Data Science Europe Presentation
Generating Trust through reliable data: International Travel with CoVID-19
goPassport Chief Technology Officer Director, Klaus Felsche, presents goPassport to Data Science Europe to explain why we need goPassport as part of the solution to making travel Covid safe.
Presentation: Data Science Europe
Klaus begins by welcoming viewers to a brief look at goPassport and some of the challenges associated with using data to reduce the impacts of the current pandemic on international travel.
He continues by explaining that there are some components of the goPassport system that are still in development. While we have an in-production real-time rules and messaging engine, an itinerary management system and most of our web-interfaces built, the mobile app is an alpha version.
Klaus emphasises that we have, however, built our Risk Analytics System which is the main focus for his presentation. There is a public version of the system via our Live Risk Map on our website.
Founded the first advanced analytics unit in Immigration and Border Protection. Developed and deployed the analytics-driven Border Risk Identification System including multiple components of the Next Generation Border Security System, the analytics-based Risk Scoring Service for major visa platforms and a risk micro-service for the Electronic Travel Authority.
Objectives of goPassport
- Keeping travellers safe.
We are focused on the traveller – aiming to use data to assist the traveller in planning a safe journey, staying safe whilst on their journey and aiding the traveller as required. To do this, we require global data about existing and emerging risks (e.g. bushfires, floods or any new, post CoVID-19 health threats that may emerge) – not just CoVID-19.
- Keeping the community safe.
Opening up borders to travellers requires a high degree of confidence that inbound travellers pose no significant risk to the local community, particularly if the community has managed to control the pandemic within their country. If this cannot be done then all foreigners will be looked on with deep suspicion: everyone knows that COVID-19 travels by air or sea, hosted by human carriers! goPassport’s technology will provide the reassurance needed that reopening borders and allowing travellers, who pose no immediate risk to the community, is safe.
Current CoVID-19 controls
Currently many countries tend to use binary and blunt instruments to contain the pandemic (or keep infection out):
Quarantine, first described over 600 years ago, is still considered best-practice. New Zealand is a good example: all arrivals are quarantined for 14 days.
Other countries have effectively given up and are open to all – except that some end up going into a severe lock-down when health systems become overwhelmed.
Some countries rely on on-arrival COVID-19 testing, some require a pre-arrival test. As you will see later, these tests have limited utility as asymptomatic carriers may not be identified early in their infection cycle.
Current CoVID-19 conundrum
the economy v health
mental health v physical health
quarantine v free movement
open borders v closed borders
Governments face some tough decisions, some governments have little choice, facing economic disaster or high fatality rates.
In all of this… is safe travel even possible?
Can we guarantee that a traveller represents no risk? Can we guarantee that the journey is safe for the traveller themselves?
The answer is Yes.
We have the technology to allow individualised pathways to travellers, based on the risk they represent. For example, we need not apply the same process to every traveller because each traveller represents a different risk profile. And we now have the technology to manage travellers based on the individual risk profile: no more: “one size fits all”
If any of you have travelled to or from Australia since 2010, you have been risk-assessed by real-time, high-performance advanced analytics systems designed to identify risk.
These systems resulted in accurate identification of a range of risks in visa and border processing and enabled many travellers to access fast-tracked processing as they were identified as low risk.
All of these processes rely on our capacity to apply accurate, complex, real-time analytics. Our Border Risk Identification System allowed over 270,000 travellers to be treated as low risk, saving nearly a million hours of inconvenience for travellers and staff time for our border staff in its first year.
“You can probably see where I am going with this: If we can do the same for COVID-19, safe travellers can be fast-tracked safely!” Klaus explains.
Managing most risk pre-travel
Before travelling we assess and mitigate the risk through multiple control layers. Klaus learnt during his time working in Immigration and Border Protection, that any control by itself has gaps and that a good strategy is to have multiple layers of control to compensate for the limitations of one layer.
Here is the goPassport model:
Please note that not all the layers are necessary for everyone. For example, New Zealanders travelling to Australia only require symptom screening and flight risk assessment because the country risk is extremely low. Travellers from higher risk ports will be required to pass through more controls to ensure that they represent no risk.
Klaus then describes some of of the model layers in more detail.
- Country Risk: we assess the current risk down to province/state level using widely available data.
- Travel History: this is important as a person may well have, over the last two weeks, visited locations that are higher risk than their home. For example, a New Zealander who has been to the US in recent days will be considered at a much higher risk.
- Symptom Checking: Where symptom checks are carried out by health professionals, we would expect that travellers with symptoms can be screened out – for example at airports. We also include a self-check function to allow travellers to assess their own health and, yes, it is recorded and evaluated within goPassport.
- COVID testing. It is a powerful tool when used correctly, powerful but not perfect.
- When vaccines become available, we will have the capacity to record these and reassess the risk posed by travellers. Like influenza vaccines, they are not silver bullets: their effectiveness is unknown.
- We assess airports, aircraft and airlines. Airport polices that reduce risk of infection, airlines that have cleaning policies and traveller management practices (eg mask wearing, seating arrangements, separating cleared travellers from those who are not, crew health procedures, luggage handling, etc)
- Importantly, we look at the route taken, We consider length of flights, traveller mix (eg mixing low and high risk travellers at transit hubs).
- goPassport support monitoring self-isolation and quarantine compliance through GPS location monitoring. This allows travellers to isolate hat home or in supervised locations before departure, reducing any on-arrival quarantine period.
In the end, we ask two fundamental questions. How likely is it that a person is infected? How likely is it that a person visiting a location will become infected?
Data enables international travel
Making data-informed decisions requires trust in data validity:
- goPassport has deployed a range of analytics techniques designed to overcome limits imposed by available global and regional data.
- Getting it wrong can have serious consequences for both, the individual traveller and the general community.
It’s about location data
You can view the Interactive Live Risk Map here: https://covid-19.gopassglobal.com/public
At first glance, you may just think this is ‘just another CoVID map..’ however, closer examination will show you some value-adds such as descriptive analytics and trend predictive analytics!
The real use of the data is less visible, what you see in the example above it just the tip of the iceberg. For example, when we ingest a travel itinerary, every location is checked for current and emerging risk and it’s not always about COVID. It could be about storms, floods, bush-fires, earthquakes, volcanoes or even civil unrest. Our database accesses a range of different data sources to assess an itinerary and then alert a traveller along the way.
It’s about clinical data: test regimes & attack rates
There are two types of test and administering them at the right time is critical to enhancing the reliability of the test process: Conducting an antibody test (Ab) within the first 6 days following infection is a futile activity – the accuracy is only 40%. Conducting a Nucleic Acid/PCR test from day 14 on is similarly flawed. The right combination of testing can more than double the reliability. goPassport understands this and guides a traveller through the best sequence and monitors outcomes.
Knowing where a person sits on a flight, train or bus and for how long is also important. All of these factors are included in the risk assessment for a journey.
It’s about the traveller’s current risk
goPassport’s job is to identify the most appropriate path for a traveller based on the policy settings for the countries visited and the potential risk they represent to those societies.
It’s about the journey risk
We have also done work on determining the impact of every potential risk vector on the traveller risk – location risk data, airline & airport risk, a person’s travel history and immunisation are just a few of the components that provide a journey risk status. For example, the importance of an airline risk increases with longer, multi-stop flights.
Making the data work
None of the data is of any use unless it is put to work. It has to be in the right place at the right time and it has to be in a form that can inform action.
Border authorities and airlines just need to know if a traveller is cleared. They need it for automated systems and they need it before boarding to ensure that any channelling of low risk and unknown risk travellers at airports and departure and arrival gates can happen to minimise cross-infection.
Health authorities need to know when a traveller is in need of assistance or should be arriving in quarantine. They will also need to know when and where a traveller should be in home-quarantine if that is required.
Travellers will need to know if their journey has health risks. They will also need to know if they get close to a high-risk area, particularly if the area was previously considered to be safe.
All of this requires a high level of automated, real-time messaging based on constant analysis of the traveller’s intended and actual location and the traveller’s relationship to risk locations.
Looking to the future
Once goPassport starts to collect traveller data from around the world, we will be in a great position to enhance our capacity to ensure safe travel. Collecting anonymised data about test results alone will help us identify emerging risks early. Our predictive models will look further into the future and become more accurate. We will be able to more accurately identify problematic locations, flights, accommodation providers and other risk vectors by tracing back any occurrences of infection during travel.
What about privacy?
We built-in ‘privacy by design’. End-to-end encryption and the ability to delete all personal data at the request of a traveller is built into our core. The depersonalised data is then able to support system enhancements and legitimate research.
Klaus finishes his presentation by thanking all for watching and listening. He invites anyone with questions to contact him via email firstname.lastname@example.org or visit our website.