Royal Statistical Society, 12 Errol Street, London, EC1Y 8LX
29th May 2014
– Welcome by Simon Burall (Involve) and summary of the process so far, and what it aims to achieve.
– Simon outlined that the Cabinet Office ideally wants civil society to come up with concrete proposals to ensure that data sharing has the appropriate safeguards. However, this process is not a juggernaut to data sharing. It may be that we don’t agree on the circumstances where it should be done and how. If this is the case, it will be a good outcome to identify where disagreement lies.
– Peter Lawrence (Cabinet Office) outlined the three different strands of activity: Research and statistics (ONS and de-identified data for research; Fraud error and debt; Identifiable data to improve services for individuals.
– He outlined the timeline of the process and the key milestones anticipated along the way. These include a draft policy proposal paper, signed off by government and civil society by mid September, a White Paper before Christmas, then a period of wider consultation until the end of March 2015. He stressed the commitment of Francis Maude to the process.
Updates from strands
Each of the three strands gave brief updates on the work that they had been undertaking.
ONS strand (research and statistics)
– The ONS is seeking easier, speedier and more flexible access to de-identified administrative data, so it can reuse it for statistical purposes only. They have been undertaking workshops with CSOs and moving forward with departments who own the data, privacy and impact assessments around sharing the data. One of the issues is that the process for approving data sharing currently requires affirmative resolution in both houses, and there is a concern that this is not an effective use of parliamentary time – and an effort to find a more effective alternative system. This involves not only the ONS but the whole of government statistical services.
– The difficulty that academics have in getting access to administrative data. There are legal barriers for some departments to sharing data, and the effort to find a common approach and apply standards with greater consistency. The problem of civil society participation and the need for greater engagement was highlighted.
Fraud, error and debt
– The challenge is for this group to agree a proportionate way of using data sharing to address fraud, error and debt issues. There has been strong engagement but a challenge of continuity of participants.
– There are a range of estimates on the scale of the problem from different bodies, ranging from £47 – 68 billion. There is a greater need for clarity on the sources of these figures.
– Data sharing happens already, but the process is cumbersome and can involve numerous bilateral bodies.
– Public attitudes to data sharing are segmented, with younger people more at ease. There is frustration when information isn’t shared across public bodies in what the public perceive as ‘the same’ problem. The need for more systematic approach to gauging public opinion, using more rigorous methodologies.
– Many of the issues from fraud error and debt strand have been around what the evidence base was for the need to increase data sharing; whether the existing gateway system is adequate and detailed evidence on how the system works. There was a request for clearer policy papers on where this stands.
Tailored public services for individuals
– This strand involves using shared data to tailor public services.
– Particular application might be on the unemployment and skills agenda for local authorities. David Coleman explained the potential applications for his work in Southend Borough Council. A single data sharing process would reduce the need for separate data sharing agreements between departments.
– There have been discussions about what other kinds of services could be tailored – particularly for carers, vulnerable people and older people. There is a question of scope and where best to draw lines.
Small group work
Participants were asked to break into three groups, one for each strand, and answer:
– Where is there currently consensus?
– Where is there disagreement?
– What would a useful next step be?
The three groups then reported back to plenary
Fraud, error and debt
– There is agreement that there’s an issue with fraud, error and debt that should be tackled.
– There is disagreement on whether data sharing is the answer. There is a need to gather more evidence on levels of fraud, error and debt, and for more systematic pilot studies on the effectiveness of data sharing.
– The next steps is to go back to original proposals of Cabinet Office to see where government thinks it might want to get to, and then to think more about what types of information might need to be shared. Is it just financial information, or designatory information (e.g. names and addresses) etc.
– Need to explore what kind of evidence is required for firm decisions to be taken.
Research and statistics
– There is a shared commitment to explore whether a ‘gold standard’ for data sharing safeguards might exist.
– There is agreement that de-identified data is less of an issue. Have to make sure that what is a new power doesn’t weaken everything else that exists regarding safeguards
– There is disagreement on questions of scope, and how far to enlarge the reach of the safeguards.
– Next step is to share work and develop the safeguards (with way of looking back to see the proposal is better than what we have already).
Tailored public services for individuals
– There was agreement that it wasn’t good practice that each time a new problem arises it is necessary to go to parliament to develop a new gateway.
– Agreement that there is a need to streamline the process of data sharing, and that this should be some kind of generic process (not necessarily a gateway).
– Agreement that parliament doesn’t necessarily provide optimum scrutiny and there may be stronger and more appropriate alternatives.
– Need to know what the purposes for which this generic processes should be used (there is currently disagreement on this). Then once know the purpose, the next step will be to look at powers and then move on to safeguards.
– Need more information on how existing gateways are created, where are the hurdles, and what ‘powers’ means in a practical sense.
Simon Burall identified common themes between the groups:
– The concern with how to future proof proposals
– How to make it flexible enough to be used for future events but not too vague.
– The need for pilots. How might we pilot and test proposals, and what would the pilot need to tell us to generate confidence to move forward?
– There are a number of issues that aren’t related to safeguards or evidence but to wider political issues. For example the role of the private sector. This is a part of the conversation that also needs to be brought out.
Problems and solutions
In small groups, participants identified problems and solutions with the process so far. These were then reported back in plenary:
– Lack of a forecast schedule of meetings there’s a continuity problem with attendees.
– Very London centric.
– Need for a pre-pack of information so everyone can be on the same page when arrive.
– Make it a longer day (a full day) to maximise value of travel.
– Output should we summarise what we came up with we can share the evolution of process with our colleagues and communities.
– Leave more time between meetings so people have the opportunity to assimilate materials.
– An agreement at end of each meeting that we won’t go over again at that meeting. A ground rule that don’t go over old ground.
– Some sort of template that demonstrates where we’ve got to – an agreed premise of what that strand was trying to do. A clear way to managing it – template that people could comment on, and that would set out where the next level of question is.
– Meetings too short and ambitious, too unwieldy. Solution is better structures to meetings and clearer objectives to meetings.
– Larger meetings, roving microphones.
– Make use of online methods, live streaming? Clear that we could do this but important that this has an audience and isn’t into the ether.
– For next session helpful to deep dive on one theme and work through and process the barriers rather than being so broad.
– Civil society not fully represented. Not much spectrum of understanding. Who can be the link people to translate the data people to the nuts and bolts people? Participants were asked to link us to people who should be invited to participate.
– Need to open up to all different perspectives – not just all one type of people. Need people in the policy areas who aren’t data people but whose issues will be affected by data.
– Private sector gap.
– Question of what the end point necessarily is? What is a white paper? Should this be assumed?
– The need for more frequent outputs and involve other orgs – blog on datasharing.org.uk site. Only one public document on datasharing.org.uk
– Datasharing.org.uk presumes the destination where we are going to get to.
– Need for consultation on the format of the day.
Simon Burall summarised the discussion. He reflected the need for a longer, all day meeting with more preparation up front, greater clarity on the objective and what comes out at the end. He summarised the desire for the Cabinet Office to communicate more clearly before and after meetings.
He noted the caution of sensitivities around private sector participants, alongside the need to ensure this group was appropriately involved at the right stage. He appealed for participants to inform himself and Tim Hughes of people who should be invited into the room.