Joint Project Helps Combat Fraud
Effective fraud prevention and detection within the remit of procurement and commissioning is an issue of concern across the whole of the NHS. Last year, the NHS Counter Fraud Authority (NHSCFA) estimated the fraud loss within NHS procurement and commissioning to be approximately £351million annually - the largest fraud risk identified within the NHS. To help tackle this, NHSCFA and NHS SBS have worked together to build and develop new data models to detect and assess fraud.
To help tackle this, NHSCFA and NHS SBS have worked together to build and develop new data models to detect and assess the outliers within which fraudulent transactions can be found, and review the controls that can prevent them.
The work began in 2017, when NHSCFA's Information Analytics team reviewed invoices drawn from two years' worth of transactional data from two NHS providers in the South West. This was the first time NHSCFA had access to large scale procurement data to examine within a fraud context, which was used to develop complex data models that could detect the outliers within which fraudulent behaviour could be found.
Following the success of this proof of concept, it was necessary to identify an appropriate next step to develop and build on these models, whilst also adapting them to alternative data sources and widening the scope of both the analysis and the data itself.
NHSCFA and NHS SBS agreed that, although there was vast amounts of data available, it would be a risk to attempt to extrapolate the data models too widely and that the temptation to 'go big' in terms of accessing as much data as possible would have carried a high risk of failure. Despite the "big data" buzzword of the modern age, the risk of seeking datasets too vast and varied to explore can result in having too little understanding of the meaning behind them, making engagement difficult and in terms of setting an appropriate scope - see https://cfa.nhs.uk/about-nhscfa/latest-news/Big-data-and-the-benefits-of-starting-small for more information.
With guidance and advice from NHS SBS, NHSCFA identified 18 NHS providers with whom it would provide a natural and logical expansion of the data models they had drawn from the pilot and sought to seek access to their invoice data.
Accordingly, NHSCFA engaged with the Director of Finance for each organisation, clarifying their intention to perform a proactive exercise and seeking authorisation to access their data. The response was extremely positive and NHSCFA was swiftly able to begin exploration and analysis.
The findings of the analysis were encouraging - concluding that the system and process controls in place at NHS SBS can be considered robust, as only a small number of outliers were identified. There was also evidence of correction and validation of data that indicated the validation was working as intended.
Where the analysis returned results that may warrant further investigation - for example where results indicated areas of potential weakness that could be facilitating bypassing of the current controls - NHSCFA and NHS SBS will pursue further joint work. A key recommendation is that further investigation is required in order to validate them or consider if and how these may be considered weak to exploitation. Additionally, NHSCFA will seek to expand their data models further by acquiring additional data in order to carry out further analysis. Modern techniques such as using data mining to proactively identify anomalous transactions or patterns of behaviour can also be used.
David Dixon, Analytical Intelligence Lead for the NHSCFA said, "Although in many respects this exercise was about the gradual expansion of data models from our proof of concept, this was an important piece of work in and of itself. It has been helpful in terms of developing our analysis, expanding our understanding of invoicing and guiding our next steps for the future. Many people assume data analytics starts with taking on vast amounts of data, but the truth is that it's always important to start small and build on it - and this has proved a very strong stepping stone to our next steps and wider objectives to combat fraud across the NHS."
Stephen Sutcliffe, director of finance and accounting at NHS Shared Business Services added: "Fraud is a high-profile and damaging problem in the NHS - drawing valuable resources away from patient care.
"Working with organisations like NHSCFA is just one way in which we're helping to tackle this across the whole of the NHS, and provide additional assurance to all the organisations we work with. We're particularly pleased by the feedback we're received from NHSCFA on our processes and controls, and look forward to building on this work in the future."
NHSCFA and NHS SBS would like to thank the following NHS organisations who agreed to provide their data for this exercise:
- Bradford District Care NHS FT
- Bristol Community Health
- Central and North West London NHS FT
- Derbyshire Community Health Services NHS FT
- Derbyshire Healthcare NHS FT
- East Cheshire NHS Trust
- East London NHS FT
- Kent & Medway NHS Social Care Partnership Trust
- Leeds and York Partnership NHS FT
- Leeds Community Healthcare NHS Trust
- Medway Community Healthcare
- North Somerset Community Partnership
- North West Boroughs NHS FT
- Solent NHS Trust
- Staffordshire and Stoke-on-Trent Partnership NHS Trust
- Surrey and Borders Partnership NHS FT
- Tavistock & Portman NHS FT
- Worcestershire Health & Care NHS Trust