About the project
Respiratory Disease detection and prediction
Respiratory diseases such as COVID, Influenza and COPD are deadly. Yet, the true number and location of respiratory disease cases often remains hidden from official statistics. This is because such disease disproportionally hospitalizes vulnerable demographics across the UK, with those able to self-medicate simply doing so – and getting on with their lives. Because true incidence unrecorded, decision makers often have to act in the dark, outbreaks cannot be detected ahead of time, and valuable health resources can be incorrectly allocated. The UKRI’s CIVIC program aims to address this.
The CIVIC Program & Shopping Data
CIVIC a first-of-its-kind partnership between Leading UK academics, the NHS, and major UK Health Retailers committed to transforming UK resilience. CIVIC is based on a single observation – that hidden disease incidence is detectable, even at community levels, through simple signals such as over-the-counter medication sales (e.g. cough mixture, decongestant and pain relief purchases). Working with over 1.5 billion medication sales logs aggregated at community levels, CIVIC asks the following key questions:
- ◼ To what extent can we detect unrecorded cases of Respiratory disease across UK neighbourhoods using AI to analyse sales data?
- ◼ Can we predict disease outbreaks ahead of time using medication sales data, in order to support early-warning systems at scale?
- ◼ What insights can such datasets give us on the impact of COVID to vulnerable, potentially hidden, communities (e.g. food poverty) to help long-term intervention strategies?
- ◼ By working with consenting individuals via “data donation”, can we identify antecedents Respiratory illness in their purchasing trajectories?
◼ CIVIC’s AI models (built in partnership with NHS-X/Boots) accurately predicted Respiratory Mortality over 17 days in advance every week between 2016-2020 in every on of the UK’s 314 local authorities (R2=0.78***). These results reflect significant advances in fidelity and preparation time (NHS). 
◼ When applied to COVID-19 (2020-21), expected to be far more challenging, CIVIC’s XGBoost models accurately predicted mortality over 21 days in advance (R2=0.71***) again at local-area levels – far outperforming models based on official case data alone (R2=0.44**) as used in traditional epidemiological models. 
◼ Work undertaken with both OLIO (Food Sharing app) and Co-op Ltd, have shown the ability to examine impact to vulnerable communities of outbreaks at neighbourhood levels. Interfaces being piloted with Havering County Council, London, and expanded to other councils through support from Guy and St. Thomas.