A crisis can shine a spotlight on whether governments are mobilising data effectively, write Sir Geoff Mulgan and Oliver Marsh of the International Public Policy Observatory
Mobilising data is essential for government decision making, both in normal times and during a crisis. In our research report, Navigating the Crisis, we examined how governments across the world used various forms of intelligence to make decisions during the Covid pandemic.
We saw that the crisis catalysed widespread innovation in uses of data, but it also revealed obstacles and raised questions that should be addressed well in advance of such emergencies.
Responding to Covid required mobilising data across many domains. This obviously involved health data, including from tests on citizens, vaccination centres, reports from hospitals and other types such as on wastewater or serological testing.
Governments used a variety of data to monitor changes in behaviour, from financial transactions to mobility data from phones. They also needed data on citizen sentiments and wellbeing, collected through surveys, digital technologies and more qualitative and narrative methods such as local conversations and forums.
Deployment of tools
Governments deployed a variety of tools to store and manage this data. Some built on existing assets, such as NHS patient data stores in the UK or local databases of poverty held by regional governments in Bangladesh. Many benefitted from open source tools such as the DIVOC platform for issuing vaccination certificates or the District Health Information Software 2 (DHIS2) platform for managing health data.
The NHS integrated Covid health data into a Covid-19 Data Store to track the spread of the virus, allocate resources and patients, and support clinical research. According to the NHS Transformation Directorate website this created “a central data store would have taken years under normal circumstances”.
Digitisation and automation played an important role. A UK official said: “The ability to have data coming in in a timely and machine readable way, and to be able to ‘turn the handle’ and make slide packs” was valuable for efficiently briefing decision makers.”
Bangladesh had been digitising records, particularly around poverty, in the cause of sustainable development, but the need to understand the impacts of Covid upon the poorest citizens accelerated digitisation.
In various countries, including the UK and South Korea, natural language processing technologies – which algorithmically draw out trends from large quantities of text – were used to summarise qualitative views from citizens collected from telephone hotlines, online forms, social media and other sources.
Obstacles and challenges
However, there were limits to technical methods for data sharing. For instance, digitisation encountered issues of standardisation and interoperability.
This was particularly strong in countries with federal systems such as Australia, Switzerland and Germany, where different regions had varying approaches and standards that could not easily interface with one another. One interviewee said the Berlin Senate sometimes struggled to fully synthesise data from health centres and schools in different districts within the city.
Centralised systems, including in the UK, also faced problems with standardisation. One research and regulatory body told us that its pre-Covid steps of digitising processes had helped it to rapidly respond to requests from governments. But it felt this could have been improved by integration across, rather than simply within, regulators – which began later in the pandemic but required “a lot of retrofitting”.
Systems like those of the UK arguably have the advantage of a central authority to define standards; but the discussions should take place and issues stress tested in ‘peacetime’ rather than crisis.
Legal issues
Questions also arose around legally permissible data sharing arrangements. Our interviewees did not raise these as a substantial obstacle, but indicated that they certainly played a role.
One from Germany raised the General Data Protection Regulation as a limit to what economic data could be shared. The New Zealand government reported that a learning from the experience of Covid was that “it would be beneficial if good information sharing agreements were in place ahead of time, or as an early priority”.
The widespread use of contact tracing apps led to debates over whether such data should be stored centrally by governments (raising privacy concerns), or whether to use decentralised approaches through technologies such as the Google/Apple Exposure Notification (GAEN) System, which was eventually used in contact tracing apps in over 20 countries.
An intriguing alternative was Taiwan’s digital fencing system, which monitored locations based on triangulating a phone’s position relative to nearby telephone masts. Monitoring was conducted by telecoms companies, based on a list of phone numbers of quarantining individuals provided by the government.
Digital Minister Audrey Tang argued that the digital fencing approach preserved privacy by re-using data, as phones “already have its signal strength checked by the nearby telecom towers anyway”, plus a constitutional limit of 14 days on tracking an individual.
Beyond legal requirements, public consent can be vital for maximising the benefits of innovations. Work by US not-for-profit consultancy The GovLab found that innovative data collection initiatives were often implemented without the necessary social licence to do so, which led to public concerns and initiatives being discontinued.
Capabilities and partnerships
Innovative technologies, backed up by shared standards and legal bases, do not guarantee effective mobilisation of data. It also needs the right personal capabilities and interpersonal relationships of people within government.
Digitisation and automation of data for decision making requires an understanding of what the end result should look like, which requires a skilled liaison between officials to iteratively develop the most effective methods of presenting data.
However, we argue that too often data reached these teams through chains siloed between different departments and domains. During Covid this often took the form of different aspects of the crisis – health, economy, wellbeing – appearing to be in opposition, with health related data and expertise often being prioritised.
We argue that capabilities which cut across domains and are skilled at synthesising multiple forms of intelligence for decision making should be more widespread throughout government, allowing for more effective synthesis for decision making.
Crisis dynamics
Effective data sharing amongst government personnel also requires understanding the dynamics of working in a crisis. Multiple interviewees noted how the initial experience may have spurred improvements in working relationships.
South Korea had, following the experience of MERS in 2015, put in place the Infectious Disease Control and Prevention Act to encourage collaboration between departments, and Covid galvanised cooperation in the manner anticipated by the original law.
But as the crisis wore on many countries – the UK included - there was growing competition between various camps: whether health focused versus economy focused departments, or different local areas.
Effective data sharing therefore requires building strong, trusting personal relationships, which can resist and persist even in the face of a crisis. These can be inculcated through interpersonal engagement, ranging from crisis simulations to shared training centres, which bring together multiple departments, tiers of government and external partners.
It is vital that governments facilitate such relationship building well in advance of future crises.
The International Public Policy Observatory is hosted by University College London. The full Navigating the Crisis report is available from here.
