ICHP collaborating with The World Bank on Big Data and AI

Andi OrlowskiAndi event image, ICHP’s Deputy Director of Business Intelligence, is proudly part of the World Bank Group’s event in Poland sharing his expertise on an international panel discussing the potential of big data and artificial intelligence (AI) in the health sector.  

Being part of this event is a great opportunity for me to represent us and our partners in the global discussion on realising the full potential of big data and AI in the health sector. This event organised, by the World Bank Group in Poland, is focussed on the potential for big data and AI to fundamentally change healthcare. The aims in Poland are much the same as they are in the UK – preventing disease, monitoring the safety of healthcare systems, and supporting the implementation of improved care pathways. They, like us, are striving to grow the conversation on how data can underpin these aims.

For example, to create big data sets we must obtain data that we can utilise in a number of ways, such as linking it at identifiable patient level for direct care or using it on an aggregated level for planning purposes. Key to this is information governance, and the need to store and utilise data in ways that maintain the public’s trust.

Our current big data/AI projects

Here at ICHP, we’re currently working on the feasibility of a flu/flu vaccination data tool that we hope will help predict the numbers of people with flu each year and the subsequent demand on primary and secondary care. The tool describes where people have had flu jabs and where the remaining opportunity lies, so that public health and the NHS can prioritise and reallocate resources for further vaccinations.

The tool looks at communities, so we can see geographically where people haven’t taken up the opportunity to have jabs and tweak the promotional and communication efforts to target local populations. To do this, we pull together data from primary, secondary, community and social care as well as mental health, local authority, Office for National Statistics, Google and the Met Office data to create ‘big data’ to analyse any correlations.

In a separate project, we’re working on a ‘bow-tie’ analysis using national hospital data to map variation in care pathways and outcomes. This involves selecting a ‘knot’ for the bow-tie – e.g. a key event like sepsis or an atrial fibrillation-related stroke – and analysing national secondary care data to show care pathways leading up to the key event (the knot) and then the following time period. We partner with a French academic organisation to use machine learning and data mining techniques to analyse these pathways to show where interventions have been missed (the gaps in care) and compare different pathways and how they are related to outcomes.

We are also working to validate predictive models created by AI and neural networks. For example, we are testing an AI tool that claims to be over 90 per cent predictive of an unplanned attendance/admission to hospital, as well as predictive tools that claim they can identify atrial fibrillation from primary care and secondary care data.

Opportunities moving forwards

As part of the event, I’d like to highlight that while much of the AI and machine learning is focused on care, it could also be used to support the running of the health service, including aspects such as the supply chain, employment, staff retention and procurement.

I firmly believe that the key is partnership working; that by being able to come together at events like these we can work together to target areas where we can all make a measurable impact throughout our healthcare systems.