserve the objection it would get.
Expect disagreement here, and let it run. "Supporting someone at a medical appointment where they would struggle and there is nobody else" is either an errand or the relationship that makes the next difficult conversation possible, and reasonable practitioners will split on it.
Where the team cannot agree, leave it with the practitioner. An audit that under-claims is far more useful than one that over-claims, because the first survives scrutiny and the second does not.
The money tasks
Ten tasks are flagged in the tool because they involve a person's money, bank access, benefits or bills. Score them the same way as everything else, but treat the result separately.
Those tasks carry financial abuse and governance risk regardless of who does them. They need a written procedure, clear recording and a second pair of eyes. That argument is independent of capacity, it usually moves faster than the capacity argument, and it is worth taking to the safeguarding lead on its own.
Reading the results
| Figure | What it means, and what it does not |
|---|---|
| Hours originating elsewhere | The figure to take outside your own organisation. It is the only one in the tool that names somebody else. Split by source so you can see which single relationship is worth fixing first. |
| Hours a month per practitioner | The headline. Compare it with the audit median of 23.6 hours. Above 111 puts you in the top quarter of the original audit. |
| Hours a year across the team | The figure to put in a paper. It is the one people remember. |
| Full-time posts that time represents | The amount of practice time being absorbed, expressed as posts. Not a headcount reduction. The work still has to be done. |
| Support worker posts those hours would fill | What it would take to absorb the work you have marked as moveable, with its cost shown alongside. It is a staffing requirement, not a saving. |
| Value at practitioner cost | What the freed practitioner time is worth. It is not cash unless a post is genuinely released or a vacancy is left unfilled. Say so out loud when you present it, before someone else says it for you. |
| The difference | Value of the time, less the cost of the posts needed to absorb it. Both use the same denominator so they reconcile deliberately, which is the point a finance reader will test first. |
What usually goes wrong
- Under-reporting. By a distance the most common error. People absorbed these tasks to get something done for someone, so it stopped feeling like work. If your total comes out well below 24 hours a month, ask the team directly whether that is real or whether it has just stopped registering.
- Answering for the team rather than yourself. Produces a flat, average-looking result and hides the distribution.
- Scoring a bad month. A crisis fortnight is not a typical month. Ask people to picture an ordinary one.
- Marking everything as moveable. If the result says 95% of it could go elsewhere, the team has not engaged with the Who should do this column. Go back to it.
- Leading with the money. The cost figure is the weakest argument in the pack, because it is a value-of-time figure and any competent finance officer will say so within a minute. Lead with the posts-worth-of-practice figure and the governance risk.
- Presenting it without the team. If practitioners have not seen the result before the head of service does, expect it to be read as management doing something to them.
Presenting it
- Team first. Share the result with the people who filled it in, and take their corrections.
- Print the whole page to PDF. The print view opens the method section automatically, so the caveats travel with the numbers. A figure circulated without its method will come back at you.
- Split the ask in two. The governance ask goes to the safeguarding lead. The capacity ask goes to the head of service. They move at different speeds and joining them slows both down.
- Take your biggest single source to that organisation. Not a complaint, just the number and a question about what changed at their end. One relationship at a time, starting with the largest.
- Take the top three tasks separately. For each one, find out whether an existing service, a voluntary organisation or another team already does that job. Often one does, and nobody knew. Those are free wins that need a conversation rather than a post.
- Ask for a decision and a date. Not for agreement in principle.
Capture sheet A · Task scoring
One sheet per person. Tick one box in each half. Bring it to the pooling session.
| Task | How often | How long each time | Who should do it | Where from |
|---|---|---|---|---|
| Money and benefits | ||||
| Bank visits | ||||
| Setting up bank accounts | ||||
| Benefit checks | ||||
| Benefits and debt | ||||
| Bills and utility calls | ||||
| Financial assessments, gathering post | ||||
| Grants | ||||
| Collecting and delivering money | ||||
| Utility debts and repayment plans | ||||
| Gas and electricity top-ups | ||||
| Practical and household | ||||
| Bulky waste | ||||
| Deep cleans and hoarding support | ||||
| Emptying properties, inventories | ||||
| Empty property checks | ||||
| Furniture and clothing | ||||
| Housing applications and forms | ||||
| Helping people move house | ||||
| Post redirection and collection | ||||
| Repairs, reporting and waiting in | ||||
| Shopping calls | ||||
| Food parcels and vouchers | ||||
| Feeding pets | ||||
| Health and medical | ||||
| Blister packs | ||||
| Prescriptions | ||||
| Urine samples | ||||
| Health equipment runs | ||||
| Medical appointments | ||||
| Booking patient transport | ||||
| Dental | ||||
| Opticians | ||||
| Placement and transition | ||||
| Day centre and social group visits | ||||
| Extra care viewings and moves | ||||
| Care home to care home moves | ||||
| Funerals | ||||
| Other | ||||
| Welfare checks | ||||
| Joint visits needing two staff |
Anything happening in this team that is not on the list: ______________________________________________________
Capture sheet B · Team and cost figures
Get these from HR and finance before the session. Guessing them is the fastest way to lose the argument.
| Figure | Value | Source and date |
|---|---|---|
| Team name | ||
| Registered practitioners in the team (headcount) | ||
| Contracted hours per week | ||
| Annual leave plus bank holidays, in days | ||
| Practitioner salary for those hours | ||
| Support worker salary for those hours | ||
| On-costs multiplier |
Capture sheet C · Follow-up actions
Fill this in during the session while people are still in the room.
| Task | Who else could do it | Who will find out | By when |
|---|---|---|---|
Where the benchmark comes from
The 36 tasks and the comparison figures come from a survey AmiHC designed and ran with 23 adult social care practitioners at a city council in August 2024. Respondents covered social workers, senior social workers, occupational therapy, wellbeing assessors, team managers and support workers, across mental health, safeguarding, learning disability, front door, review and locality teams.
- 17 of the 23 reported doing at least one of these tasks. Six reported none.
- Among those 17, the median was 23.6 hours a month, the top quarter reported 111 hours or more, and the highest single response was 405 hours.
- Those three figures count only tasks where the respondent gave both a frequency and a duration, so they understate rather than overstate.
- The tasks carrying the most hours in total were medical appointments, joint visits, health-related transport and patient transport.
- The tasks most often reported at all, on a wider count that includes those who gave no duration, were joint visits (65% of all 23 respondents), then blister packs, food parcels and repairs (57% each), and medical appointments (52%).
Every task on the list is something a practitioner reported actually doing. Three questions were rewritten because the original survey form had corrupted or truncated them, and the wording throughout has been tidied for consistency.
In that council, the audit contributed to the creation of six full-time equivalent team support worker posts within an adult social care restructure, funded from within the existing establishment.
Indicative only. This is a structured estimate, not a time-and-motion study, and it is not employment, HR or financial advice. Creating or regrading posts involves job evaluation, staff consultation and trade union engagement. Use the audit to open the conversation, not to close it.