Biometrics for vaccine delivery

The WHO estimates that more than 50% of vaccines may be wasted globally every year, and do not reach their intended beneficiaries. Further, data on immunisation rates vary, with WHO estimates often more conservative than and drastically different from administrative reports. In such a scenario, contactless biometrics can help promote verified delivery of vaccines, ensuring that each vaccine reaches its intended beneficiary. When adopted at scale, our solution can provide accurate data on vaccination rates, and minimise wastage by allowing governments to tie each vaccine dose to a child who needs it.

Across the developing world, there are often huge gaps between administrative and WHO survey vaccine coverage data: for instance, the average gap in Bangladesh is over 17%, and in Nigeria the discrepancy in numbers rises to almost 34%. However, often administrative data rely on health cards for vaccine history, which are likely to be lost or misplaced: in Nigeria, only 1 in 3 children have such vaccine cards to prove their vaccine history. In such a situation, parents’ recall determines data logged into health records, which is likely to be inaccurate. As these aggregated estimates guide policy building, it is extremely important for them to be as accurate as possible, as each percentage gap indicates millions of children slipping through the cracks and missing out on life-saving vaccinations.

Our solution is a camera-based, contactless, biometric product, which utilises facial recognition to identify people. An advantage of using contactless biometrics is that it protects health workers from potential transmission by fomites, while allowing them to do their job efficiently. On the ground, this involves a health worker using our mobile app to register people’s biometrics against a record on a data collection app where their vaccination history is recorded. The next time a beneficiary turns up, the health worker can pull up their record using face biometrics, ensuring that the right record is identified, and continuum of care is maintained. 

While the AI revolution has accelerated the potential of mobile-phone camera-based biometrics like facial recognition, progress has been hampered by the crippling racial and gender biases found in camera-based algorithms. Simprints’ contactless biometrics solution harnesses inclusive, ethical AIto improve accuracy, cost-efficiency, and scalability in future impact projects. 

Biometric data is extremely sensitive - unlike a password, you can’t change your biometric. We extend the principles of the EU’s GDPR (the strictest in the world) to any country where we operate, and implement the strictest security measures when processing and storing biometric data.

Vaccine cards and registers are a common sight in health clinics across the developing world. However, they pose a significant challenge for health workers to maintain, as paper records are easily lost or misplaced. For policy-makers, paper records make it extremely hard to obtain accurate data on immunisation rates to shape interventions. For beneficiaries, most often children, lost health cards often mean missing out on life-saving doses of essential vaccines. Another consequence of these challenges is wastage of valuable vaccine doses, as they often go unaccounted for.

To address these challenges, Simprints has been working with Ministries of Health and health workers over the past 5 years to build a rugged solution suitable for the last mile. As a result, our contactless biometrics solution can be used in the field without any internet connectivity, and has been designed keeping health worker needs in mind. By tying beneficiary health records to biometrics, Simprints can help capture those who fall through the cracks between household, health post, and centre. Simprints leverages biometrics to create an integrated digitised healthcare system that links everyone to a single patient record across the entire immunisation process.

We cannot solely treat disease outbreaks reactively, but a fast proactive response requires an accurate database of those who have been reached by vaccinations, and more importantly, those who have fallen through the cracks. Absence of accurate data on vaccine history, including follow-ups, as is common in many low- and middle-income countries, can not only lead to new infections, but also render policy interventions ineffective as the actual scale of missed or zero-dose populations remains unknown. Further, in the case of a new disease where vaccines are scarce, identification can also help minimise wastage, by accounting for every dose administered.

Simprints started off with a fingerprint biometric system, which was the world’s first designed for frontline health delivery in remote, offline environments. Given our prior experience delivering programs on the frontline, ours is the only biometric company that can unequivocally verify that every vaccine has reached its intended recipient - and do it whilst respecting and protecting the privacy of beneficiaries. And it works: Brown University’s evaluation of the deployment of Simprints across clinics in Malawi offering clinical care for pregnant women with HIV, and found that the biometric system captured nearly 50% of HIV visits that were missed by the current EMR system used to monitor HIV visits.

Our competitors and existing biometric technologies are largely tailored to developed-country contexts and unsuitable for the frontlines of global health. Additionally, most of these are not interoperable with major digital health systems and are typically inaccurate due to scarred fingerprints and biased algorithms. Simprints has remained laser-focused on last mile technology. We build our solutions to be rugged, wireless, long-lasting, and highly-accurate and we continuously optimise for frontline contexts, such as reliable syncing in unreliable areas, wraparound data analysis, and streamlined workflows.

Developments in AI have accelerated the potential of mobile-phone camera-based biometrics, eliminating the need for separate hardware, cutting costs and increasing scalability of digital identity. However, progress has been hampered by the crippling racial, ethnic, and gender biases found in camera-based AI algorithms. Responding to these challenges, in July 2018 Simprints began exploring inclusive, ethical AI-powered biometrics to improve accuracy, cost-efficiency, and scalability in future impact projects. Our aim is to reverse the current Western data "bias" within camera-based AI, as well as the widespread oversight in protecting beneficiaries’ privacy rights and data security, to provide a powerful, ethical, and inclusive biometric identification system at scale.

Our product is a mobile phone identification app, built using a commercial face recognition and matching Software Development Kit (SDK). The current prototype and eventual product both allow for upsync to and downsync from a cloud server. To put privacy-first, images that are uploaded to the cloud are deleted from the user’s phone, and face-related biometric information is stored on the phone as a signature/template (strings of numbers), which cannot be converted back into the original image nor used without the algorithm. This means that while the product works offline, beneficiary privacy remains protected.

To date, other biometric vendors have made use of proprietary biometric systems, effectively “locking in” buyers. Simprints promotes sustainability by using open template standards (ISO 19794-2) to prevent vendor lock-in, and have developed basic integrations with many of the most widely-used digital systems, including DHIS2 and CommCare.

Simprints is already in use in 12 countries, and has reached 400,000 people to date.  In 2016, we deployed Simprints in a pilot project with BRAC HNPP to reach 28,000 expectant mothers and newborns with essential care. A longitudinal study funded by DFID and the Global Innovation Fund showed that the deployment of Simprints increased postnatal healthcare coverage by 11% and antenatal healthcare coverage by 38%. Because of that increase, 9% more women gave birth with a skilled birth attendant and 19% more newborns received essential care. Brown University’s evaluation of the deployment of Simprints across clinics in Malawi offering clinical care for pregnant women with HIV, and found that the biometric system captured nearly 50% of HIV visits that were missed by the current EMR system used to monitor HIV visits (https://link.springer.com/article/10.1007/s10461-019-02748-6)

Prevention of, and response to disease outbreaks requires tools that support and protect health workers, and strengthen the overall health system. While vaccines go a long way in ensuring both, incomplete, outdated and unreliable population and health data, compounded with reliance on fungible, paper-based health records make it extremely hard for health workers to identify those who have been missed out in immunisation campaigns. Simprints’ contactless biometrics, integrated into frontline worker workflow, can help create a biometric identity linked to each vaccination record, providing a unique link between individuals and their vaccine record that cannot be lost or forgotten. This results in reliable data on number of individuals vaccinated, avoidance of duplicate records, and reduction in wastage by prevention of unnecessary double doses. At the policy level, accurate digital records can produce reliable population-level immunisation coverage data to inform public health decisions, which, in the long run, can help achieve equitable and robust vaccination coverage, protecting lives by conferring herd immunity against infectious diseases. This breaks down into:

1. Issues to address:

   1. Incomplete, outdated and unreliable population and health data, 

   2. Methods to track vaccine records rely on cards or ID numbers that can be lost, forgotten, stolen, swapped or destroyed

   3. Need for identification methods that do not risk transmitting highly infectious diseases

2. Inputs

   1. Simprints’ contactless biometrics integrated into frontline worker workflow

   2. Training for frontline workers on using smart devices and contactless biometrics

   3. Community sensitisation for using contactless biometrics

3. Outputs

   1. Biometric identity created for every individual receiving the vaccine, without risk of infection

   2. Biometric identity linked to vaccination record, providing unique link between individuals and their vaccine record that cannot be lost or forgotten

4. Outcomes

   1. Reliable population-level immunisation coverage data to inform public health decisions

   2. Maximisation of resources by ensuring individuals do not receive unnecessary double dosages

   3. Biometric IDs provide basis upon which to build further health system strengthening initiatives

5. Impact

   1. Equitable and robust vaccination coverage, protecting lives by conferring herd immunity against infectious disease

Simprints’ fingerprint biometrics technology currently serves over 400,000 beneficiaries, and is on track to reach 700k by the end of the year. Our camera product has presently been piloted with 656 people in Kenya for a program that was distributing aid in schools. We're going to be running numerous facial recognition projects in the next year, including partnering with Gates for Operation Fistula in Madagascar, aiming to reach over 100k people in total. Additionally, we’re in talks with MoH Ethiopia and Gavi to deploy our solution for the verified delivery of a COVID-19 vaccine, once it becomes available. Through these partnerships and others, in the next five years we expect to reach 10m people.

Simprints’ mission is to transform the way the world fights poverty. We build technology to radically increase transparency and effectiveness in global development, making sure that every vaccine, every dollar, every public good reaches the people who need them most. In the next year, we are projected to reach 2M beneficiaries, and have a goal of 10M for the next five years.  

We anticipate being able to serve NGOs, aid organisations and governments with our contactless solution. This includes our current partners such as the Ethiopian Federal Ministry of Health and BRAC, who will be able to expand their use of biometrics to include facial recognition without needing to pivot away from their existing data-collection tools. We will also be able to meet the needs of new partners more effectively, particularly those who require touchless biometrics (such as a government responding to a pandemic), or those who require a rapid deployment (such as an aid organisation responding to a humanitarian crisis). Specifically, the world’s increasing vulnerability to epidemics and pandemics provides both an opportunity and an imperative to scale by partnering with governments to ensure effective and equitable national-scale vaccination programs.

A key barrier for us in accomplishing our goals is obtaining core funding. While most project budgets cover program costs, core costs are often barely accounted for or often overlooked. Core funding can enable us to invest in continuous R&D, build an incredibly talented team, and ensure our organisation stays sustainable in the long run. 

One technical challenge is ensuring that the structure of our system will allow us to operate effectively at scale with government partners. For biometrics to have long-term sustainable impact, there needs to be a high level of local capacity by partners to choose and maintain their systems. However the requirements of Ministries of Health are very different from businesses or NGOs. While for-profit biometric companies often focus on end-to-end, proprietary, and cloud-based systems, our interviews with multiple Ministries suggest the priorities are actually systems with modular hardware, modular mobile software, server hosting with local storage, and significant capacity transfer services for local staff. This gives Ministries the power to maintain data sovereignty and replace components at any time without major re-builds, and therefore rapidly scale secure, cost-effective, and locally-maintained biometric systems that fit their needs. 

To overcome the core funding barrier, we are working with our network of current partners to raise catalytic funding for Simprints. While this has been put on hold due to the COVID-19 pandemic, we are hopeful that our conversations will progress, and we will be able to raise the required core funding.  

In order to meet requirements for system modularisation, we will continue retooling our existing architecture to be fit-for-purpose for Ministries. We are currently seeking innovation funding to help us accelerate this process, so that we can prepare to deploy biometrics as and when a COVID vaccine becomes available.

Simprints currently has 43 full time employees that work on its fingerprint and camera products, across product development, project delivery and business development.

Simprints has been working with biometrics on the frontlines for the past five years. In this time, we have learnt that biometrics are not a standalone solution, but instead can add great value in collaboration with the right partners. Today, we have 43 team members around the world, with offices in the UK, Bangladesh and Ethiopia. These are people we hired from leading companies all around the world: BCG, the UN, or Amazon to name a few. And we have purposefully focused on team diversity to build the most inclusive solutions with over 20 different nationalities represented. In addition, our board of directors brings together world-leading experts from Salesforce, Google, and Harvard Medical School, among others.

Simprints, as a solution, has been designed with a partnerships model in mind. On most projects, we work with an implementing partner and plug into the workflow of a data collection platform. To make our projects sustainable, Simprints has also set up regional offices in Bangladesh and Ethiopia to equip our partners with the ability to run and maintain systems with minimum support from us once the program is fully set up. We currently have three groups of partners:

• Governments: MoH Ethiopia, MoH Bangladesh

• Private funders: CIFF, Botnar, Gavi, ARM, Cisco

• NGOs: BRAC, D-Tree

Simprints focuses on building tech that is optimised for the development context, with frontline workers in mind. Our platform is rapidly deployable, with minimal set-up required, making it easily scalable for larger projects. In addition to program accountability, we can provide deep insights into program activity through customised M&E reports. 

 Additionally, each project is allocated a dedicated project manager, who provides advice on program design, implementation, and workflows. They also conduct customised training for users to be onboarded on our platform, and are available round the clock if any assistance is required. 

To sustain our business, we are committed to being an earned revenues-driven organisation. We partner with governments and private philanthropy organisations to introduce verified impact in a range of programs, including MDA for neglected tropical diseases and continuity of care in MNCH services. We also raise innovation funding to improve our product and keep up with the latest developments in technology. 

From the beginning, we fully committed to the social enterprise model of putting impact over profit, while ensuring that we build a sustainable tech company that can achieve scale through B2B and B2G sales. Simprints became cashflow positive in 2019, four years after our founding, having already won over £11M of funding at a 65% compound annual growth rate by focusing exclusively on lean product development, earned revenues, and financial discipline. In keeping with this drive toward financial independence, we aim to concurrently scale up our business development with increasingly large-scale project contracts. We expect to continue to fund our work through a mix of innovation grants, government projects, and catalytic funding. 

Our commitment to strengthening health systems to enable them to better respond to disease outbreaks directly aligns with the goals of the Challenge. Often, the responses to outbreaks are short-sighted, and only focus on fixing the issue at hand, rather than strengthening the health system as a whole. Our interest in applying to Solve stems from the contest’s interest in sustainable, long-term solutions that prevent, rather than reactively respond, to crises by strengthening health systems as a whole.

Further, MIT is a hub and a frontrunner in all things tech. As our solution builds on existing AI models, Solve can help us access a network of renowned experts at MIT to refine our product and enhance its market fit. 

Finally, an ‘MIT Solve’ accreditation adds a lot of credibility to a new product like ours, helping us build partnerships and expand our reach, having an outsized impact on our finances way beyond the prize funding itself. 

Simprints would be keen to partner with JPAL, an off-shoot of MIT, for M&E. While the impact of biometrics has been proven in smaller studies, we are still looking to conduct a full-scale RCT to show the impact of integrating biometrics into global health programs.

We would also be keen to engage with MIT faculty as advisors who can help us explore new regions for scale, improve our product-market fit, and develop our product further. This could be in the form of ad-hoc advisory calls, or adding them to our advisory board for more regular interactions.

Simprints’ mission aligns with that of the Elevate prize: elevating opportunities for all people, especially those who fall through the cracks of service delivery because of their lack of ID.

So far, Simprints’ fingerprint solution has proven its impact on the ground: a longitudinal study funded by DFID and the Global Innovation Fund showed that the deployment of Simprints increased postnatal healthcare coverage by 11% and antenatal healthcare coverage by 38%. Because of that increase, 9% more women gave birth with a skilled birth attendant and 19% more newborns received essential care. We have also succeeded in making our solution sustainable and scalable, both financially and operationally. 

We understand that we, alone, cannot fulfil the needs of the 1 billion people across the world that lack any form of identification. To drive action, we’ve been working with the World Bank’s ID4D (through the Mission Billion Challenge) as thought leaders in the field, ensuring that there is awareness and adoption of best practices in the digital identity space.

The Elevate Prize will allow us to apply our learnings from fingerprint biometrics to our contactless biometrics product, and enable us to have it ready when a COVID-19 vaccine becomes available, and before the next pandemic strikes. We will also use the funding to test the accuracy of our solution for toddlers. In the long run, contactless biometrics will help strengthen health systems by digitising health records, and as a consequence, make better data available to influence decision making.

The AI revolution has accelerated the potential of mobile-phone camera-based biometrics like facial recognition, and has the potential to become a powerful tool for impact organisations trying to reach the most vulnerable populations. However, progress has been hampered by the crippling racial, ethnic, and gender biases found in camera-based AI algorithms: a US government study released in December 2019 found that Asian and African American people were misidentified up to 100 times more frequently than white men by the world's top facial

recognition systems. In addition, the vast majority of solutions currently available focus on Western markets: tools require online systems or high performance computing tools, which cannot run without reliable internet, or on low-cost Android phones typically used in last mile settings. Furthermore these providers do not offer implementation services, instead relying on their customers to have a high level of technical ability in order to incorporate the SDKs within their mobile or desktop applications. Primarily developing products for self-use (e.g. authentication for financial services using a personal, high-powered smartphone), these companies are not tailored to support frontline service delivery programmes that work at the last mile for some of the most vulnerable and marginalised populations. 

For example, when a COVID vaccine becomes available, effective and equitable distribution among vulnerable populations will be of the utmost importance, to ensure that no vaccine is wasted, and that no lives are needlessly lost to this pandemic.

Simprints’ aim is to reverse the current Western data "bias" within camera-based AI, as well as the widespread oversight in protecting beneficiaries' privacy rights and data security, to provide a powerful, ethical, and inclusive biometric identification system at scale.

The end goal is for large impact organisations to use our solution to reach their most vulnerable populations who are currently "falling through the cracks" of their service delivery programmes worldwide. In time, we will have set the standard for similar AI-powered biometric technologies for delivering services in global development.

We would use the AI for Humanity Prize in two ways:

1. To prepare Simprints’ tech stack to allow our AI solution to be used effectively in  national-scale vaccine programs, retooling the infrastructure to allow improved syncing of biometric data 

2. To ensure that beneficiary privacy is not compromised when biometrics is used for pandemic response, by creating enhanced data protection and privacy policies, audit logs, and risk registers specifically for pandemic response/vaccination programs