Addressing the elephant in the room

Like many areas of the public sector, Essex Partners are exploring how we can share and use the data we have to improve the services we deliver. As part of the Essex Data programme we are sharing personal pseudonymised data, between partner organisations, to try and understand some of our most challenging issues and develop a place based approach to predicting risk to enable early intervention.

It won’t surprise you that because we are sharing personal information and because we are looking at sensitive issues relating to vulnerable groups, our stakeholders have been consistently focused on doing the right thing ethically.

Let me give you an example. If we knew where those at high risk from experiencing domestic abuse were, before it occurred, we could do something to reduce the risk. But we are talking about risk, not actuality, so might we be pre-empting and labelling, and affecting an outcome that would otherwise not have occurred? What if we don’t share the data and therefore don’t prevent a negative outcome that could have been foreseen - is this ethical?

A research ethics process - ensuring that ethical considerations of research projects are given appropriate attention - is not new for us but applying it in the context of data science and predictive analytics is. We were able to use the Cabinet Office Data Science Ethical Framework to address this new challenging area.

The quick checklist that is part of the Cabinet Office Framework has been particularly useful. So far we have used it to help us analyse the ethical issues around our Domestic Abuse prototype and are in the process of using it to support the development of our Gangs, Violence and Vulnerability prototype. We have used the checklist to facilitate a structured conversation with key stakeholders, finding the questions, on the whole, easy to interpret and the scale response a useful prompt to explore risky issues.

The output generated from these conversations has been used to demonstrate that the right people have had the right conversation and to build appropriate modifications into the project approach. For example in answering the question 'what is the quality of the data?' we explored issues around potential bias which led us to further strengthen our approach to triangulate the insight from the data with other sources of information in producing the analytical output for decision makers to use.

The Framework has offered us a practical tool, helping us understand the ethical issues at play and enabling us to ensure they are managed effectively. However more broadly it has helped us build confidence in the Essex Data programme by providing reassurance that innovation won't compromise ethics. Because our work is new and pushing boundaries it is an essential part of the programme to enable a culture change around how people perceive data projects and how willing they are to share data.

Liz Ridler

Deliver and Evaluation Lead, Essex Partners

4 December 2017