Risk Stratification

The Suffolk Multi-Agency Safeguarding Hub (MASH) is a collaboration of organisations in Suffolk who are responsible for safeguarding vulnerable children and adults. The MASH triages notifications of need from agencies and individuals with safeguarding responsibilities and enables shared information and a joined up approach between the relevant organisations.

The aim of this initiative is to streamline the time consuming process of notifications, decrease the number of false negatives and support the MASH decision making in general. Using advanced machine learning techniques to their data the Risk Team has built a machine learning system which assesses the level of severity of the notifications received.

• Read our progress up-date (.pdf)

We are currently testing the accuracy of the model using live data.

The numbers of children entering care in England have been increasing significantly year on year since 2014 and at the same time the funding supporting local authorities to tackle the issue has being reduced over recent years. This initiative aids the efficient and timely identification of children in need (CiN) who are at high risk of going into care (CiC), so that preventative interventions can take place to avoid the complexities and costs involved with entering care.

Using machine learning methods the team has built a predictive model which assigns a risk level to individual cases and will support the decision making of case workers and help them to prioritise their workload. The accuracy and specificity of the model has now been verified using historical data.

Collaborating with Essex Police in Colchester this initiative uses data analysis to:

• increase understanding of the characteristics relating to knife crime in Colchester.
• increase ability to focus resources
• assess the effectiveness of knife crime interventions

Essex Police have observed an increase in serious sexual offences over the past 3 years and this is particularly prevalent amongst young people aged 18 and under. Essex Police’ observations also indicate an increase in unhealthy relationships and grooming of young girls.

The aim of this project is to map serious sexual offences across the whole of Essex unpicking a variety of parameters in order to enable Essex Police to implement effective interventions where they are most needed.

The Risk Team have created an on-line platform in order to help Essex Police with their report writing and the easy visualisation and breakdown of data.

The Catalyst Risk Stratification Team have worked alongside Essex Partners on the School Readiness pilot initiative. The Team applied data analyses and advanced machine learning techniques to anonymised data to build a predictive model which was applied to identify children who may not be ready to start school by the age of 5 in a selected pilot area.

As part of this project the Catalyst Risk team has contributed academic expertise to the integration of machine learning techniques in the Council’s decision making processes. The model has been used by Essex Partners to further develop the Essex Data (ED) programme, which is part of the county’s ‘Future of Essex’ strategy.

Commissioners have been able to use insight from the ED programme to apply for £3.35 million funding for interventions, to inform work with Ofsted, and to form a ‘New Generations’ community group consisting of parents, head teachers and charity representatives, which has implemented a new nursery and family boot camp in the pilot area.

• Read more in the Case Study.

The Catalyst Risk Stratification Team identified and merged relevant datasets for Suffolk County Council to enable the identification of local children who were eligible for, but not receiving free school meals (FSM).

By notifying local schools of this shortfall and encouraging the uptake of FSM by eligible pupils, schools would subsequently be able to increase their applications for pupil premium payments. It was predicted that this work had the potential to result in between £4 and £15 million of extra funding from Central Government for local schools.

The Catalyst Risk Stratification Team analysed data on reported instances of mass marketing fraud for Suffolk Trading Standards, producing a list of postcode hotspots. This was supplemented by a literature review detailing the common setbacks to conducting research on this topic.

The Catalyst Risk Stratification Team conducted exploratory analysis on a cohort of assessment data for the Youth Offending Team at Suffolk County Council.

This project uncovered patterns within the data and enabled the identification of protective and harmful characteristics in the context of youth offending. The team were also able to recommend how data collection and recording practices could be improved in order to enable more sophisticated analysis to be undertaken.

School readiness is a term that describes a child’s ability to engage in and benefit from early learning experiences. Evidence shows that children who are not ready for school and do not meet key developmental milestones experience additional challenges compared to their peers. Suffolk County Council (SCC) would like to ensure that all local children are ready to start school at the age of 5 and are measured as having a ‘good level of development’ as specified by national guidelines.

Therefore SCC would like to know if they can uncover any factors in their data that makes a child less ready so that they can recommend targeted interventions. However, they would also like to assess the effectiveness of these interventions so that the less effective ones can be scaled back and cost savings made.

The Catalyst Risk Stratification team is analysing 5 education datasets to identify the interventions administered by SCC health and children’s centres that are most effective and have a positive impact on a child’s readiness for school, and secondly to establish the attributes that contribute to a child not being ready for school.

The Catalyst Risk team is working with Suffolk County Council to develop a tool using data analysis of a range of anonymised pupil’s educational data to aid front line teams who work with young people who are either NEET or at risk of becoming NEET.

The analysis aims to answer a range of questions to provide information that could help identify future preventative measures to reduce a person’s chance of becoming NEET.