Sharing routine data for research - Data linkage through Connected Bradford 

City of Bradford Metropolitan District Council (CBMDC) contributes and uses data from the database, alongside the Department of Education, Yorkshire Ambulance Service and primary and secondary care organisations. Data shared has included: 

• individual, pseudonymised data from adult social care, children’s social care, education and the National Child Measurement Programme 

• non-identifiable personal data such as housing adaptations 

• non-personal data such as environmental aspects e.g. air quality sensor data 

By linking routine data collected by the council to healthcare and the wider determinants of health data, Connected Bradford can provide the council with valuable insights on the factors contributing to population health and inequalities. This new evidence can be used to inform council decisions and policy and to target provision to vulnerable and disadvantaged groups. External researchers can also apply to access data from Connected Bradford, increasing the opportunity for cross-sector research and collaboration.  

When NIHR Health Determinants Research Collaboration (HDRC) Bradford was launched in 2023, a new dedicated data management team was created. This provided an opportunity to increase data sharing across council departments and linkage with the Connected Bradford platform.  

Building capacity to share data

The aim of HDRC Bradford is to sustainably transform the council, so it becomes evidence-led and data driven, at all levels of decision-making, to reduce health inequalities. Over the last 18 months, the HDRC Data Manager and information governance colleagues have reviewed processes within the council around data sharing with the Connecting Bradford database and have implemented improvements to build confidence. These improvements have included: 

• the development of clear, documented processes for pseudonymising data 

• the introduction of a central log to record all data that has been shared 

• 6 monthly reviews with council officers who are sharing data to ensure ongoing review of the processes 

• proactively identifying additional data that could be shared 

• supporting Connected Bradford researchers to understand and maximise use of council data by providing expertise, links to subject matter experts and developing guides 

• developing training materials for council officers around the programming languages of R and Python to enable them to analyse data from the Connected Bradford database 

Keeping council data secure

A 4-step process is followed by council officers to ensure all personal data is kept secure during the sharing process: 

1. Data professionals at CBMBC collect data  

2. Identifiable information is removed at source by the data providers (the council team providing the data) so that personal information is protected.  

3. The de-identified dataset is securely transferred to trusted partners who link it with other datasets for research and log all data sharing activities 

4. Researchers use the data on secure software to spot trends, patterns, and connections to other information, findings from this research are shared with council decision-makers 

All officers are asked to complete a Data Protection Impact Assessment for the dataset they intend to add to Connected Bradford. This ensures there is no risk of a data breach and that all steps will be taken to ensure it is safe to share the data.  

All officers also follow a data extraction and sharing process that has been designed by the Bradford HDRC Data Manager and approved by CBMDC’s Data Protection Officer (DPO). This process enables each service to extract its routine data and delegate to an approved data officer to encrypt personal identifiers using the SALT key and OpenPseudononymiser tool. This data is then transferred to Kiteworks, a Private Data Network (PDN) to ensure the secure exchange of the final dataset. This is collected by the Connected Bradford Database Manager at BIHR for the data engineers to validate, before completing the data linkage process to its additional datasets, e.g. primary care.  

Governance and oversight

Researchers wishing to access the Connected Bradford database must complete an expression of interest application form. This is discussed the Connected Bradford Learning Health Systems Board comprising of representatives of all the organisations contributing data (including CBMDC) and the Connected Bradford team. The Board can choose to give the researcher access, deny the request or ask for additional information. If approved, researchers are given secure access to specific and limited data that is necessary for their research.  This is to ensure control over who has access and to ensure the data remains protected.  

Researchers access Connected Bradford via a Virtual Desk Environment (VDE) on a secure Council device. The CB VDE allows each researcher to access their approved datasets on the CB SQL database work in their preferred data environment and programming language, e.g. R or Python. Researchers are unable to extract or save any data or information from the VDE. Completed work is requested via the CB team. CB also has a GitHub site, where each dataset has a data dictionary, with researchers also encouraged to share any code to ensure projects can be replicated. 

Projects facilitated by data sharing: 

Reducing school absences 

Research combining council data with data sets on the Connected Bradford platform found a 38% increase in unauthorised school absences between 2012/2013 and 2018/2019 with an uneven distribution of these across the Eccleshill, Tong and Bowling and Barkerend wards. It also found that school absences in Bradford are highly concentrated with around 11.5% of students generating 80% of unauthorised absences. Connected Bradford researchers have worked with council officers to form a task and finish group to address factors that influence school absences. Outputs can be found on the N8 Research Partnership site here. 

Exploring development milestones and young people’s educational outcomes 

Through combining datasets, researchers have been able to explore the relationships between young people not reaching early development goals and those with a late diagnosis of autism and later risks of exclusion and becoming NEET (not in education, employment or training at ages 16-24). By identifying the potential impact of earlier intervention, the analysis provides decision-makers with important information about where resources can be used most effectively to best support young people.  

Improving air quality 

In September 2022, Bradford launched its Clean Air Zone (CAZ).  The CAZ aims to reduce pollutants by increasing the number of CAZ compliant cars, ultimately decreasing GP and hospital attendances for cardiorespiratory illnesses linked to poor air quality. 

The HDRC Bradford data science team analysed millions of records of vehicle data from CAZ traffic cameras. The team extracted and cleaned the data and used the programming language “R” to investigate patterns and detect noncompliant vehicles. Interactive dashboards were created, allowing the sustainability team to examine vehicle types and CAZ compliance status by time and location. 

The dashboards confirmed the hypothesis that Manchester Road had a high frequency of retrofitted buses contributing to poor air quality. This insight enabled the sustainability team to optimize CAZ compliance strategies and target interventions more effectively.  This analysis contributed to the Bradford CAZ attracting investment and £8m new zero emission bus funding through the Government’s NO2 Programme. 

More information 

For more information on HDRC Bradford, please see: Bradford Health Determinants Research Collaboration | Bradford Council 

For more information on the Connected Bradford database: Connected Bradford - Bradford Institute for Health Research 

For more information, please contact: Rob Shore, HDRC Data Manager on rob.shore@bradford.gov.uk 

Explanation of terms: 

• pseudonymised data = when personal details are removed from data, so an individual is no longer identifiable  

• SALT tool – a tool that generates random data to add to data or a password to make it more secure  

• R = software for statistical analysis 

• Python = programming language sometimes used for data analysis 

This learning story was prepared with support from NIHR RSS Specialist Centre for Public Health delivered by Newcastle University and Partners. With thanks to HDRC Bradford for providing the content.