Attendance and Inclusion Services and Interventions

Setup

1 Introduction

A programme of analytical work was began by the SCC BI team in September 2023, with the aims of understanding the key drivers of school absences in Sheffield and the reach and effectiveness of existing services and interventions. The first of those two requirements is written up the report Attendance in Sheffield Schools. This report addresses the second requirement.

A note on terminology

The terms services and interventions are generally used interchangeably. Some of the things we’re evaluating here might be better classed as events. Even so, in each case we’re treating them the same, using the data to evaluate the impact of on attendance.

1.1 Data sources & processing

This is only a brief overview of data sources & processing used in this report. Please contact the BI team giles.robinson@sheffield.gov.uk if you require further detail.

• Attendance data is held in the PAS Oscar database tables
• Involvements are held in PAS Oscar database
• Child social care episodes are retrieved from Oscar database but are recording on LiquidLogic children’s system.
• Other services and interventions which are not stored as involvements (school attendance orders, parenting programmes, managed moves) are also retrieved from the Oscar schemas.
• Economic deprivation information is retrieved from local indicators data
• Demographics are from the Oscar database
• Special Educational Needs data is from the school census
• School information is from the Oscar database

A separate data model R script retrieves the data listed above and transforms it into clean processed data for analysis in this report script. Crucially for this analysis, attendance data is categorised by DoE attendance codes and calculated as % of available sessions attended and these metrics, and mapped onto the involvements data according to involvement start dates. Average ‘before and after’ attendance metrics are then calculated. This creates involvement summary which can be generated for any demographic group for which we have data.

1.2 Analytical approach

Measures of effectiveness & finding suitable comparators

In this report our aim is to evaluate how effective different things are at increasing school attendance and reducing exclusion rates - though the primary focus is attendance. There are a few difficulties to this:

• Services are allocated according to need, and not on a random basis, so we often do not have a “control group”
• Attendance tends to reduce over time as children progress through secondary school - so our background rate is (in older children at least) always reducing, and we shouldn’t necessarily be surprised to see that our interventions are associated with a reduction in attendance, even if the stated aim is to increase attendance

To tackle this we’ll take a few different approaches, explained here with example plots:

Average attendance before & after

The ideal approach is to compare a baseline period (prior to intervention) to a test period (following intervention) for the same children. We can then see differences in overall attendance and coded absence reasons, and we can see how these differ for different groups.

This shows the overall trajectory of changing attendance, and how this is affected by our involvements - though it’s important to remember that the points on these charts are just the overall average, behind which lie a lot of variation.

For example, this is the plot for Attendance Advice. We see attendance declining prior to involvement and improving afterward.

Important

The pattern seen here is repeated for many of the interventions covered in this report. The overall trend here changes direction, but attendance levels do not return back to what they were prior to involvement. Moreover, if we add up all the attendance on the left hand side of the chart above and compare it to that on the right, we see a net negative change. Attendance advice is associated with a decrease in attendance levels.

Categorised attendance before & after

This view looks at the same indexed time periods as the average attendance before & after plot shown above, but here children are categorised according to attendance brackets. For each time period (again, relative to the involvement start or closure date) the % of children in each attendance category is calculated: 0 - 50% 50 - 80% 80 - 90% 90 - 100% Also included here are several important other categories: “pre-school” - children who have not yet reached school age at the time period in question “leaving age” - those who have completed Y11 and left school at the time period in question “COVID” - half term periods that fell during COVID lockdowns, the attendance data for which is incomplete or unreliable “out of scope” - time periods that have not elapsed yet or for which data is not yet available. These are categorised but excluded from the % calculations used in the plots.

This can reveal patterns of change in the distribution of attendance that lies behind the averages. Here’s an example categorised stacked bar plot, in this case for Attendance Advice

Coded absence reasons

Comparing levels of difference coded absence reasons before & after involvement may reveal changes in patterns of behaviour. We can do this at the level of the year prior and year post involvement, or at the level of the half term prior and post involvements.When looking at the year prior & post we restrict the data to only children who have at least 3 half terms’ worth of data available in each of those one year periods.

As an example, here is the coded absence before & after plot for the ‘Current EHCP’ involvement:

With & without analysis

Comparing the attendance levels (and coded absence reasons) of groups who had the intervention with those who did not. This is more problematic because finding suitable comparator groups is difficult, but useful for interventions where no prior period data is available - such as the readiness for school team, or 0-5 SEND team.

Here’s an example ‘with & without’ plot, in this case plotting attendance from year 1 through to year 6 for those with and without the Inclusion & Attendance Y4 team involvement:

1.3 Random sampling

To contextualise the above, we can compare the outputs of analysis as described above with a similar analysis for a randomly selected children with no involvement. Before we get into the analysis of the actual involvements, it’s worth looking at a couple of these.

Firstly, a random sample average trajectory plot. Here we’ve taken 200 random children who consistently hit below 75% (for 3 or more half terms in a row), since 2021, and who did not have any involvements.

These random samples often show better apparent recovery than many of the involvements. The best explanation for this is regression to the mean, the idea that when conditions are in a more unfavourable state, they will often revert back to a less unfavourable state, even in the absence of any intervention. This can lead, for example, to people concluding that the medicine that they began taking when they were feeling particularly ill was the cause of their recovery, when in fact they would have recovered naturally without the medicine.

Next we’ll look at the categorised attendance of a random sample, before & after a random date. When evaluating involvments we will typically remove the pre-school, COVID, leaving age and out of scope categores, but they’re included here. The profile of these plots is usually very different for children with involvements.

Finally we’ll look at an example codified reasons of a random sample. These four plots each show the year on year change for a random sample of 200 children, based on a randomly chosen date since 2021.

The takeaway from these 4 charts is that year on year increase in ‘no reason’ absences of about 1% is to be expected in many cases, as are small variations in the other coded reasons. Large chages beyond this are likely behaviour of the cohort receiving the involvment and/or a result of the involvement itself.

2 Services & Interventions

The rest of this report details the findings on effectiveness, broadly following the methodology given above. We cover the following services, interventions & events, and aside from the child social care episodes (CIN/CPP/CLA) which are taken together, we’ll cover in order of decreasing volumes of activity (based on a count of children receiving each during 2024).

Involvement	count
CIN	2971
MAST	2900
Ed. Psych.	922
Attendance Advice	843
Reduced timetable	788
Current EHCP	758
Autism team	648
Inclusion Advice	541
EHE	540
EY Inclusion	537
Exclusion Concern	481
PA Cohort Tracking	406
Consultation	401
Progressions Team	370
CPP	306
penalty notice warning letter	287
SA Cohort Tracking	221
CLA	220
Parenting	218
Other	154
I&A - Y9	112
HI team	110
I&A - Complex SEND	109
Non PIP/SIP - Think for the Future	107
Non PIP/SIP - Theraputic Outreach	101
SIP - Theraputic Outreach	77
Secondary Inclusion Panel	77
I&A - Vulnerable Learner	60

 

Involvement	count
PIP - Theraputic Outreach	59
Primary Inclusion Panel	59
Portage	58
Rowan outreach	53
VI team	53
UCAN	42
I&A - School Readiness	35
I&A - Y4	32
S437- School Att Order EHE	32
PNOR 50+	31
S437 - School Att Order CME	23
School Att Order Breach	23
I&A - Reintegration	13
GP Protocol	10
Nurture - BLT KS1	9
Nurture - BLT KS3	7
Nurture - Bumblebee	6
Nurture - BLT KS2	5
Nurture - Step Out	5
Exclusions - Managed Move Involvement	4
Managed Move	4
SIP Screening	4
I&A - Traveller Education	3
PIP - Think for the Future	3
Nurture - The Hive	2
EHE Meeting	1
Nurture -	1
PIP Screening	1

2.1 Family Intervention Service

Family intervention Service (FIS), formerly known as MAST.

This service is targetted at families rather than children, but in our data model we mapped FIS activity to the records of about 3000 children per year:

Immediately prior to FIS involvement we see a large drop in average attendance - perhaps this reflects some family crisis that prompts the services involvement, and also affects school attendance? The involvement is associated with an immediate increase in attendance, and a change in the direction of travel, with attendance steadily improving term-on-term once the team is involved. Although the overall net year on year change is still negative - attendance does not recover to prior levels.

Exclusion rates remain high for children involved with FIS:

The categorised reasons also show the immediate effect of FIS involvement, with an increase in the highest attendance category, and a drop in the lowest bracket:

Since the significant change following FIS involvement happens at the half term level, we’ll look at coded reasons at the half term level. Here we see a decrease across all coded absences, especially in illness rates.

Finally, we look at how the half-term to half-term shifts in attendance vary by some chosen characteristics:

• Younger children see a greater net improvement than older children; those going into Y7 and Y11 see a net reduction
• girls see a bigger improvement than boys
• Deprivation makes a difference with children in more deprived wards seeing a bigger increase in attendance - but only to a point

2.2 Educational Psychology

This involvement marks the child’s placement on the caseload of an Educational Psychologist. There are around a thousand of these in the city each year:

The majority of these are for children who require SEN support, and those with speech, language and communication needs:

Educational Psychology
counts of children with involvements in 2024, by SEN level
SEN Support	475
No SEN	296
NA	102
EHCP	49

 

Educational Psychology
counts of children with involvements in 2024, by primary specific need
NA	407
Speech, Language And Communication Needs	260
Autistic Spectrum Disorder	89
Social, Emotional and Mental Health	85
Moderate Learning Difficulty	21
Physical Disability	11
Other	10
Hearing Impairment	9
Specific Learning Difficulty	7
No Specialist Assessment	6
Visual Impairment	6
Profound And Multiple Learning Difficulty	4
Severe Learning Difficulty	4
Behaviour, Emotional And Social Difficulty	2
Multi Sensory Impairment	1

The two year run up to the start of an Educational Psychology involvement shows a quite steep decline in average attendance, and the involvement startis associated with a turnaround in the direction of travel.

Exclusion rates show a rapid and sustained drop following Educational Psychology involvement:

The coded reasons in the year before & after educational psychology doesn’t reveal any movement beyond what we see in the random sampling, so is not included here. The categorised attendance analysis also doesn’t reveal any significant impacts of Educational Psychologist involvement.

Finally, we also looked at a ‘with & without’ analysis, tracking children with various SEN levels and needs, and children in deprived areas, but across all of these variables children with the Ed. Psych. invovlement consistently attend lower than those without it - and clearly there are factors at work beyond what we have access to in the data.

2.3 Attendance advice

844 attendance advice involvements were started in 2024:

The age profile shows children of all ages receiving this involvement, but with a peak at ages 13 - 15.

Calculating the average (mean) attendance in the terms prior to, and following the opening of attendance advice involvement, we see the pattern below. Average attendance declines steeply towards the point of intervention. After the involvement starts we see a turnaround, though attendence does not recover back to levels seen two years prior (given the overall 4 year timescale here, this can perhaps be attributed to the tendency of attendance to reduce with age)

The categorised attendance analysis shows more detail than the overall average, as we get steady growth in both those entirely missing ‘NA’ and the highest attendance bracket, while the severe absence category steadily decreases. Here those leaving school or at preschool are removed:

Another approach is to compare the relative effectiveness of attendance advice on groups of children with different characteristics. Attendance advice is associated with a reduction in attendance year on year (and this remains true across all groups) but if we look half-term to half-term there are groups who see a net improvement:

1. Attendance advice makes more of a difference in younger children.

2. There is a big difference in gender - with girls seeing a net increase at the half term level, and boys seeing an overall decrease

3. There is also a trend by deprivation - children living in poorer areas of the city see a net increase whereas children living in more affluent wards see a decrease on average.

It’s worth also looking at the school level - since attendance advice is essentially a notification from the authority to the school, and the school’s response to this may vary. The confidence intervals (shown as error bars here) are very wide for this data, but there are big differences between schools that may be indicative of different responses to the attendance advice intervention:

Finally for Attendance Advice, looking at the coded reasons a full year either side of involvement only shows increases across various absence reasons, but there are small improvements at the half term level, mostly in terms of illness and ‘no reason’ absences:

2.4 Reduced timetable

Reduced Timetable volumes

Reduced timetables are associated with a reduction in all absence codes:

The categorised attendance data reveals large increases in the 0-50% attendance bracked, presumably as a result of the reduced timetable arrangement. There are no lasting improvements to the higher attendance brackets:

2.5 Current EHCP

The ‘Current EHCP’ involvement simply tracks children with an Education, Health & Care Plan. And although not an intervention in & of itself, it is a plan to meet the child’s needs, so will be associated with other changes.

There are around 800 EHC of these starting each year - though changes in demand & the council’s response to that demand, may impact volumes. 2023 seems to have been a peak year.

The current EHCP start date sees a sustained increase in average attendance levels:

The EHCP start date is associated with a reduction in almost all coded absence reasons:

The categorised attendance plot shows a sustained increase in all higher attendendance brackets, and a reduction in the severaly absent bracket:

2.6 Autism Team

We start around 600 of these involvements per year:

The ‘before & after’ plot only shows a small change in attendance levels. But since the team become involved around age 5 most children do not have much of a ‘before’ period to consider.

In the coded year-on-year veiw, the Autism team start date is associated with an increase in almost all types of absence.

The assumption here has to be that these year on year increases reflect the steeper line seen in the ‘with & without’ analysis below, and are explained by the same greater severity of need.

To create a “with vs without the team” analysis, we take all children in the attendance data, with a primary specific need of Autism at some point in the attendance data (so we’re taking a “lifetime diagnosis” approach). This cohort is then categorised into those who had involvement from the team and those who did not. We see a consistently lower attendance for children involved with the team, as well as a consistently steeper drop in attendance.

Once again, we have to assume that this reflects differences in the severity of need - with the team becoming involved in the more severe cases - and that there are factors at work here that are not recorded in the data.

But before we move on it’s worth considering the age on in involvement start date. The autism team generally become involved around age 5, but can be at any age - and the team can become involved before a formal diagnosis is in place:

If we categorise children according to their age on first involvement date, there are some interesting differences in patterns of attendance. Here we’ve created three groups: those aged 5 and under when the team first become involved, those aged 6 to 10, and those aged 11+.

2016 seems to have been a peak year for all age groups (which may be a data quality issue) but 2016 aside, the voumes of all 3 of theses groups have increased recently:

Plotting attendance by national curriculum year for these three groups shows that children who have inolvement with the team earlier in life have better attendance throughout their school career than those who have team involvement starting later. This may result from the actions & support of the team, or perhaps the impact of having a diagnosis at different ages, or it may result from effects upstream of these differences in diagnosis age, none of which we can unpick from the data, but the effect is there:

2.7 Inclusion Advice

Inclusion advice involvements are associated with a reduction in exclusion rates, but exclusion rates persist for these children, remaining far above the average rate (shown here with a dotted line)

Inclusion advice is associated with an increase in “no reason” absences reasons, and absences due to exclusions. This fits with what we see above - exclusion rates are higher year on year for children who receive inclusion advice.

2.8 Elective Home Education (EHE)

Warning

It seems unclear from the data whether the attendance data following an EHE involvement is attendance back at school following a short period of home education, or if it shows attendance in the home education setting. But do home educators submit attendance data?

Because of this uncertainty, the analysis included here is limited, but from the available data it seems a period of EHE results in increased attendance at school.

2.9 Early Years Inclusion

The Early Years inclusion team work almost entirely with children younger than school age, so a ‘before & after’ approach will not work here.

Since the team work with SEN children, it makes sense to compare the attendance levels of children who require SEN support or an EHCP plan, but with and without the EY inclusion team involvement. In both cases, (after year 0 at least) we see better attendance in all years when the team are involved. (note that the ‘with the team’ line here cuts off in year )

2.10 Exclusion Concern

This involvement identifies children with high levels of temporary and permanent exclusions. This is growing over time, presumably as exclusion rates are also growing.

The focus here is exclusions, and looking at the before & after plot for exclusions, we see extremely high exclusion rates prior to involvement, and a significant reduction afterward, but rates continue to be well above average:

Exclusion concern involvements often correllated with other involvement types - inclusion panel, think for the future and reduced timetables.

2.11 PA Cohort Tracking

This involvement simply means the child is persistently absent, and is being tracked as such. (see also SA cohort tracking, below)

The PA cohort tracking involvement shows no particular improvement in attendance levels:

2.12 Consultation

N.B. At the time of writing this involvement is not understood - we had feedback that it was not in use in late 2023, but there are volumes of activity in 2024 and 2025.

The “Consultation” involvement before & after plot shows very low attendance prior to involvement, and the trend continuing - though we have to consider that activity prior to 2024 is very limited, and so there simply isn’t enough elapsed time to give data here.

2.13 Progressions Team

The progressions team are involved in matching children to alternative provision.

Progressions Team
counts of children with involvements in 2024, by SEN level
No SEN	653
SEN Support	354
EHCP	40
NA	25

There is no improvement to attendance levels associated with progressions team involvement:

The progressions team involvement is associated with an increase in ‘no reason’ absences and absences due to exclusion:

2.14 Penalty notice warning letter

Penalty notice warning letters don’t appear to have any effect on overall attendance (this is in contrast to the Section 437 notices covered later on):

Penalty notice warning letters are associated with year on year increases in most coded absence reasons, particularly ‘no reason’ absences:

2.15 SA Cohort Tracking

This involvement simply means the child is severely absent, and is being tracked as such. (see also PA cohort tracking, above)

Children with the SA cohort tracking involvement show a general improvement in attendance levels:

2.16 Parenting

The parenting data comes via the Early Help data model, and is mapped onto attendance data like the other involvement types. These are children whose parents are enrolled on parenting programmes. :::{.callout-warning} N.B. the volumes here drop off in 2022, and work is needed to check we’re pulling in all available data. :::

Parenting programme don’t appear to have any effect on overall attendance, indeed average levels continue to decline after the involvement start:

Looking at the coded reasons, there are no shifts here that are not comparable with those seen in the random samples - parenting programmes seem to have no significant effects:

2.17 Other

N.B. At the time of writing, I don’t know what this actually means. There is an involvement cohort name of simply “other”, with over 150 involvements starting in 2024:

“Other” type involvements see a substantial change in average attendance levels:

Looking at the coded reasons, there are no significant year on year shifts associated with the ‘Other’ type involvements.

2.18 Hearing Impairment Team

Involvement with the Hearing Impairment team has no apparent effect on attendance - which for this cohort is healthly throughout.

The analysis of coded reasons shows no changes that are not comparable with the background changes seen in the random samples, and so is not included here.

Taking children with a primary specific need of Hearing Impairment, we can do a ‘with & without team involvement’ analysis. This shows consistently lower attendance through all school years, for children with the team involvement. Presumably this reflects differences in severity of need within the cohort of hearing impaired children.

2.19 Inclusion & Attendance Y4 / Inclusion & Attendance Y9

2.19.1 I&A Y4

The Inclusion & Attendance Y4 team get involved from Y4 onwards in order to assist with the transition to secondary school in Y6-7 which is generally associated with a significant drop in attendance. Activity is fairly low in recent years, with only about 40 involvements starting each year.

The team work more with children in more deprived areas, and those who require SEN support.

Inclusion & Attendance Y4 team
counts of children with involvements in 2024, by SEN level
No SEN	304
SEN Support	243
EHCP	15

 

Inclusion & Attendance Y4 team - counts by IMD quartile
4 = most deprived wards; 1 = least deprived
1	34
2	105
3	157
4	264

And the children who the I&A Y4 team work with tend to have lower attendance throughout primary school

Since the team’s aim is to assist with the transition to secondary school (Y6 to Y7), it makes sense to compare this transition point for children who have involvement with those who do not. We’ll do so across a some other factors, like SEN levels, deprivation and prior attendance levels.

The charts below show this analysis from years 4 to 8 for all pupils, those with SEN support, EHCP, those in the most deprived wards of the city, and those who were severely absent in Y4 or in Y6:

The above plots show that attendance throughout years 4 to 7 is consistently lower for all children the team work with. And looking at the transition to secondary school for all pupils on average, for children on SEN support and for children in the most deprived wards, the drop from Y6 to 7 is more severe among children the team work with. This presumably represents the targetting of this service to children with the highest needs, according to factors not visible in the data.

For children with an EHC plan, and children who were severely absent in Y4 or Y7, when the team are involved attendance actually improves into Y7. The plots below show the net difference in overall % attendance between Y6 and y7 for the same groups given above, again split by whether the I&A Y4 team were involved:

We should also look at exclusion rates for the Y4 team

2.19.2 I&A - Y9

Overall, children involved with the team show lower attendance through secondary school than those without:

We’ll take the same approach for the Y9 team as we did for Y4 - comparing attendance with & without the team across a number of characteristic groups. This shows a similar pattern to what we saw with the I&A Y4 team - at the Y9 to Y10 boundary, children involved with the team show a more severe drop in attendance than those without, across all the groups we’ve looked at here. The unfortunate conclusion is that we can’t say much about the team’s effectiveness from our data, and we have to assume that there are factors at work that are not present in the data.

Exclusion rates beofre & after the I&A Y9 team involvement

2.20 I&A - Complex SEND

Note

Although the team is called ‘Complex SEND’ almost half of the children involved with the team do not show up as having any special educational needs, according to the school census data.

Inclusion & Attendance Complex SEND
counts of children with involvements in 2024, by SEN level
No SEN	217
SEN Support	211
EHCP	63

 

Inclusion & Attendance Complex SEND
counts of children with involvements in 2024, by SEN level
NA	242
Speech, Language And Communication Needs	80
Autistic Spectrum Disorder	56
Social, Emotional and Mental Health	43
Moderate Learning Difficulty	16
Behaviour, Emotional And Social Difficulty	15
Specific Learning Difficulty	12
Other	9
Profound And Multiple Learning Difficulty	6
Hearing Impairment	4
Severe Learning Difficulty	4
Physical Disability	2
No Specialist Assessment	1
Visual Impairment	1

The I&A complex SEND team involvement shows a sustained improvement in attendance levels:

Although the direction of travel changes for average attendance, when we look at coded absences for the year either side of the involvement start, overall absences increase. Though it’s worth considering that children with complex SEND show a general decline in average attendance beyond the decline for other children. There are also small improvements in lateness, illness & exclusion levels here:

We’ve also looked at a ‘with & without’ analysis, for both children with SEN support and an EHC plan. In both cases, children involved with the team have consistently lower attendance than those without. This must reflect the ‘complex’ nature of the cohort’s needs.

Exclusions

2.21 Non PIP/SIP - Think for the Future

Think for the future is focussed on behaviour, resilience & inclusion. There are around 100 of these involvements each year, for children of all school ages, but mostly around age 11.

The focus here is on exclusions, but for completion we’ll also show the attendance plot. The children receiving this involvement show very poor attendance, which continues to worsen after the involvement:

This change in average attendance is driven by growth in those severely absent - 0-50% attendance; the orange bars here:

Looking at exclusions, you can see the extremely high exclusion rates immediately prior to involvement here, and although there is some reduction in the average rate, it’s slight and over a long time.

Finally for this involvement, the coded reasons show a significant year on year increase in overall absences, driven mostly by an increase in the ‘no reason’ category:

2.22 Non PIP/SIP - Theraputic Outreach

Tip

The volumes and age profile, categorised attendance profile and before & after plots here all look very similar to the non PIP SIP think for the future involvements given above.

Analysis shows that these are the exact same cohort of pupils! Is this an issue with the data, or to be expected?

There are around 100 involvements starting per year with the age profile peaking at 11:

Attendance levels continue to decline after non PIP/SIP theraputic outreach:

There is also no improvement in rates of exclusion absences for children with this involvement, with rates remaining well above average following involvement start:

2.23 SIP - Theraputic Outreach

SIP theraputic outreach involvements are associated with a continued decline in attendance:

SIP theraputic outreach involvements are associated with a significant improvement in exclusion rates:

2.24 PIP - Theraputic Outreach

After PIP theraputic outreach involvements attendance is very poor - although the picture here is better than at Secondary Inclusion Panel theraputic outreach, and shows a slow improvement:

Though rates do not return to previous levels or the overall average, PIP theraputic outreach involvements are associated with a significant improvement in exclusion rates:

2.25 Secondary Inclusion Panel

Caution

As with Non PIP/SIP Think for the Future and Non PIP/SIP Theraputic Outreach, SIP theraputic outreach and Secondary Inclusion Panel seem to be the exact same cohort of pupils

TO DO - check this in the source data

Secondary Inclusion Panel involvements are associated with a continued decline in attendance:

Secondary Inclusion Panel involvements are associated with a significant improvement in exclusion rates:

2.26 Primary Inclusion Panel

Caution

As with several other pairs of involvements, PIP and PIP theraputic outreach are the exact same cohort

Primary Inclusion Panel involvements are associated with a continued decline in attendance:

Primary Inclusion Panel involvements are associated with a significant improvement in exclusion rates:

2.27 PNOR 50+

Volumes of this involvement are fairly low, and many do not appear to have a Date of Birth recorded, so the age profile could not be calculated.

Volumes of this involvement prior to 2024 are very low, and the involvement generally pertains to older children, and specifically to those not on roll. There is no attendance data after involvement start date for this cohort:

The stacked plot is useful to understand how recent these involvements are, how much of the attendnance data is either not present (or NA - the red bars here) or out of scope (black bars) meaning this relates to periods of time that have not yet occurred, or for which the data has not been released:

2.28 I&A - Vulnerable Learner

I&A - Vulnerable Learner involvements are associated with very low attendance levels, and show no apparent improvement once the involvement is in place:

The categorised stacked plot reveals more of what’s happening with this cohort - some are of leaving age, but more become just missing from the data. There are also large volumes who become severely absent (0-50% - the orange bars here) after the involvement date.

2.29 Portage

Portage service is associated with very young children with special educational needs.

The team work more with largely children with EHC plans, and across a range of needs, principally those who require SEN support, principally speech, language & communication needs and autism.

Portage
counts of children with involvements in 2024, by SEN level
SEN Support	388
EHCP	334
No SEN	74

 

Portage - counts by IMD quartile
4 = most deprived wards; 1 = least deprived
1	181
2	256
3	276
4	280

Portage
counts of children with involvements in 2024, by primary specific_need
NA	329
Speech, Language And Communication Needs	243
Autistic Spectrum Disorder	195
Severe Learning Difficulty	64
Physical Disability	62
Profound And Multiple Learning Difficulty	48
Social, Emotional and Mental Health	39
Moderate Learning Difficulty	23
Specific Learning Difficulty	19
Other	17
Hearing Impairment	4
Visual Impairment	4
Multi Sensory Impairment	2
No Specialist Assessment	2

For portage we’ve taken attendance data from year 0 to year 6 for children with EHC plan; SEN support; Speech, Language and Communication Needs, and Autism, in each case comparing children with & without a Portage involvement. The differences here are small but children with SEN support have better attendance in all primary years except year 1 when involved with the service. There is also evidence that children with Autism show better attendance. For children with SLCN (which is the most prevalent primary need of children involved with the service) or an EHC plan, the picture is quite mixed, with no clear evidence of a significantly better attendance. (Year 0 is removed here due to low availability of data)

2.30 Rowan outreach

The majority of children with the Rowan outreach involvement have primary needs of Autism or Speech, Language & Communication needs:

Rowan outreach
counts of children with involvements in 2024, by primary specific_need
Autistic Spectrum Disorder	113
Speech, Language And Communication Needs	61
Social, Emotional and Mental Health	24
NA	21
Profound And Multiple Learning Difficulty	2
Hearing Impairment	1
Moderate Learning Difficulty	1
Multi Sensory Impairment	1
No Specialist Assessment	1

The ‘before & after’ plot for Rowan outreach shows a change in average direction, but a slightly messy picture, since there isn’t a lot of full “before & after” data available:

Given that Rowan outreach work with children at or before school readiness age, it makes sense to look at the same primary school attendance profile that we did for Portage above.

Here year 0 is removed due to very small data availability. Although generally children involved with the outreach team show lower attendance than those without, children with Speech, Language & Communication Needs, children with Autism, and children with an EHC plan all show significant and consistent improvement in attendance through years 1 to 3 when involved with the team.

2.31 Visual Impairment team

Involvement with the Visual Impairment team is associated with a small improvement in attendance:

The stacked categorised plot shows an improvement in the highest attendance category over the first year of involvement with the team:

As with the Hearing Impairment team, if we compare attendance levels for children with a Visual Impairment with and without the team involvement, we see consistently lower attendance through all school years for children involved with the team. This presumably reflects a higher level of need, or other factors. Children involved with the team do show improvements in attendance levels through primary school:

2.32 UCAN

(this description copied from Sheffield Directory) Sheffield Early Years Language Centre (Ucan) is funded jointly by the NHS and Sheffield Local Authority. The centre is staffed by speech and language therapists from the NHS and a teacher and assistant from the 0-5 SEND Support Service.

The centre provides intensive early intervention for pre-school children with identified developmental language disorder (DLD) and training for parents and Early Years practitioners in meeting children’s needs.(Please see more detailed information below about admissions criteria and the work of the centre).

Since involvement with UCAN is exclusively prior to school starting age, we cannot do the ‘before & after’ analysis.

The majority of children involved with UCAN have special educational needs:

Portage
counts of children with involvements in 2024, by SEN level
SEN Support	388
EHCP	334
No SEN	74

 

UCAN - counts by IMD quartile
4 = most deprived wards; 1 = least deprived
1	15
2	22
3	32
4	25

The majority have speech, language and communication needs:

UCAN
counts of children with involvements in 2024, by primary specific_need
NA	53
Speech, Language And Communication Needs	37
Autistic Spectrum Disorder	1
Physical Disability	1
Severe Learning Difficulty	1
Social, Emotional and Mental Health	1

Some of this may be due to the recency of the UCAN involvements and the general recovery in attendance for younger children, but the data shows that children with UCAN involvement show better attendance in primary than their peers who have similar needs:

2.33 I&A - School Readiness

Warning

Looking at the age profile, there is clearly a data quality issue with some of these involvements.

The team work more with children in more deprived areas, and those who require SEN support.

I&A - School Readiness
counts of children with involvements in 2024, by SEN level
SEN Support	270
EHCP	146
No SEN	94

 

I&A - School Readiness - counts by IMD quartile
4 = most deprived wards; 1 = least deprived
4	230
3	166
2	124
1	51

I&A - School Readiness
counts of children with involvements in 2024, by primary specific_need
NA	171
Speech, Language And Communication Needs	163
Autistic Spectrum Disorder	136
Social, Emotional and Mental Health	72
Moderate Learning Difficulty	9
Other	8
Specific Learning Difficulty	6
Physical Disability	5
Severe Learning Difficulty	5
Hearing Impairment	4
Visual Impairment	3
Profound And Multiple Learning Difficulty	2
No Specialist Assessment	1

If we look at all pupils, we see that children who the school readiness team work with tend to have lower attendance throughout primary school, including a bigger dropoff into Y6. Children involved with the team show an overall sustained improvement in their attendance through primary school:

Breaking this down for the same characteristic groups we used for various other involvements above, we see generally lower attendance for the children accessing the service, which probably reflects factors not captured in the data - i.e. significant need beyond what we’re controlling for here. The exception seems to be children with Autism, who appear to significantly benefit in attendance.

2.34 S437- School Att Order EHE

Attendance orders related to electively home educated children. In 2024 Sheffield City Council has just 35 of these recorded. The orders can be issued to the parents of children of any age.

S437 Attendance Orders are followed by a sudden and significant improvement in attendance levels.

Behind those averages though, there are still large volumes of children with entirely missing attendance data - are these those still being home educated?

2.35 S437 - School Att Order CME

School attendance orders relating to a Child Missing Education (CME). As with the EHE attendance orders, Sheffield issues just a few dozen of these per year. CME orders or mostly issued to parents of secondary school pupils:

In the ‘before & after’ analysis, we see that prior half terms are blank due to children missing school entirely for this period - hence the attendance order. Average attendance after the attendance orders improves dramatically (but remains below overall average levels).

(note that some half term -1 data here was removed, presumed incorrectly indexed on the wrong date)

All coded absence reasons show a large decrease following the CME attendance orders:

The categorised attendance data shows a dramatic change following this involvement, with a big reduction in the NA category and increases across all ‘present’ categories - though there are also signs that effects have a limited timespan, as the % in the highest attendance category drops away, and the NA category increases once again.

2.36 I&A - Reintegration

The plot above shows a significant shift immedately after the involvement. The categorised reasons show a similar shift whether we look at it half-term to half-term or year on year - a reduction in overall absences, mostly driven by a reduction in exclusions, while illness & lateness levels rise slightly:

Finally, the stacked categorised bar chart shows more detail than the average before & after plot - with the most significant movement being between those severely absent and those in the highest attendance bracket (although these fall away over time):

2.37 School Attendance Order Breach

The plot above shows a significant shift immedately after the involvement. The categorised reasons show a similar shift whether we look at it half-term to half-term or year on year - a reduction in overall absences, mostly driven by a reduction in exclusions, while illness & lateness levels rise slightly:

Finally, the stacked categorised bar chart shows how, despite the increase in average attendance, the majority of children are absent from the attendance records entirely both before and after the attendance order breach involvement:

2.38 GP protocol

There is no apparent affect on attendance levels from this involvement:

The categorised reason plot is not included here, as it shows no particular movement. However, although there is no change in overall average attendance, there is significant movement on the stacked bar plot of attendance brackets. This happens immediately following the involvement, with growth in both the highest and lowest attendance brackets:

2.39 Nurture

There are five different involvement types here all under the banner of ‘Nurture’. It looks like these cover different sites in the city, but the assumption is that they are all broadly the same service and so are covered here together.

Those 5 involvement types are: Nurture - BLT Hinde House Nurture - BLT Yewlands Nurture - Bumblebee Nurture - BLT Earl Marshall Nurture - Step Out

Volumes are quite low and except for Bumblebee much of the activity is very recent:

There are differences in the age profile of the different nurture type involvements.

Nurture type involvements show continued low attendance after involvment - but with some signs of improvement over time:

Nurture type involvements show and immediate and dramatic reduction in exclusion rates:

And returning to attendance, the stacked bar plot of attendance brackets is more revealing than the average attendance levels. Preschool half term periods and some COVID periods have been removed here. Although the overall average doesn’t change, the involvment start date sees a big reduction in the worst attendance brackets (0 - 50%) and growth in both the highest (90% +) and middle attendance brackets (50 - 80%).

2.40 Managed moves

The volumes available for managed moves are very low:

Because the volumes are so small, the confidence intervals here are wide. Managed Moves are associated with a continuing decline in attendance levels:

Looking at coded reasons before & after a managed move, the overall increase in absence seems to be driven by a an increase in ‘no reason’ absences:

3 Child social care

There are three levels of child social care: Child in Need (CIN) Child Protection Plan (CPP) Child Looked After (CLA)

Although these are not involvements in the same sense as the other services & interventions, they have been treated as such in the data so that the start date can be used a zero point to compare attendance levels before and after the episode starts.

Unlike the other involvement types, which have been taken in order of volume, we’ll look at CIN, CPP and CLA together

3.0.1 Child Social Care Volumes

There are around 5000 new CIN episodes start each year.

3.1 Child In Need - effects on attendance

Prior to a CIN episode we see declining attendance levels. The CIN episode is associated with a turnaround in the average direction of travel, but attendance levels remain very low for the following two years. It’s worth noting that the average attendance level for all these children is below 90% for the entire four year window of time we’re looking at here.

CIN episodes show an increase in overall absence in the year following episode start date. Most reason codes stay the same, with a small increase in “no reason” absences:

Before we move on to the CPP episodes, here is the stacked bar plot of bracketed attendance. First we’ll include all categories, since it illustrates how many children with CIN episodes are of pre-school age both before and after the social care start date.

Secondly, here is the plot with the preschool, leaving age and COVID periods removed, so that we can focus on the changing patterns of attendance:

3.2 Child Protection Plan - effects on attendance

Child Protection Plan episodes are associated with a more significant turnaround in attendance levels:

CPP episodes are associated with decreases in all coded reasons. These decreases are much more apparent at the half term level than the annual level:

Here is the categorised stacked bar plot for Child Protection Plans, showing an immediate and sustained increase in the number of children with the highest levels of attendance:

3.3 Child Looked After - effects on attendance

CLA episodes show the most dramatic change in attendance levels of anything we’ve covered in this report:

CLA episodes show a significant decrease across all all coded absence reasons except family holidays (which is less of an issue for this cohort in any case):

Finally here is the stacked categorised attendance plot for Child Looked After episodes, showing an immediate and sustained increase in the proprtion of children with the highest levels of attendance:

3.4 Conclusions

This is the end of this report. The picture in terms of effectiveness is very mixed, though a few themes have emerged:
- Most interventions see a change in overall direction of change in attendance, but rarely do we see a net gain
- Some interventions are effective on exclusion rates, but not on overall attendance.
- Interventions are more effective on younger children.
- There is some evidence that girls have better responses than boys.
- Children’s Social Care episodes are associated with particularly strong improvements in attendance.
- The Family Intervention Service is associated with an immediate improvement, and sustained improvement follows. The very high volumes of this service, and relatively low attendance of children in the run up to an episode of FIS, together suggest an opportunity.

Finally, we should restate that where we have found no evidence of a direct increase in attendance levels following intervention we must emphasise that this is not evidence that such interventions are failing or not worth doing - there may be benefits beyond what is visible in the attendance data, and the likelihood is that these interventions are having a limiting effect on what would otherwise be an even worse situation.