Methodology

Children in need: A focus on re-referrals

• Published: 26 February 2026

Source

The ad-hoc analysis of re-referrals to children's social care is based on data collected in the children in need census. For a full explanation of the methodologies for data collection and processing see Children in Need methodology. 

In summary, the children in need census is an annual child-level data collection by the Department for Education (DfE) which collects individual assessment information and child characteristics for each child that has been referred to children’s social care services. 
 
The census is a statutory requirement under section 83 of the Children Act 1989 and data is collected from each local authority in England. 
 
The data items collected in the children in need census include child identifiers and characteristics along with the dates of any referrals to children’s social care services; assessments; section 47 enquiries; child protection plans and reviews; and from 2022 onwards child in need plans. 

Data quality and cleaning

Data  cleaning

The children in need census data is collected through the DfE COLLECT (collections online for learning, education, children and teachers) system. 
 
When local authorities upload their data to COLLECT a series of validation rules are run to identify any errors or queries with the data. Local authorities then undertake cleaning of their data to resolve these issues; with the department providing help and support throughout this period. A local authority can also add notes to their data return to explain any outstanding issues. These notes are examined by the department during the quality assurance process and, where applicable, followed up and resolved with the individual local authority. 
 
Once all errors and queries have been resolved or explained, local authorities submit a final return. A full list of validation rules for checking queries and errors is available in the children in need census guide to submitting data (opens in a new tab) (opens in new tab). 
 
DfE then carry out further quality assurance of the data so that it can make an assessment of the data quality and highlight any issues to users. 

 
Further validation checks are built into COLLECT to highlight where key measures (for example the number of referrals) have increased or decreased substantially between census years. Where changes exceed the threshold DfE requests local authorities to check their data and confirm that it is correct. It should be noted that a large change between years does not necessarily imply the information provided is inaccurate and may be explained, for example, by changes in local practice.

For a full explanation of the methodologies for data cleaning see Children in Need methodology. 

Issues relating to specific local authorities

Hampshire, 2024

Hampshire local authority moved to a new case management and reporting system and their return in 2024 had significant data quality issues and was assessed to not be sufficiently reliable to use. Therefore, their data for 2024 is presented in all of the underlying data files as ‘u’ to indicate low reliability and 2023 figures for Hampshire are included in the 2024 totals for the South East region and England.

Hackney, 2022, 2021

Hackney had a cyberattack in December 2020, which had a significant impact on their information management systems. As a result, 2020 figures for Hackney have been included in the 2021 and 2022 national and regional totals, but data in the underlying data files for Hackney has been presented as ‘x’ to indicate not available.

Factors identified at the end of assessment

Factors identified at the end of assessment were collected and reported for the first time in 2014; however, data was only published at national level due to concerns about its quality. Since 2015 data has been published at local authority level, with coverage continuing to improve.   

It should be noted that not all episodes have factors recorded, but this has improved over time. Nonetheless, there can be differences in the recording practices between local authorities therefore this data should be treated with caution and shouldn’t be taken to represent the national, regional or local authority level prevalence of particular issues. 
 
In addition, local authorities have been encouraged to report all relevant factors identified at the end of assessment, which may account for some of the increases seen in recent years and means the majority of children have more than one factor recorded for each episode of need. This should be taken into consideration when assessing time series assessment factor data.

Longitudinal Analysis

Re-referrals

Within the release we have applied longitudinal analysis to better understand re-referrals over longer time periods, duration of episodes, levels of intervention and assessment factors associated with referrals. This analysis uses the Longitudinal CIN dataset (LCD) which means that some data may not exactly match published Children in Need statistics due to matching methods. 

This analysis shows data from 2017 to 2023. This is because the LCD only contains data up to 2024 and the analysis requires a full year of data after the reference year to see whether initial referrals will become or not become a re-referral.

The analysis disaggregates referrals into three groups:

• Initial referrals that will lead to re-referrals;
• Initial referrals that will not lead to re-referrals; and
• Re-referrals

Child Protection Plans (CPPs)

New analysis has been conducted linking together 5 years of data from the Children in Need census between the years 2019-20 to 2023-24 and using unique identifiers to determine how many children aged up to and including 16 were subject to a Child Protection Plan (CPP) at some point during the reporting period. This linking took account of changing identifiers across years and local authority reorganisations. The linking accounted for children who moved across local authorities where the Unique Pupil Number (UPN) was recorded in both authorities, therefore cases where a child moved but the UPN was not known will appear as distinct children in the dataset. 

The estimate includes children from pre-birth to age 16; those aged over 16 at the end of the reporting period have been excluded.

This analysis has shown there were 291,600 children subject to a child protection plan at least once over a 5-year period (1 April 2019 to 31 March 2024), representing around 3% of 0-16 year-olds (or 1 in 39). Of which, 237,370 children were not on an active child protection plan at the end of the reporting period, representing around 2% of 0-16 year-olds (or 1 in 48).

Definitions

Referrals

A referral is defined for the purposes of the children in need census as ‘a request for services to be provided by children’s social care services’. This is in respect of a case where the child is not previously known to the local authority, or where the case was previously open but is now closed. A referral should not be recorded if there is a case for the child already open. A referral can be made by a professional from one of many different agencies but the term as used here is a broad one which encompasses referrals from any source, including self-referrals. 
 
A referral may result in: 

• an assessment of the child’s need
• the provision of information or advice
• referral to another agency
• or no further action.

If a child is referred more than once in the year then each referral is counted in the figures in this publication. 

Assessments

Local authorities undertake assessments of the needs of individual children to determine what services to provide and action to take. An assessment should be completed within 45 working days of a referral. 
 
An assessment is undertaken for any child who has been referred to children’s social care services with a request that services be provided. Statutory guidance ‘Working Together to Safeguard Children’ was revised in 2013 (and again in 2018) giving local authorities more flexibility when assessing children. Previously, local authorities carried out an initial assessment within 10 working days and (where needed) a more in-depth core assessment within 35 working days. All local authorities now carry out a single continuous assessment within 45 working days.

Factors identified at the end of assessment

Factors identified at the end of assessment are additional factors that social workers record as being relevant in a case. 

The reported figures are based on assessment factors recorded against individual episodes of need, which begin when a child is referred to children’s social care services and is assessed as being in need of children’s social care services. Each unique factor is counted once against a given episode, irrespective of the number of times the same factor was recorded in that episode. However, as a child can have more than one episode of need during the year (ending 31 March), the same child can be recorded more than once for a given factor. 

The majority of children have more than one unique factor recorded for each episode of need. 

Data processing

Referrals and assessments

The method used to calculate referrals and assessments is outlined in the table below.

Measure	Definition
Number of referrals	A count of referrals, where the referral date is within the collection year 1 April to 31 March.
Number of assessments completed	A count of assessments completed, where the assessment authorisation date is within the collection year 1 April to 31 March.
Duration of assessments	The duration of an assessment is calculated as the time in working days between the assessment start date and the assessment end date.
Referrals within 12 months of a previous referral (re-referrals)	A count of the number of referrals within 12 months of a previous referral. Re-referrals are calculated based on data from the latest and previous collection years. For example, re-referrals for 2024 are calculated as follows: Select the referrals that occurred between 1 April 2023 and 31 March 2024 in the 2024 data. Select the referrals that occurred between 1 April 2022 and 31 March 2023 in the 2023 data and remove any duplicated children, keeping only the latest referral for the child. Sort the referrals for each child by date. Calculate the time in days between the last referral and the penultimate referral for each child. The referral is added to the count of re-referrals if the difference between the pair of dates is 365 days or less.
Referrals resulting in no further action	A count of referrals in the collection year 1 April to 31 March where the referral resulted in no further action. Referrals resulting in no further action are not included in the other children in need population measures (eg, the number of children in need at 31 March).
Assessed not to be in need (of children's social care services)	A count of referrals within the collection year 1 April to 31 March, where: The referral does not end in no further action before an assessment has taken place  The referral has an assessment, but no further activity such as a section 47 enquiry or child protection plan; and The referral has reason for closure recorded as ‘RC8, case closed after assessment, no further action’ or 'RC9, case closed after assessment, referred to early help' Referrals which led to an assessment where the child was assessed as not in need of services are not included in the other children in need population measures (e.g. the number of children in need at 31 March).
Factors identified at the end of assessment	A count of factors identified at the end of assessment for each episode where the assessment authorisation date is within the collection year 1 April to 31 March. Duplicate factors are removed so that an individual factor is counted only once against an episode.

Confidentiality and rounding

Confidentiality

The following symbols are used in the underlying data files as follows: 

Symbol	Meaning
c	to protect confidentiality - secondary suppression may also be applied to figures to protect confidentiality
z	data not applicable
x	data not available
u	low reliability
k	values rounds to zero but is not zero, for example where a percentage is less than 0.05%

Where any number is shown as zero, the original figure submitted was zero.

Rounding

England and regional totals are rounded to the nearest 10, therefore totals may not equal the sum of their component parts. 

Percentage are rounded to one decimal place, therefore component parts may not sum to 100. 

Rates per 10,000 children aged 0 to 17 years are rounded to one decimal place.

Related statistics

Characteristics of children in need in England:  Information on referrals, assessments and children who were the subject of a child protection plan (taken from the Children in Need census) are published annually.

Outcomes for children in need, including children looked after by local authorities in England: a range of outcome measures at national and level authority level for children in need  including children looked after by local authorities in England.

Help and support

Contact us

If you have a specific enquiry about Children in need: A focus on re-referrals statistics and data:

Children's Social Care Analysis and Research

Email: Bruce.GOLDING@education.gov.uk
Contact name: Bruce Golding

Press office

If you have a media enquiry:

Telephone: 020 7783 8300

Public enquiries

If you have a general enquiry about the Department for Education (DfE) or education:

Telephone: 037 0000 2288

Opening times:
Monday to Friday from 9.30am to 5pm (excluding bank holidays)