Month: July 2020

The Wight way or the wrong way

This is a piece I contributed to a book on the Coronavirus pandemic, called Pandemic Where Did We Go Wrong? It is very slightly rewritten to suit the web, for example some footnotes have turned into links and others into text. The brief was to provide a summary of what happened to England’s contact tracing app for “second year journalism students”. It is deliberately non-technical and non-opinionated.

Intro from the book

The UK was not ready for Coronavirus. The UK needed to build a test and trace capability suitable for the scale of the pandemic. A key part of the Government’s strategy, a contact tracing app, broke on the Isle of Wight.Peter Wells explains why…

The need for contact tracing

As the pandemic struck it became clear that the UK had existing public health capability but not to the required scale. It would need a massive programme of work to strengthen it. Public health is a devolved matter, all four national administrations and local public health authorities across the UK started work.

Contact tracing is part of the public health response to any infectious disease. Expert contact tracers interview infected people to discover to whom they might have transmitted a disease , they then interview those contacts to discover if they might have the disease and to whom they could have transmitted it, and so on. The contact tracer uses their expertise to recommend what activity each possible contact should take.

Research into the Coronavirus showed that people became infectious before they showed any symptoms. Some researchers became concerned that this would lead to traditional contact tracing methods being ineffective. The disease would spread before contacts could be traced. This would lengthen the lockdown, cause more economic damage, and cost lives.

A group of scientists in Oxford led by Professor Christophe Fraser, the Coronavirus-Fraser group, started to explore whether technology, in the form of a mobile phone contact tracing app, could help.

The rise of the NHSX app

The app would use Bluetooth signals to detect which phones had been in contact with each other. If a user told the app on their phone that they had tested positive for Coronavirus, other app users who might be at risk would be informed and advised to take action. The methodology had previously been explored in theory, but had not been used in an actual pandemic response anywhere in the world.

In early March 2020 Professor Fraser met with Matthew Gould, the CEO of NHSX – a unit of the Department of Health and Social responsible for digital transformation – to explore the idea. The Government agreed that the scientists would develop the theoretical framework and models, while NHSX would work with partners to start developing an app.

In late March Matt Hancock, the Secretary of State for Health and Social Care, announced that NHSX had started developing a contact tracing app.

It quickly became clear that other people had views.

A group of technologists, including myself, asked the government to be wary of trusted untested technologies and recommended that it act transparently. The government said that it would act transparently and that people would be able to trust the app.

European academics specialising in  privacy were concerned that an app which traced people’s contacts might reveal too much information about citizens to their governments. They started developing their own approach to digital contact tracing, known as DP-3T. This approach was termed ‘decentralised’ and ‘privacy-preserving’, in contrast to the UK’s approach which was framed as ‘centralised’ and as endangering privacy. They tried to persuade the public, European governments and the EU to support their approach.

Google and Apple, who make the operating systems for the vast majority of smartphones in use, announced that they supported this de-centralised, privacy-preserving approach and would release their own protocol that would make it easier for governments to build apps. They said that they would not support centralized apps, like the one that NHSX had started developing.

The Ada Lovelace Institute, an independent research body, performed a rapid evidence review and found “an absence of evidence to support the immediate national deployment of the technical solutions under consideration”.

The Coronavirus-Fraser group’s mathematical models showed that if the app tracked the disease well enough and the people using the app behaved as expected then the spread of the virus would be reduced, but it was not clear what would happen in the real world. Mathematical models help us understand what might happen. They do not tell us what will happen. The Government continued on their course.

Updates were given to the national press but despite the promises of transparency detailed information was not shared, whether with the general public or with experts outside the Government. In the absence of information public debate continued, often fuelled by social media speculation about who was building the app and what their motives were.  

In late April 2020 NHSX told the BBC that they were running a limited trial on an RAF base in Yorkshire. No results from this trial were published, but the Government concluded that a bigger trial was needed to understand how the app might work in the real world.

The Isle of Wight trial

In an interview Dave Stewart, the leader of Isle of Wight council, suggested that the Isle of Wight could be used for a trial of the app. If the trial was successful then the island’s tourism industry could be one of the first to be reopened and the Government would owe a favour to the island.Matt Hancock agreed with Dave.

On 4th May he announced that a trial would take place on the island before the app was rolled out nationwide, saying “where the Isle of Wight leads, Britain follows”. Some island entrepreneurs started selling t-shirts with the slogan on it. It is not known how many they sold.

The government’s announcement of the trial said that its overall test, track and trace programme was being trialled on the island, but the press release and accompanying press coverage focussed on the app. The Isle of Wight Green Party spokesperson, Vix Lowthion, asked the Government how they would  be measuring the success of the ‘test, track and trace’ programme. There was no answer.

By now it seemed like there were daily press stories about the app. It dominated questions in the Government’s press conferences.

Changing direction?

In the week following the launch of the trial it was reported that the UK was building a second contact tracing app that would follow the Google-Apple de-centralised model; that any contact tracing app would face technical challenges in being accurate enough to be useful; that there were plans to expand the app functionality, and that Parliament’s Joint Human Rights Committee recommended new legislative safeguards before the app could be safely used.

Civil servants spent an increasing amount of time trying to understand the public debate and get their messages to cut through the noise, while the people on the Isle of Wight were still deciding whether to use the app at all. 

In the circumstances it must have been pleasing that 55,000 people downloaded the app in the first few days. Unfortunately it was not clear how many were on the Isle of Wight and how many were curious journalists, technologists and politicians. Public surveys showed that a sufficiently high proportion of UK citizens would be willing to download the app, but surveys do not always reflect actual behaviour.

Unfortunately after the early enthusiasm about the number of downloads there were few official updates from the trial.

The promises of transparency were not being met. The reasons gradually became clear. 

Time for a U-turn

On 19th May, two weeks after the trial had started, Government Health Minister Lord Bethell told the House of Lords that the trial had taught Government “one important lesson: that people wanted to engage with human contact tracing first, and quite reasonably regarded the app as a supplementary and additional automated means of contact tracing”. 

Government publicly shifted their focus to the need for manual contact tracing and delivering on the Prime Minister’s statement that England would have a “world-beating” test and trace programme by the start of June. 

The aftermath

Local news site NewsOnTheWight wanted more detail. They started asking questions of both Isle of Wight and NHS officials about how the app had performed on the island. They were promised answers, but none were given.

The UK Government launched England’s new national test and trace programme on 28th May. Like the new programmes in Scotland, Wales and Northern Ireland the programme did not have an app.

The Government said that the app would be added at a later date and would be the cherry on the cake.  But it was unclear when the cherry or the cake would be in place.

There was talk of launching a second version of the app on the Isle of Wight, but it never appeared. Ministers said the app would launch at the end of June, and then that it would launch in Winter

It became clearer that the “world beating” test and trace programme was facing problems too.

What went wrong?

While central Government officials had built a new national contact tracing service, a new central data store and a new Joint Biosecurity Centre to help make decisions, local involvement was lacking and the system failed to cater for the many languages spoken in England. A more human approach was needed.

On the same day that news started to emerge of a major Coronavirus outbreak in Leicester, the Government announced that the original plan for the contact tracing app had been abandoned and that England would start to move to the de-centralised model proposed by Google and Apple .

As well as the important lesson that humans wanted to talk with humans, the Government’s announcement confirmed rumours that there were technical problems with the app. It would not work on Apple iPhones unless it used the technology that Apple had announced two months previously.

The wrong strategy?

Apart from these ministerial announcements there is still no official output from the Isle of Wight trial. We do not know whether the app found any cases of the virus that would not have been found by manual contact tracing.

Press stories continue to emerge about why the app had failed and how it could come back stronger than before, but there also seems to be a growing realisation that the app had been the wrong place to start all along.

Other countries, like Germany and France, had begun the pandemic in a similar situation to the UK and had built strong test and trace programmes rooted in local public health institutions that were trusted by the people they served. They were now trialling apps, but the apps were built on strong foundations. They really were just a cherry on an already baked cake. The cake was tasty, the cherry was a bonus.

The Government had trusted in untested technology solutions to respond to the pandemic. Perhaps putting more trust in institutions and humans would have been a better strategy.

(Reminder: this is from a book that has other things written by people who are not called Peter Wells. The contributors include Dave Miller, James Ball, Juliet Rix, Paul Corrigan, Angela Antetomaso Forbes, Dr Alex Connock, Matthew d’Ancona, Dr Steven Mccabe, Simon Morioka and Claire Kennedy, Dr John Lister, Dr Paul Davies, Dorothy Byrne, Paul Connew, Professor Barnie Choudhury, Neil Fowler, Ben Parsons, Professor Vicky Pryce, Professors David Bailey and Phil Tomlinson, and Professor John Howson. You can get the book printed and delivered by Amazon.)

The UK address mess: a way forward?

After two years and £80m the UK recently published its first Geospatial Strategy and committed £963m to a new 10-year Public Sector Geospatial Agreement.

Shortly afterwards Facebook acquired Mapillary, increasing their bet on OpenStreetMap, while the UK Government invested £500m in satellite firm, Oneweb. The 16,800 word strategy is out of date.

It needs to be more disruptive if it is going to help deliver on Government Ministershopes that data will be the backbone of the UK’s recovery from Coronavirus. It needs to do more to encourage the UK’s geospatial institutions to adapt to and thrive in the 21st century.

This blogpost contains an idea to help.

Understand the address mess and then start cleaning it up with a new government service.

UK address data is a mess

UK address data is a mess. The data is low quality. It is hard to understand what can legally be done or how to fix things.

The problems that affect citizens, particularly with new build properties, are well known. Mismatching and missing addresses have caused problems during the Coronavirus pandemic. 

The legal complexities are well known. A lot of money is spent on lawyers.

And governance is a mess. The Geospatial Commission, GeoPlace, Scottish Improvement Service, Northern Ireland Land & Property Service, local authorities, Ofcom, Land Registry, the VOA, Ordnance Survey and the privatised Royal Mail are all involved. Organisations find themselves talking with several of these institutions to get things done.

Other countries faced similar issues and are cleaning them up. While there are lots of dedicated people in the UK we are not making fast enough progress in doing the same.

UPRNs will not clean up the address mess

Government has many levers to try and clean up this mess.

The Geospatial Commission has agreed to get UPRNs published as open data. UPRNs are a unique number for every property. The Commission will learn that this is an ineffective lever to pull on its own.

The new FindMyAddress website by GeoPlace is designed to help citizens find the UPRNs for their homes and places of work.

The new website has a poor user experience. Because of the address mess the very first thing you see is 5300 words of legal text. 

But there is a more fundamental problem. UPRNs are boring and should be invisible. UPRNs help data geeks, like me. UPRNs are useful when they connect together services.

Citizens know addresses. If a citizen needs to know their UPRN then something somewhere has gone wrong. Every visit to this website is a failure.

“Take me to 10033569645, James”
Image by Elliot Brown CC-BY-SA 2.0 

Build an address service

Government needs to build a service, not a website

The website has a nice bit. The text area to search for addresses. A user starts to type in an address and it helps them find an address and UPRN. A neat bit of work by GeoPlace and Aligned Assets.

Turning that neat address search functionality into a common component that any service designer can easily reuse, let’s call it GOV.UK FindMyAddress, would be a place to start cleaning up the address mess.

GeoPlace and Aligned Assets should be working with local authorities, Government Digital Services and the Digital Land team at MHCLG to do this.

The search field would become an API designed to improve the flow and data quality from a user entering an address to a UPRN being passed to back-end services.

The code should be open source. Many people will help maintain it, while anyone can fork the code and do their own thing. Openness supports competition as well as collaboration.

The API would also help to discover errors or omissions in the data. It can feed those back into Geoplace and local authority’s address management processes.

Collaborative maintenance will gradually improve both data quality and the service. Win win.

Other things that a service team should explore

A service team would also explore business models and governance.

Perhaps the Office of National Statistics should be involved. It would provide one of the most foundational parts of statistics and the planned public sector integrated data infrastructure with long-term stability. 

Rather than regulating Royal Mail and paying them £16m for access to postal addresses, the new service could publish postal addresses too. The same data foundations can support both digital and physical services.

Collaborative maintenance will reduce costs and this new service will outcompete Ordnance Survey on some of the planned uses for their OS Places API and AddressBase products.

This will reduce the need for some of the £963m 10-year Public Sector Geospatial Agreement. Hopefully the contract allows the price to change as needs change.

This competition from within the public sector will help provide the OS with the impetus to change and provide better services in other parts of the UK’s geospatial data infrastructure.

A way forward?

Creating a GOV.UK FindMyAddress service that helps deliver better services and better address data might be a good idea. Or it might not.

But it is more likely to clean up the address mess than anything in the UK’s Geospatial Strategy.

We need to fund teams to explore innovative ideas if the UK’s geospatial data infrastructure is going to thrive in the 21st century.

© 2020

Theme by Anders NorenUp ↑