A game of snakes and ladders: the world of complex health and social care data linkage

Objective and ApproachResearch on social care services requires large, comprehensive, routine datasets. Launched in November 2019, Developing resources And minimum data set for Care Homes’ Adoption (DACHA) study aims to develop a prototype minimum dataset as proof of concept and to propose implementation. 
We describe our experience since 2021 of identifying, applying for, and linking care home, local, integrated care system (ICS) and national datasets, including direct-care software-provider data, GP and NHS England (NHSE) datasets. 
ResultsOur key challenges: 
 
Risk aversion and fragmented information governance within and between organisations. 
Complex information governance structures must be understood before participants are recruited. 
Shifts in the data landscape: data ownership moved from clinical commissioning groups to ICSs; NHSE merged with NHS Digital. 
Research funding: research is required of ICSs, but it is not a system priority, so processes and funding aren’t in place, and research burdens an already strained system. 
 
Our key lessons: 
 
Formalising agreements early on and obtaining, and maintaining, senior, committed buy-in. 
Studies should value the data sharing process through the inclusion of data controllers and information governance staff in the system in research project budgets. 
The facilitation of GP and social care data collection nationally must be balanced with its administrative burden. 
 
ImplicationsThe DACHA study overlapped with the COVID-19 pandemic and could only engage a limited number (n=3) of ICSs within project resources. Nevertheless, the complexity of the system and the plurality of actors delay or even block legitimate research interests.

We were not able to access some of the datasets we originally aimed to use, particularly GP data What makes up the MDS?
▪ GP patient data are held by individual GP practices ▪ Worked with three ICBs to explore potential data flows 1.One ICB were not able to prioritise the IG work due to competing priorities 2. In another, we had productive discussions on accessing risk stratification data collected from GPs, but were unable to align timescales for setting up data sharing agreements 3.In the third, we secured some data sharing agreements with GP practices by working through a CSUhowever the number of consented residents from these practices was too low to proceed while maintaining patient anonymity 1. Changes in the data landscape • CCGs -> ICSs.ICBs set up in 2022 after initiation of DACHA.
• Merger of NHSD and NHSE, restructures • Impact of pandemicrapid set up of linkages at all levels 2. Risk aversion DACHA participants provided explicit consent for their data to be linked and shared.
Despite this, NHSE assessed at the highest level of risk and the DPIA required Caldicott Guardian review.
We think the merger of NHSD and NHSE meant that processes were reviewed and led to a different attitude to risk.
NHSE are committed to supporting research but did not prioritise this.IG was not set up to respond to external research.
Key contacts in one organization did not necessarily trust or accept decision making made in other parts of the system.We encountered different views on legal bases for data sharing from different organizations (for example pushback on legal bases for data flows from DPOs for GP practices in one ICS).
Ideally we would get signoff from all data controllers on legal bases used in e.g.data sharing agreements with care homes at the start.

Research funding frameworks
• DACHA required significant work from busy teams

Conclusions
▪ Resource demand for building an MDS must be considered to make sure it can be most useful and adaptable (high quality, timely, responsive) ▪ The current data landscape raises 4 key challenges to data linkage ▪ Lessons learned 1) Formalising agreements early on and obtaining, and maintaining, senior, committed buy-in.
2) Studies should value the data sharing process through the inclusion of data controllers and information governance staff in the system in research project budgets.
3) The facilitation of GP and social care data collection nationally must be balanced with its administrative burden.

Contents 1 .▪
The DACHA project: why do we need a minimum dataset?snakes and ladders: the world of complex health and social care data linkage 4 DACHA (Developing resources And minimum data set for Care Homes' Adoption) started in November 2019 and is divided into five work packages with PPIE throughout Objectives: Identify data required for a minimum dataset, through review of literature and engagement with stakeholders ▪ Link data sources to create a proof-of-concept minimum data set (MDS) that is usable and informative for different user groups ▪ Collect quality of life (QoL) data in care homes and link to MDS, and assess value added of QoL in informing local picture including for better care for residents

3.
Legal bases for data sharing DACHA MDS for Direct Care vs Service Evaluation vs Research Our prototype MDS was developed for Research A real MDS would be used for all three purposes Linkages at ICSs set up for specific purposes, not flexible NHSE did not have a legal basis to hold the care home data