Leveraging national health datasets to investigate changes to acute admissions of adults with type-2 diabetes.
The NHS Transformation Unit (TU) is an internal consultancy team that focuses on driving and facilitating transformational changes within the NHS. These transformations often involve designing and implementing new strategies, technologies, and processes to enhance patient care and outcomes.
The TU works primarily to improve the efficiency, effectiveness, and overall quality of healthcare services within the Midlands and Lancashire geography. Additionally, the TU frequently works with teams in other areas such as Cheshire & Merseyside, and actively participates in national healthcare programs.
The TU are experts in areas such as programme management, clinical service transformation, stakeholder engagement and system development and has developed its in-house analytical team to support digital and financial workstreams. Recently, much of the analysis performed by the TU has been focussed on supporting finance and business intelligence problems. Consequently, this has resulted in most of the analysis being undertaken in Excel or other proprietary software. Whilst there was some experience within the team of the application of programming languages and visualisation software, these were not routinely used.
A student from the University of Salford collaborated with the TU for an extended period during the concluding phase of their MSc in Data Science. This collaboration took the form of a Live Project which the student undertook to inform their final dissertation. Live Projects are typical of the MSc Data Science course at the University of Salford and allow students to develop their analytical and technical skills in an industrial setting whilst also contributing to their host organisation.
The project required the development of numerous tools to explore how admissions of adult diabetics to secondary care sites had changed both during and after the COVID-19 pandemic. Primarily, a pipeline was developed which pulled the most recent admissions data for diabetic adults before being imported into a bespoke Tableau dashboard. The dashboard presented several key performance metrics from 2017-2021 to allow users to explore changes in admissions, as well as patient demography and pathway utilisation.
Additionally, a series of predictive tools were developed using the R programming language. These included clustering patients into risk categories for several clinical outcomes of interest such as re-admission and their length of stay in hospital. Projections for future diabetic admissions were also made for the next five years using publicly available data on demographics and other health factors from the Office for National Statistics and UK Health Security Agency.
This innovative approach to addressing clinical questions marked a novel shift within the TU at that time, sparking a heightened interest in advanced programming and visualisation software solutions. The expertise acquired by the TU in this domain has significantly expanded over time, with high-level programming languages now routinely employed for tasks such as handling disease prevalence data requests. Additionally, the team’s proficiency in tools like Tableau and Microsoft PowerBI has experienced substantial growth. They consistently engage in geospatial mapping analysis, which has become a fundamental element in the TU’s strategy for gathering and visualising population health intelligence. Since the initiation of the project, the analytics team within the TU has seen significant progress in both skill and confidence and is actively involved in supporting numerous projects with their analytical expertise. This continual growth underscores the team’s integral role in contributing to the success of the TU’s various initiatives.