By Aaron Atkinson, Associate Director of Business Intelligence, Helen Duckworth, Director of Business Intelligence Transformation, and Jess Hicks, Associate Director of Advanced Analytics, at NHS Arden & GEM.
Data, when used effectively, is perhaps the most crucial tool available to help NHS organisations deliver the Triple Aim duty of improved population health, improved service quality and improved value.
The complexities of information governance and other necessary hurdles that must be overcome to enable data sharing can make system-wide data projects feel unattainable. Yet, we are working with an increasing number of Integrated Care Systems (ICSs) that are seizing the opportunity, making positive improvements to patient outcomes and system resilience through data-led transformation programmes. So how and why should systems prioritise data analysis, and what criteria are essential for success?
The reason for using data to inform decision-making is clear. We cannot afford to make commissioning choices which are not built on good evidence – we need confidence we’re supporting the right communities with appropriate solutions. By linking disparate datasets and using population health management methods such as segmentation and predictive risk modelling, we can use data to gain a much clearer insight into the needs of our populations and create much more targeted interventions. The data also enables us to model likely outcomes based on different scenarios, supporting both commissioning and service delivery. This is not quite crystal ball territory, but it does give ICSs sufficient insight to make and monitor decisions.
The ‘why’ is easier than the ‘how’, of course, but this is where it’s getting exciting. We are seeing real progress with ICSs sharing and utilising data across partners. This includes the first example of an ICS enabling a voluntary sector provider to access linked patient data to provide proactive patient care. Bedfordshire, Luton and Milton Keynes (BLMK) ICS is using data to identify high intensity users of accident and emergency (A&E) services and put in place interventions such as social prescribing to reduce that usage. The ICS wanted to develop a truly cyclical pathway which would intervene to reduce the risk of admittance, as well as providing supported discharge from hospital for those who are admitted. Arden & GEM worked in partnership with the ICS to develop a High-Intensity Users Outcomes Report, drawing on data from various sources including primary care systems, secondary care, daily discharge datasets, community services, mental health, health and social care, and hospice care.
Crucially, Turning Point, a voluntary organisation providing local health and wellbeing support in Luton, has access to the same data as the community discharge teams and other partners, and is responsible for running reports, within the secure data environment, to identify patients in need of social prescribing and wellbeing support, as well as providing daily or next day updates on their activity to inform the programme’s effectiveness. In line with information governance requirements, access to the various elements of the report is managed using role-based access control, which was developed in close liaison with data protection officers, IT and digital leads, and leaders from participating organisations. Although there were challenges to overcome, the ICS is now reducing high-intensity use of A&E and improving patient experience by providing targeted support and timely interventions within the community. This is a promising example of how systems can work together to make meaningful changes that deliver much more joined up care across service providers.
Mid and South Essex ICS has also invested in business intelligence and data analytics to support care planning. The ICS identified the need for a system-wide approach to capturing, analysing and using data through a collaborative data platform. Working alongside Arden & GEM, the new platform enables Mid and South Essex ICS to combine, analyse and interpret national and local data from multiple sources, all within a single platform. Data is presented to clinicians and commissioners through intuitive dashboards that allow the data to be used effectively, with information governance arrangements enabling safe data sharing to support system-wide working. Initial dashboards focus on urgent and emergency care, population health management, and diabetes, cancer and long-term care, based on priorities identified during the platform’s development.
Building confidence in data sharing
Recent research suggests that the public are largely supportive of the use of data to support healthcare. According to the Public Attitudes to Data and AI Tracker Survey, 81% respondents said they would be happy to share personal data with the NHS to develop new healthcare treatments, while 62% would be happy to do so with government to deliver public services. Enabling data sharing is simpler now that ICBs are custodians of system data, with the freedom to sub-license data to system partners with the appropriate safeguards in place. The Triple Aim also provides a clear incentive to use population health to deliver better patient outcomes and more cost efficient services. However, since the removal of temporary allowances granted during the COVID-19 pandemic, NHS organisations are naturally cautious about data sharing. So how can this be overcome?
Key to building confidence is comprehensive engagement with stakeholders - something that has been common to both examples cited above. This is as much about developing the business case for data-led transformation as it is about understanding your end users' needs. Engagement with stakeholders is essential in capturing the right data as well as understanding how it needs to be presented and what dashboard tools will enable users to gain valuable insights. Mid and South Essex ICS, for example, engaged with over 70 people in developing their platform, including digital and data leads from system partners. A data summit enabled stakeholders to discuss their individual and shared data and intelligence needs and share priorities that fed into the platform development. This collaborative working is built into the ICS’s way of working, with monthly BI and Data Delivery Board meetings supporting ongoing development of the platform. Similarly, BLMK involved stakeholders, including patients, in co-designing, planning and implementing their project.
Often when we look at making better use of data, we are also talking about shifting the culture of an organisation. Although system working has been in place for some time, traditional organisational structures and working methods remain. Organisations within systems will often be at a different place on their data journey, and the quality of data and IT systems varies significantly. Against this backdrop, data-led projects need to consider the varying levels of knowledge and – more importantly – capacity to take on change. In our experience, for example, delivering new data projects in primary care is often best done in incremental phases, recognising resource pressures and the multiple systems in use. This is about ensuring clinicians and administrative staff fully understand the benefits to their patients and their organisation and are equipped to use the tools available to them to make meaningful improvements in patient care.
Data sharing must continue to go hand in hand with data security. Our communities expect us to keep their data safe, and the stringent Information Governance Toolkit all organisations must regularly maintain is a key part of that. Building secure data platforms with role-based access and well-managed authentication is enabling systems to keep a much tighter rein on their data, while sharing appropriate information to support service delivery.
Artificial intelligence (AI) and machine learning
We are starting to see recent improvements in data and analytics advance even further with use of machine learning and AI. Working with BLMK ICS, we have developed a case finding tool which is helping their practices and Primary Care Networks to identify patients who may be suitable for social prescribing support. The model uses machine learning to calculate the likelihood that an individual would be referred to social prescribing based on historical patterns in social prescribing referrals and uses this data to identify and prioritise new patients who may benefit from a referral.
There are also growing examples of AI being used to support clinical decision-making and it has the potential to enhance how we use population health management to provide effective preventative care. But just like other data improvements, these developments will need to be supported with effective engagement – specifically with patients – and safeguards that build confidence and change.
Data and analytics is one of the fastest-moving areas in the NHS. The COVID-19 pandemic demonstrated the value of effective data, both in protecting the most vulnerable and in rolling out the vaccination programme. How we apply that learning to preventative care will be critical in achieving successful delivery of the Triple Aim.
This article was originally published in National Health Executive.