MD3007 SafeStats data portal and ISTV programme

1.1.    This decision covers the costs of running the SafeStats data portal and the Information Sharing to Tackle Violence (ISTV) programme for the next three financial years, namely 2022-23 to 2024-25.

1.2.    SafeStats is a secure data platform hosted, managed and maintained by the GLA. The platform hosts London-based community safety data from nine different agencies, including the Metropolitan Police Service (MPS) and the London Ambulance Service (LAS). None of the datasets hosted on SafeStats contain directly personally-identifiable information, due to either the limited data fields shared or the data having undergone a process of pseudonymisation. Access to SafeStats is limited to named individuals within approved organisations, who have been authorised by their organisational administrator as having a clear purpose for access. SafeStats enables these authorised users to download the hosted data for legitimate and lawful purposes under section 17 and section 115 of the Crime and Disorder Act (CDA) 1998. 

1.3.    The primary aim of SafeStats is to support partnership work on crime and disorder reduction by enabling community safety analysts working within local authorities, emergency services, and other organisations involved in community safety work such as MOPAC that have a legitimate purpose under section 17 of the CDA 1998, to access multi-agency data in a secure, standardised and lawful way. 

1.4.    The London-based Information Sharing to Tackle Violence (ISTV) programme is sponsored by the Mayor’s Office for Policing and Crime (MOPAC) and delivered by the GLA’s City Intelligence Unit (CIU). The programme involves the monthly sharing of violence-related Accident and Emergency (A&E) attendance data for the purposes of violence reduction and prevention.

1.5.    As part of the delivery of the London-based ISTV programme, the GLA lead on all aspects of the daily running of the programme, including hospital engagement and training, liaison with the NHS Violence Reduction Programme, collation of the ISTV data submissions from A&E departments and the hosting of the data through the SafeStats data platform. 

1.6.    The data shared is de-identified and is governed by the NHS Information Standard, ISB 1594. The specific data items shared are the time and date of arrival in A&E, the time and date of the violent incident, the primary means of assault (for example the type of weapon that was used), and the specific location of the incident. 

1.7.    Funding for 2022-23 through to 2024-25 is being provided separately by the MPS and MOPAC, the latter of which is on behalf of the London-based Community Safety Partnerships (CSPs) and the Violence Reduction Unit (VRU).  

1.8.    This decision is preceded by approvals through MD2772 (2021-22) and Director Decisions DD2485 (2020-21) and DD2457 (2020-21).  

1.9.    This specific decision seeks approval to receive and spend grant funding of £555,000 from MOPAC and £50,000 from the MPS for the GLA to continue running the SafeStats data portal and the ISTV programme. This expenditure will be received and spent across the 2022-23, 2023-24 and 2024-25 financial years.

1.10.    MOPAC has confirmed that they have sufficient budget to cover the grant set out in this decision. 

1.11.    The GLA’s dual involvement with the SafeStats data portal and the ISTV programme aligns to the principal purposes of the GLA set out in the GLA Act 1999 by furthering the promotion of social development in Greater London. Through informing interventions to reduce/prevent crime and the associated harms, SafeStats and the ISTV programme allow for the well-being of individuals within Greater London to be improved. 

1.12.    The proposed expenditure related to SafeStats and the ISTV programme concern the GLA cooperating with, as well as facilitating and coordinating the activity of MOPAC and cooperating with Transport 

2.1.    The data held on SafeStats (inclusive of the ISTV data) is primarily used by organisations working within community safety to help reduce and prevent crime and disorder. The data is used within a series of analytical outputs such as Joint Strategic Needs Assessments (JSNAs), Community Safety Partnership tasking presentations and problem profiles. These outputs are used to improve the lives of Londoners by reducing harms, reducing victim rates and reducing re-offending. 

2.2.    SafeStats provides support to Mayoral priorities, including the Police and Crime Plan, as well the MPS’s ‘Met Direction’ strategy and local authorities’ strategic priorities. 

2.3.    SafeStats enables users to analyse crime and disorder incidents that have been reported directly to the police, as well as those that have not, such as ambulance dispatches to violence-related injuries. This allows for a more insightful and comprehensive understanding of London’s community safety issues, while helping to inform upon the associated drivers and risk factors. While there is likely to be some overlap between the incidents reported upon by the MPS and the LAS, there is no means of determining which incidents are contained within both datasets from the data held on SafeStats. This is because there is no consistent incident number or reference that runs across these two datasets, nor across any of the other datasets held on SafeStats. 

2.4.    None of the datasets held on SafeStats contain directly personally-identifiable information, due to either the limited information recorded/provided or the data having undergone a process of pseudonymisation (e.g., the spatial suppression of dispatch location provided by the London Ambulance Service). However, due to the common categories of information featured in the datasets hosted within SafeStats, and their subsequent ability to be cross-referenced, the data, when considered as a whole, should technically be considered as potentially ‘containing data that may indirectly identify an individual’. 

2.5.    The London Ambulance Service dispatch dataset is the only dataset containing any characteristics related to the data subjects (specifically age and gender). However, the combination of information within these records is not sufficient to directly enable the identification of an individual. It may, however, be possible, through ‘jigsawing’ together a combination of relevant information across multiple datasets relating to the location, category and/or date/time of an incident (as well as any personal characteristics that exist) to identify an individual. 

2.6.    However, the SafeStats Conditions of Use document, which users agree to abide by upon every system login, explicitly prohibits this, stating “SafeStats datasets should never be used with each other or with other available data to ‘match’ or ‘jigsaw’ with the intention of identifying an individual”.

2.7.    The potential for identifying the data subjects featured in the SafeStats datasets is significantly limited at the GLA processing stage. The data is very limited at source, securely stored, accessible only by a small number of highly trained and authorised GLA staff, and only has the potential to identify an individual data subject under very specific circumstances that rely upon additional information being obtained from data sources external to SafeStats.

2.8.    It must also be considered that even in a ‘worst-case scenario’, where all the data available on SafeStats concerning either a specific individual or incident is combined, the collective information would not be greater than could be derived from an eyewitness account or a media report. 

2.9.    Despite this, a Data Protection Impact Assessment (DPIA)  was completed for SafeStats to cover the collection, collation and storage of personal information. The DPIA specifically relates to the GLA’s responsibilities as both a Data Controller and the initial processor of the datasets. All SafeStats-accessing organisations (‘end-users’) are considered Data Controllers for the SafeStats data that they download, and as such are recommended to complete their own DPIA.

2.10.    The community safety data hosted on SafeStats is often unobtainable by any other means, or where it is obtainable, requires extensive data governance discussions and documentation to be drafted and ratified by individual organisations. 

2.11.    While the data derived from SafeStats is used by the accessing organisations to inform a wide array of different outputs and decisions, relevant examples include:

• demand analysis to help inform service provision and early intervention programmes
• analysis to support the tasking of community safety partnership funding and resources
• analysis to support local government policy and strategy
• undertaking public health analysis to look at criminogenic needs. 

2.12.    The MPS are currently undertaking a phased review and withdrawal of direct access to their systems for non-MPS personnel. This is related to the imminent introduction of Connect, the MPS’ new integrated data solution. The intention of the MPS is to withdraw the in-scope third party access prior to the implementation of Connect. The MPS have proposed that data could then be made available through alternative data feeds as part of the development of their Open Data Strategy. This includes increasing the number of MPS datasets being hosted on the SafeStats data portal.

2.13.    The additional MPS datasets proposed for inclusion in SafeStats refer to de-personalised victim data (ethnicity, sex, and broad age range), perpetrator data (ethnicity, sex, and broad age range), recorded crime outcome data, and data regarding anti-social behaviour (ASB) incidents (location, temporal information, response times, and the type of incident). These datasets will only include the de-personalised details of the victims and perpetrators (where relevant) and will not include data relating to any associates, witnesses or informants.

2.14.    The income received from the MPS will fund both the system process work and the website application development work required to enable the GLA to securely host these four additional MPS datasets on SafeStats. 

2.15.    The Assistant Director of the City Intelligence Unit at the GLA, is responsible for the achievement of the below SafeStats-specific objectives for 2022-23: 

• providing authorised SafeStats users access to timely crime and community safety data, in-line with the sanctioned Data Sharing Agreements
• continuing to develop and maintain the SafeStats data platform with a focus on improving the tools and search query functions available to users to enable accurate, bespoke outputs
• pursuing the acquisition of additional data sources to host through SafeStats, including the negotiation of additional datasets from the MPS that will enable users to analyse victims, perpetrators, recorded crime outcomes, and ASB 
• developing regular analytical outputs using SafeStats data to monitor trends and identify areas for further analysis, including quarterly community safety trend reports and a borough core dataset for violence
• through the Strategic Crime Analysis Team at the GLA, working with MOPAC, the VRU and GLA policy teams to maximise the internal use of SafeStats data and insight.    

2.16.    The Assistant Director of the City Intelligence Unit at the GLA, is also responsible for the achievement of the following ISTV-specific objectives for 2022-23:

• facilitate the delivery of hospital A&E data on violent incidents, by: 
  • working with hospitals who aren’t currently supplying ISTV data on a regular basis
  • providing training for A&E staff in the collection of ISTV data
  • liaising with hospital A&Es to improve the consistency and quality of their ISTV data submissions 
• ensuring that appropriate data governance arrangements remain in place and are up-to-date to reduce the risks to MOPAC/GLA
• engage with the London-based CSPs, MPS and other relevant stakeholder groups to promote awareness of the programme and encourage the utilisation of the ITSV data at both tactical and strategic levels
• expand upon the evidence-base for the beneficial impact of ISTV in relation to violence reduction.

3.1.    Under section 149 of the Equality Act 2010, as a public authority, the GLA must have ‘due regard’ to  the Public Sector Equality Duty; that is, the need to: 

•    eliminate unlawful discrimination, harassment, and victimisation
•    advance equality of opportunity
•    foster good relations between people who share a protected characteristic and those who do not.

3.2.    This involves having due regard to the need to remove or minimise any disadvantage suffered by those who share a relevant protected characteristic that is connected to that characteristic, taking steps to meet the different needs of such people; and encouraging them to participate in public life or in any other activity where their participation is disproportionately low.

3.3.    SafeStats meets this public sector equality duty in two ways: advancing equality of opportunity and fostering good relations between people who share a relevant protected characteristic and those who do not through analysis conducted using the data as well as through robust data governance to prevent discrimination.

3.4.    The Mayor’s Equality, Diversity and Inclusion Strategy sets out how the Mayor works to create a fairer, more equal, integrated city where all people feel welcome and are able to fulfil their potential. The analysis completed on SafeStats-derived data furthers this strategy by contributing to the aims of London being a “safe, healthy and enjoyable city”, “a great place for young people”, and a place that people feel (and are) safe “getting around”. 

3.5.    Both the SafeStats data portal and the work informed by SafeStats data, feed into several of the Equality, Diversity and Inclusion Strategy’s objectives, including: 

•    To work through TfL and with the London boroughs, transport providers, the Metropolitan Police Service (MPS), the British Transport Police (BTP) and the City of London Police (CoLP) to help reduce crime, and the fear of crime, on London’s streets and transport system.

The SafeStats data portal hosts data from four of these explicitly-mentioned agencies (TfL, MPS, BTP, and CoLP), as a means of assisting users in understanding the type and level of crime and disorder occurring on the transport network. The data enables the identification of transport hubs, routes and stops that have disproportionately high offending linked to them. This in turn allows these locations to benefit from targeted interventions to transform them in to safer and more secure locations, such as through designing out crime initiatives. 

•    To work with local authorities, the MPS, criminal justice agencies and voluntary and community sector partners to help address the impact of crime on those groups and communities disproportionally affected, particularly with respect to hate crime, domestic violence and violence against women and girls.

SafeStats enables users to analyse an array of community safety issues through the provision of multi-source data. This includes recorded and unrecorded police incident data. Since some crime types have lower than average police reporting rates, such as in the instance of domestic abuse, it is essential that community safety professionals can produce a composite understanding of each issue using data from a range of different sources. This is especially important as the victims of under-reported crime types are frequently the most vulnerable and susceptible to repeat victimisation. This also links to health inequality as the victims (and perpetrators) of violence are consistently reported to have higher health needs and worse health outcomes across a range of measures compared to the rest of society.

•    To work with all relevant partners to help understand and reduce the gap in risks of fires between different communities and engage local communities to better understand and respond to their needs.

SafeStats hosts data for the London Fire Brigade (LFB) deployments to intentional and unintentional fires. Analysts use this data to enable an intelligence-led approach to helping vulnerable/high risk communities to protect themselves from fires. 

3.6.    The users of the SafeStats data platform work within the GLA, local authorities, the police, the NHS, and functional bodies of the GLA. All these organisations must comply with the Public Sector Equality Duty (PSED) when they carry out their functions or services. Any decision or adoption of new policy that has been informed by data derived from SafeStats (including the ISTV data), is therefore subject to compliance with the PSED. 

3.7.    Data obtained from the SafeStats data portal regularly feeds into several GLA outputs that assist in narrowing social, economic and health inequalities, such as the London COVID-19 Resilience Dashboard. 

3.8.    SafeStats data is also key to informing the public health approach to violence reduction, both through the provision of routine analysis and insight to the VRU Partnership Reference Group (PRG) and the NHS London Violence Reduction Programme (VRP). 

3.9.    All these uses of SafeStats data respond to and facilitate discharge of the PSED by both recognising the disproportionate impact of crime and safety on London’s diverse communities and by facilitating a collaborative approach to developing interventions and preventative solutions. 

3.10.    SafeStats data enables a much deeper and meaningful understanding of the community safety issues affecting London than reliance upon police data alone. Through including crime and disorder incidents not reported to the police, the GLA ensures that (as part of its on-going legal responsibility to have due regard to the need to promote equality, in everything it does, including its decision-making), barriers are removed that may prevent those with protected characteristics benefiting from interventions, support and engagement. 

3.11.    SafeStats contains some limited de-personalised data on age and sex within one of the hosted datasets (LAS dispatch data), but otherwise does not contain data relating to the protected characteristics of disability, gender reassignment, pregnancy and maternity, race, religion or belief, or sexual orientation.

3.12.    The ISTV dataset contains no data relating to any of the protected characteristics stated in the PSED nor contains any information at all related to either the victim’s characteristics or the perpetrator’s characteristics.

3.13.    There is no evidence in either the Mayor’s Equality, Diversity and Inclusion Strategy or the underlying evidence-base that suggest some groups face specific barriers or unequal outcomes because of the SafeStats data portal or the ISTV programme.

3.14.    Analyses conducted using the data contained in the GLA SafeStats data portal are used to inform the Mayor on areas of London which are affected by a range of crimes (including those not reported to the police), and vulnerable groups of people/communities who are victims of crime. The driving force of analysis conducted by organisations accessing SafeStats data is to improve equality of opportunity for Londoners by creating an improved understanding of victims of crime, helping those identified as being vulnerable, and assisting in the enforcement of those who seek to harm. 

3.15.    All members of the SafeStats team have successfully completed the GLA Public Sector Equality Duty training. The team are content that there is no potential for SafeStats or the ISTV programme to discriminate against a specific group. 

3.16.    Between 2020 and 2021, the GLA (as led by the SafeStats team) took part in the Information Commissioner Office’s (ICO) Regulatory Sandbox. This Sandbox is a service introduced by the ICO to help support organisations create services/products (that include personal data) in a safe and novel way. 

3.17.    During this Sandbox, research was commissioned to help further understand the public’s view of sharing personal data for the purposes of violence reduction. The research used specified selection criteria to ensure that the individuals consulted were those most at risk of involvement in violence, either as a victim or a perpetrator. This helped the GLA to comply with Article 35 (9) of the GDPR (seeking the views of data subjects where possible, as part of a data protection impact assessment). 

3.18.    The ICO concluded that this commissioned research enabled the GLA to articulate which areas of data protection the public had the most concerns about as well as demonstrating general support for the proposed data-sharing activity in respect of the balancing of the rights of the data subjects against the wider public benefit of the data processing. The ICO further added that this consultation acted as an important step in ensuring the GLA demonstrated the appropriate care and respect for the rights of the data subjects whose data it processes and hosts on the SafeStats data platform.

3.19.    The SafeStats team has previously undertaken a pan-London needs and requirements consultation with the London-based CSPs and the SafeStats data-providing agencies. Views were sought through a combination of one-to-one interviews, consultation meetings and consultation surveys. The consultation explored the current use of the system by the CSPs as well as their desired future users of the system, including discussions of dataset availability and disclosivity, data tools and visualisations. The findings from this consultation subsequently informed the re-development of the SafeStats data platform. 

3.20.    An additional user survey was then undertaken in 2020 which looked specifically at the needs and requirements of the London-based CSPs and the VRU in relation to their work around violence reduction/prevention. The findings of this survey highlighted that there was noticeable variation in the availability and accessibility of violence-related data across CSPs, and that users needed access to more expansive and granular data which included additional demographic data. These findings have since been used alongside the necessity and proportionality principles of data-sharing to help inform upon applications for additional SafeStats data feeds. 

4.1.    The key risks and issues associated with the SafeStats data portal are: 

• the successful and timely completion of the ongoing development work of the platform within budget. This is needed to ensure that the system continues to meet user need, can create automated outputs that were previously undertaken manually in an inefficient manner and to ensure continuity of service for users during the MPS-systems access revoking for non-MPS personnel. 
• the successful negotiation of new and improved data flows into SafeStats to meet the evolving data needs of the current users and expanding user base. 

4.2.    The key risks and issues associated with the ISTV programme are: 

• the ISTV data submissions from the A&Es to the GLA being negatively affected by issues such as key personnel leaving the A&E department, no A&E resource capacity available and/or lack of support from senior A&E managers
• the data collection being impacted upon by IT issues, the roll-out of self-booking systems in A&Es and/or the roll-out of the Emergency Care Dataset (ECDS)
• the GLA experiencing difficulties accessing the ISTV data once it has been incorporated into the ECDS, with A&Es unaware of the continued requirement to engage with the ITSV programme.

4.3.    All the identified risks associated with the SafeStats data portal and the ISTV programme are documented within risk registers. These risks and associated mitigations are routinely reviewed, assessed, and updated accordingly. 

4.4.    If the ongoing SafeStats service is not maintained, this would result in significant barriers to the completion of community safety analysis across London. Should community safety analysts not be able to access multi-agency data in one system (SafeStats), this would vastly reduce the efficiency and effectiveness at which they can source, cleanse, manipulate and analyse the various datasets. Several of the datasets hosted on SafeStats are unlikely to be made available by an alternative means, which also risks them being excluded from analytical outputs. This would then affect outputs such as those used to inform tactical taskings, strategic direction, service provision and commissioning, policy development and outcome monitoring. 

4.5.    If the ISTV programme ceased to exist, it would impact greatly upon the quality of the violence-related analysis being completed across London. Without the programme enabling the sharing of violence-related A&E attendance data, it is unlikely that this data would continue to be shared directly with all community safety units for them to utilise within their violence work. Potentially, these analytical outputs would then concentrate almost exclusively on incidents of violence that have been reported to the police. This would invariably have a bigger and more significant impact when trying to analyse violence themes that have disproportionately high levels of unreported incidents, such as domestic abuse, sexual violence, and gang-related violence. 

4.6.    The SafeStats data platform and the ISTV programme link to several Mayoral strategies and priorities, including the Police and Crime Plan for London 2022-25, the London Knife Crime Strategy, the Tackling Violence against Women and Girls Strategy, Thrive LDN: Improving Mental Health, the 24-hour London Vision, and the London Health Inequalities Strategy. 

4.7.    Access to the SafeStats data portal is only granted to new organisations upon the combined satisfactory completion of the SafeStats security audit checklist and the drafting/signing of a SafeStats Data Sharing Agreement (DSA). 

4.8.    Upon every SafeStats system login, all individual users confirm that they have read the terms and conditions of use of SafeStats and consent to abide by them. 

4.9.    All individual user system log-ins and data query downloads are fully audited. 

4.10.    There are no conflicts of interest identified from any of those involved in the drafting and clearance of this report. 

5.1.    Mayoral approval is sought for the receipt and expenditure of £555,000 of revenue grant funding from MOPAC and £50,000 from the MPS for 2022-23 to 2024-25 to deliver both the SafeStats data service and the Information Sharing to Tackle Violence (ISTV) programme. The MOPAC payment is made centrally to avoid collecting individual payments from each of the local authorities within London and other accessing organisations. MOPAC will be invoiced by the GLA annually in arrears for the SafeStats and ISTV service/programme delivery. The MPS will be invoiced for the full amount at the conclusion of 2022-23. 

5.2.    The funding totalling £605,000 across 2022-23 to 2024-25 is to deliver both the pre-existing SafeStats data service and the ISTV programme, comprising of the below:  

5.3.    The SafeStats budget, income and work is managed by the City Intelligence Unit (CIU) within the GLA’s Strategy and Communication directorate.

6.1.    The foregoing sections of this report indicate that the decisions requested of the Mayor fall within the statutory powers of the GLA to do anything which it considers will further the promotion of and/or to do anything which is facilitative of, or conducive or incidental to the promotion of,  social development in Greater London and in formulating the proposals in respect of which a decision is sought officers have set out above how they have complied with the Authority’s related statutory duties to:

• pay due regard to the principle that there should be equality of opportunity for all people
• consider how the proposals will promote the improvement of health of persons, health inequalities between persons and to contribute towards the achievement of sustainable development in the United Kingdom
• consult with appropriate bodies.

6.2.    Section 31(1) of the Greater London Authority Act 1999 (the “GLA Act”) prohibits the GLA from incurring expenditure on activity which may be undertaken by MOPAC, TfL or the London Fire Commissioner. However, section 31(6) of the GLA Act permits such expenditure where it relates to cooperation with, or the facilitation or coordination of activity of, such bodies. Officers have indicated at section 1 (above) of this report that the expenditure proposed concerns such cooperation and is facilitative of, or coordinates the activity of, those bodies.

6.3.    In taking the decisions requested, the Mayor must have due regard to the Public Sector Equality Duty (section 149 of the Equality Act 2010); namely the need to eliminate discrimination, harassment, victimisation and any other conduct prohibited by or under the Equality Act 2010 and to advance equality of opportunity between different groups of persons who share a protected characteristic (race, disability, gender, age, sexual orientation, religion or belief, pregnancy and maternity and gender reassignment) and those who do not and foster good relations between different groups. To this end, the Mayor should have particular regard to section 3 (above) of this report.

Revenue grants to the Authority

6.4.    Approval is sought for a £550,000 revenue grant from the Mayor’s Office for Policing and Crime (MOPAC) to the Greater London Authority pursuant to section 121(1) of the GLA Act. In accordance with section 121(3) of the Act, no conditions may be imposed upon the GLA’s use of the grant other than that the grant must not be used for revenue expenditure. Furthermore, in accordance with section 121(4) of the Act, the GLA must apply the grant towards meeting revenue expenditure for the purposes of or in connection with the discharge of its functions. As set out in paragraph 1.10, MOPAC has informed the Authority that it has sufficient funds to make the requested grant.

Procurement

6.5.    Approval is sought for the expenditure of £605,000 on the delivery of SafeStats and the ISTV programme across the next three financial years (2022-23 to 2024-25). All expenditure and services procured must be made in accordance with the requirements of the Authority’s Contracts and Funding Code. Furthermore, officers must ensure that appropriate contractual documentation be executed by both the Authority and any relevant contractors prior to the commencement of the required services. 

6.6.    Paragraph 1.9 above indicates that the MPS contribution of £50,000 to the GLA amounts to the provision of grant funding and not payment for services. Officers must ensure that the funding is distributed fairly, transparently, in accordance with the GLA’s equalities obligations and with the requirements of section 12 of the Authority’s Contracts and Funding Code.

6.7.    Furthermore, officers must ensure that an appropriate funding agreement is put in place between and executed by the GLA and the recipient before any (a) commitment to fund is made; and (b) funding is paid to the recipient.

7.1.    Both the SafeStats data portal and ISTV programme are already in existence and are fully operational. As such there are no additional procurement or delivery activities to note. The funding relevant to this decision will ensure continuity of service for both the SafeStats service and the ISTV programme, with development and expansion as required.