By: Nelson M. Nones CPIM, Founder, Chairman and President, Geoprise Technologies Corporation
Now that the first wave of the 2019–20 coronavirus pandemic (COVID-19) appears to be cresting in many parts of the world, attention is quickly shifting to reopening the societies and economies which were locked down in order to mitigate the pandemic’s impacts on healthcare systems.
But ending a lockdown releases potentially infected people from the confines of their homes, introducing the risk of re-igniting the contagion. To prevent this, public health authorities need to implement containment strategies which:
- Quickly test as many people as possible;
- Isolate patients as soon as new COVID-19 cases are detected;
- Trace each patient’s known contacts with other people to warn, test and potentially quarantine those who are at risk of infections spread by the patient; and
- Manage the detected COVID-19 cases effectively to maximize patient recoveries.
Contact tracing is regarded as an essential tool for reopening society while reducing the likelihood as well as the ultimate severity of second and subsequent COVID-19 waves in the future. Unlike a lockdown, which inflicts great economic pain because it indiscriminately isolates both the well and the sick, contact tracing provides a mechanism for isolating only the sick and those who are at greatest risk of falling ill, while allowing everyone else to go about their daily lives as usual.
Geoprise operates in the United States, the United Kingdom, Thailand and Singapore, so right now we are locked down until at least the end of this month. Fortunately, all of us remain healthy but we are working from home. This has given us opportunities to examine how we can best contribute to reopening the places where we live and work, and we’ve come up with an idea for accelerating the adoption and improvement of contact tracing using our existing and proven GM-X technology.
Contact Tracing the Old-Fashioned Way
Contact tracing is a well-established process illustrated in Figure 1. It traditionally relies upon patient interviews, and possibly additional interviews with family members, healthcare providers and others, to discover the patient’s movements and the people with whom the patient had close contact after becoming contagious. Interviews are usually conducted by phone.
Figure 1 – Traditional Contact Tracing
The information gleaned from these interviews is typically recorded in a centralized database backend which is used by epidemiologists working at individual healthcare providers and health authorities during the course of an outbreak.
Minnesota, where Geoprise corporate headquarters are located, has a population of 5.6 million and is home to the world-famous Mayo Clinic which has its own centralized internal contact tracing database. The Mayo Clinic, in turn, collaborates with the Minnesota Department of Health which operates the Minnesota Electronic Disease Surveillance System (MEDSS), a centralized database and interactive Web-based platform used since 2006 by state and local public health departments, laboratories and clinics in Minnesota. The state uses its MEDSS system, in turn, to report nationally notifiable infectious diseases to the U.S. Centers for Disease Control and Prevention (CDC).
Like Minnesota, Singapore has a population of 5.6 million. The Singapore Ministry of Health established an Infectious Disease Alert and Clinical Database system after the severe acute respiratory syndrome (SARS) epidemic ended in 2004.
Unfortunately contact tracing done the traditional way is extraordinarily labor intensive. By late March 2020, when it was recording about 50 new cases per day, the Minnesota Department of Health reportedly employed dozens of contact tracing staffers working round-the-clock in shifts. And in early March 2020, when Singapore was reporting just 5 new cases per day, its Ministry of Health was reported to have formed a dedicated team of 140 people working on contact tracing in shifts from 8:30 A.M. to 10:00 P.M. every day of the week.
Beyond telephone interviews, Singapore’s contact tracing team also utilizes secondary data sources including mobile phone location tracking records, police records and usage of automated teller machines (ATMs) and credit cards to track the whereabouts of patients during the 14 days before they became symptomatic. Their goal is to form a complete picture, known as an “activity map,” within one day after a patient tests positive for COVID-19.
Traditional contact tracing methods may be adequate to handle typical infectious disease outbreaks, but they have been overwhelmed by the 2019-20 coronavirus pandemic which has erupted at a scale not seen in over a century. The evidence from Minnesota and Singapore suggests that each new COVID-19 case currently creates a contact tracing workload equivalent to anywhere from 1 to 24 people working an 8-hour shift, meaning that in the U.S. alone, a workforce approaching 150,000 people would need to be mobilized full-time for contact tracing after the initial wave passes. To put this in perspective, the Transportation Security Administration (TSA) employed about 45,000 screeners at U.S. airports in 2018.
Considering the time it takes to hire and train personnel, we believe it is unlikely that traditional contact tracing processes can scale up to the task of reopening society anytime soon.
Contact Tracing – Recent Developments
Now that some 45% of the world’s human beings own a smartphone, the technology embedded in those devices holds out the promise of substantially reducing the human effort needed for contact tracing.
Smartphone location tracking capabilities aren’t up to the task, however, because their global positioning systems (GPS) are typically accurate to within 5 meters (16 feet) at best which is well beyond the 2 meters (6 feet) of social distancing which health authorities are currently recommending ― and also well beyond the 4 meters (13 feet) which some epidemiologists, according to the latest news reports, believe is most appropriate in view of what recent data is revealing about COVID-19 transmission. But nearly all smartphones have Bluetooth radios which are considered accurate for measuring distances of 2 meters (6 feet).
Singapore has already released and begun to use the world’s first contact tracing application for smartphones that takes advantage of Bluetooth technology. The TraceTogether app, as it is known, automatically logs a history of encounters with other smartphones which come within 2-meter or 4-meter range. Should the phone’s owner test positive for COVID-19, the owner can voluntarily upload the history to the backend.
At present the Singapore Ministry of Health’s backend performs a risk assessment by filtering the uploaded histories to identify close contacts using known time and distance parameters for the COVID-19 virus, and then relays the filtered histories of close contacts to the Singapore contact tracing team which follows up by initiating traditional contact tracing interviews with patients.
Researchers in other countries are proposing alternative approaches to further automate the data analysis and risk assessment process. Instead of uploading the smartphone owner’s entire history, only encounters beginning on the day when the patient was deemed contagious would be uploaded. The backend, in turn, would automatically relay the uploaded history of encounters to the mobile apps installed on all the other smartphones. Each encounter in the history could include a risk score computed by the backend before the history is relayed, or the mobile app installed on the other smartphones could compute risk scores locally after receiving the history of encounters. The mobile app installed on each smartphone would then check its history to determine if it has observed any of those encounters in the past and, if so, display an alert to the smartphone's owner whenever the risk score exceeds the close contact threshold.
Smartphone owners who receive such an alert would then watch for symptoms and get tested for COVID-19. If any of their tests are positive then they too can voluntarily upload their histories to the backend, and the cycle repeats itself until no further proximity contacts exist.
Singapore has released the open source for its TraceTogether app for use by other health authorities. Building upon this and the work of other researchers, Apple and Google, which between them hold a combined 100% smartphone market share, recently announced a partnership to roll out a common application program interface (API) in May 2020 which public health authorities can incorporate into the apps they publish for Apple’s iOS and Google’s Android operating systems.
Using the Bluetooth technology already embedded in smartphones for opt-in proximity sensing in lieu of traditional contact tracing is alluring, but not without several inherent limitations.
For starters, many people don’t own or use smartphones and a sizeable number of those people, including nursing home residents and prisoners, are highly vulnerable to COVID-19 infection.
Those who do own and use smartphones can expect a large number of false positives when, for example, a pair of smartphone owners are situated within 2 meters (6 feet) or 4 meters (13 feet) of one another, but in different apartments sharing the same wall, floor or ceiling.
App installation is voluntary in most countries where apps are available, but uptake has been slow so far owing to privacy and security fears. For instance, it’s estimated that only about a quarter of Singapore’s smartphone owners had installed the TraceTogether app by early April 2020.
Importantly, in response to security and privacy concerns, Apple’s and Google’s partnership intends to maintain strong user privacy and security protections, meaning that smartphones won’t reveal any personally identifiable information (PII) to other smartphone owners or public health authorities, whether or not the other smartphone owners are deemed at risk of infection, and even when the app installed on those devices communicates with a centralized backend. Patients’ identities would be revealed to public health authorities only when those who test positive for COVID-19 voluntarily upload their phone’s encounter history to the backend. Consequently, the contact tracing records will contain very little of the PII that epidemiologists have come to expect from traditional centralized contact tracing databases.
These limitations indicate that Bluetooth technology is best suited for augmenting rather than fully replacing traditional contact tracing methods, by boosting the productivity of contact tracing teams in order to accelerate the re-opening of society.
Introducing mobile apps and Bluetooth technology for proximity sensing will require healthcare providers and public health agencies to augment or modify their existing contact tracing databases to accept, process, store and use new kinds of information never envisioned when those databases were designed and built. Most centralized backend databases and applications in use today were deployed one or two decades ago and are designed to hold phone numbers, postal addresses and perhaps email addresses ― not the cryptographically hashed ephemeral identifiers or secret day keys envisioned by proximity sensing researchers as illustrated in Figure 2.
Figure 2 – Contact Tracing Using Smartphones
By Prof. Carmela Troncoso et. al, “Decentralized Privacy-Preserving Proximity Tracing,” April 2020
Modifying those systems to accommodate the new technology may prove to be difficult, time-consuming and expensive, as many U.S. state employment agencies are discovering after their decades-old systems were inundated with new unemployment insurance claims filed by workers who were laid off or furloughed because of the COVID-19 pandemic.
At Geoprise, we believe that healthcare providers and public health authorities can address this issue quickly by deploying our existing and proven GM-X application technology.
Deploying the GM-X Application for Contact Tracing
The GM-X Party Subsystem addresses the needs of organizations to answer questions such as:
- What are the attributes or characteristics of the people who are involved?
- What relationships exist between various people, between various organizations, and between people and organizations?
- What are the addresses, phone numbers, and other contact mechanisms of people and organizations, and how can they be contacted?
- What types of communication or contacts have occurred between various parties, and what must be done to effectively follow up on these communications?
Conventional data models maintain information about organizations and people in various entities that are portrayed completely separately. Because each application within an organization has its own needs, the data modeler will usually base the model on the needs of a specific application.
The problem with keeping this information in separate entities is that people may also have different roles that change over time. Conventional systems will record redundant information about a person because they store a different record each time the person’s role changes. For example, a person’s role could change from Anonymous Contact to Personally Identifiable Contact. It is also possible for a person to have several different roles at the same time; for example, both Personally Identifiable Contact and Patient.
The solution to this redundancy problem is to have a single entity which can store information about any type of person, with details of their various roles, relationships, contact mechanisms and communications maintained on separate entities.
In the GM-X data model, attributes which are common to both organizations and people, such as current status, are stored in a Party entity, while attributes and characteristics which only people possess, such as date of birth and gender, are stored in a Person entity. The Party Role Link entity identifies each party’s past, current and even future roles for both organizations and people based on starting and ending dates.
When people must remain anonymous for privacy protection reasons, it’s highly likely that an individual person will be represented by multiple identifiers on the backend database. Consequently, the backend database must be designed to accommodate a very large number of parties. The GM-X application meets this requirement by maintaining up to 100 septillion parties per database instance.
The GM-X application allows any number of party roles such as contact, patient, doctor or nurse to be assigned to each person. It also allows any number of party roles such as public health agency, hospital, clinic, convalescent care facility, nursing home or dormitory to be assigned to each organization.
Any relationship between two parties, including encounters sensed by smartphones belonging to the respective parties, can be stored in the Party Relationship entity. This includes information about the type of relationship as well as each party’s role in the relationship. There may be more than one relationship between two parties, and each entry is only valid for a particular time period, so it is possible to have different values for different dates. Multiple relationship types can be configured within the GM-X application to differentiate between close, unsafe and safe smartphone encounters inferred from the Received Signal Strength Indicator (RSSI) and device model values captured in the history.
Each party in the GM-X database can have multiple contact mechanisms including mobile apps identified by cryptographically hashed ephemeral identifiers or secret day keys in addition to conventional phone numbers, postal addresses and email addresses.
Sharing Information Between GM-X Contact Tracing Database Instances
Traditional centralized systems and databases are vulnerable to data corruption as well as data theft risks. That's because anyone with sufficient credentials can alter the contents of a database without leaving a trace.
Blockchain, a disruptive technology which has emerged only within the past decade, is one of the most heavily tested and secured technologies in the world. It’s a technology that makes such data corruption and theft impossible.
Geoprise Technologies was the first software house in the world to step forward with a pragmatic solution to leverage this new technology for backend applications.
The GM-X application facilitates secure, permanent, tamper-proof and verifiable information exchange between different contact tracing backend systems using permissioned blockchains derived from Bitcoin blockchain technology. It includes a Blockchain Server which runs on either Microsoft® Windows® or Linux®/UNIX operating systems. It is installed at each participant node and can be clustered for load balancing and high availability.
Assuring Backend Privacy and Security
Compliance with privacy and security requirements is essential for any backend contact tracing system to be taken seriously. Inadequate privacy and security safeguards could lead to delicate information falling into the wrong hands, or being modified or destroyed by unauthorised personnel.
The GM-X application provides robust privacy and security safeguards through secure logon and logoff; optional logon using two-factor authentication (TFA), one-time password (OTP) or single sign-on (SSO); mandatory user authentication and role-based access controls at the task level; and optional field-level role-based access controls. It is hardened to resist session fixation and hijacking, cross-site request forgery (CSRF), code injection, structured query language (SQL) injection and cross site scripting (XSS) attacks.
This is a call to action for deploying our existing and proven GM-X application as the backend contact tracing database and application for a new generation of mobile apps that use Bluetooth technology to automatically record encounters between people who might be infected with the coronavirus.
As the initial wave of the COVID-19 pandemic subsides, it will be necessary to reopen societies and economies as quickly as possible in order to minimize the pandemic’s adverse economic consequences. This can only be accomplished through widespread COVID-19 testing and surveillance together with contact tracing to identify and isolate those who are infected, and potentially quarantine those who are at greatest risk of falling ill, while allowing everyone else to resume their normal lives.
Automated proximity tracing using the Bluetooth technology already embedded within most smartphones has the potential to significantly improve the productivity of COVID-19 contact tracing teams, thereby enabling policymakers to accelerate the pace of economic recovery ― but only if backend contact tracing databases and applications can be deployed just as quickly as new mobile contact tracing apps are rolled out to the public. Otherwise the time and cost of adapting existing backend applications to incorporate the new technology will delay those productivity improvements and hamper the recovery.
Having thoroughly evaluated solutions like Singapore’s TraceTogether app as well as the proposals recently put forward by researchers in other countries, we at Geoprise are confident that public health authorities and healthcare providers can deploy the GM-X application immediately as the backend contact tracing database and application for mobile contact tracing apps.