Feebris

An AI-powered mobile health platform for diagnosis and monitoring in the community.

Health systems worldwide are centred around a sparse and unevenly distributed resource – doctors. Yet, globally there is a health workforce crisis, with a shortage of 7.2 million healthcare professionals. These workforce shortages result in restricted access to care, lack of personalised and regular assessment, and delayed identification of diseases and exacerbations. Consequently, millions of young children are dying unnecessarily from treatable conditions such pneumonia (nearly 1 million deaths annually).

Feebris is an AI-powered mobile platform that puts the brain of a doctor into the hands of communities. Feebris’ AI-engine uses machine learning applied to point-of-care devices to interpret large volumes of medical data and perform diagnosis & monitoring. Feebris’ first application is diagnosis and monitoring of respiratory conditions in the most vulnerable patients – young children and the elderly. The platform empowers minimally trained users, such as community health workers, to perform evidence-based diagnostic services.

The Feebris platform has three main components: (1) off-the-shelf sensors, (2) mobile app, (3) cloud infrastructure. First, the platform connects wirelessly to affordable off-the-shelf devices, like digital stethoscopes and wearables, that can capture health measurements in the community. Two, the AI-engine interprets these health signals to identify pathologies and fuse them into diagnosis of disease or exacerbations. The AI-engine combines a range of signal processing and machine learning algorithms trained against large datasets, delivering evidence-based outcomes. Third, equipped with a cloud infrastructure, the platform can share community-generated insights with the health system and drive continuity-of-care.

The Feebris platform is a disruptor for three main reasons: (1) it captures objective clinical measurements via point-of-care devices in the community; (2) it uses powerful machine learning algorithms to perform evidence-based diagnosis & monitoring, with accuracy comparable to that of a doctor; (3) it fuses large volumes of data to perform holistic health assessment, including co-morbidity analysis. Moreover, the platform is hardware agnostic, connecting to sensors that are most appropriate for the local market, ideally locally manufactured. 

The Feebris platform is innovative for three main reasons: (1) it captures objective clinical measurements via point-of-care devices in the community; (2) it uses powerful machine learning algorithms to perform evidence-based diagnosis & monitoring, with accuracy comparable to that of a doctor; (3) it fuses large volumes of data to perform holistic health assessment, including co-morbidity analysis. Moreover, the platform is hardware agnostic, connecting to sensors that are most appropriate for the local market, ideally locally manufactured. Finally, the platform enables a new layer of healthcare - AI-powered community care, delivered by carers & community health workers

The Feebris platform has three main technology components: (1) off-the-shelf sensors, (2) mobile app, (3) cloud infrastructure. First, the platform connects wirelessly to affordable off-the-shelf devices, like digital stethoscopes and wearables, that can capture health measurements in the community. Two, the AI-engine interprets these health signals to identify pathologies and fuse them into diagnosis of disease or exacerbations. The AI-engine combines a range of signal processing and machine learning algorithms trained against large datasets, delivering evidence-based outcomes. Third, equipped with a cloud infrastructure, the platform can share community-generated insights with the health system and drive continuity-of-care

Over the next 12 months we aim to develop our platform to commercial readiness and start generating revenue. In Q4 2018, we are launching a 12-month programme in Mumbai, India, funded by the Children’s Prize, with 10,000 children to prevent 200 pneumonia deaths. Data collected will refine our algorithms, and throughout we’ll add features with user input. By the end of the 12 months we aim to have submitted our application for CE certification as a diagnostic tool.  

Over the next five years our vision is to grow the functionality of our solution to include: frailty, malnutrition, cardiac conditions, and childhood development. By adding these conditions, we will be able to help vulnerable patients manage a wider range of complex conditions and co-morbidities. For children in low-resource settings, this will mean access to a more holistic community-centric tool for early detection. These problems are global, so towards the end of this period, we also aim to scale our solution to new markets, initially the US and Rwanda, where we have established health networks.

The target market for our child health application are community organisations in emerging markets. These organisations operate large networks of community health workers who lack the medical skills and tools to perform evidence-based diagnosis and rely on observational guidelines to identify children at risk. Our technology enables community organisations to upskill their staff and turn them into a powerful healthcare workforce. We have been developing the platform alongside several such organisations (e.g. Apnalaya in India)

Over the last 2 years, our solution was developed and prototyped via global health studies. In January 2018, we were awarded the prestigious Children’s Prize to deliver our innovation to 10,000 children in Mumbai’s urban slums over 12 months. We are currently setting up the launch of this large-scale intervention

In Dec 2018 we will be launching our child health programme in India via a Mumbia-based NGO, Apnalaya, currently serving 1.2m people living in urban slum communities. Their community health workers will use our platform to perform evidence-based diagnosis of respiratory conditions. Each health worker will be equipped with a mobile phone and a set of devices which they will use to conduct examinations in the community.

We are simultaneously pursuing relationships with larger global health organisations to enable implementation at scale. We estimate that in 3 years, our technology will be used by 2,000 workers to reach 750,000 children.

Elina is a biomedical engineer and data scientists with expertise in global health. After 5 years of academic research at Oxford University, Elina founded Feebris to develop technology that democratises access to healthcare. She has previously worked as a global health and innovation consultant for several organisations including the World Health Organisation.

Adam works on business strategy, operations, financial modelling and business development. He has experience in operations and project management on multi-million-pound engineering projects and strategy consultant for multinationals. More recently, he ran another start-up, gaining expertise in sales (B2B and B2C), marketing and branding.

We will initially operate a B2B SaaS model via non-governmental organisations in India which operate networks of community health workers. These organisations want our platform to provide new services and to upskill staff. In India, we charge $27/month per health worker – this cost will either be passed on to the patient per appointment (in India 70% of healthcare expenditure is out of pocket) or donors, at significantly lower cost than current lower-quality alternatives.

By 2020, our algorithms will be CE certified for diagnosis and wearables will be integrated. We will also have collected evidence on the clinical efficacy and health economics. At this point we will start selling direct to consumers and also integrating with public health institutions. Based on our forecasts we should be profitable by year 5.

After we’ve generated sufficient data, we will also be exploring monetisation of the anonymised health data, with consent from our users.

The Solve ecosystem would be truly transformative for Feebris, enabling access to key implementation partners as well as alignment with the health delivery value chain. We are building a hardware agnostic platform, partnerships with manufacturers of physiological sensors are key for the growth of our platform. Additionally, we are pursuing partnerships with providers across the value chain (from primary care to pharmaceuticals) to ensure our solution facilitates continuity of care and delivers value for both patients and the health system. 

Our regulatory strategy involves certification for the European market (EU) via CE. This includes the development of: (1) a quality management system (QMS) compliant with medical devices standards (e.g. ISO13485); (2) a technical file; and (3) a clinical report on the performance of our algorithms in a real clinical setting. Getting our AI engine certified is a big barrier to entry. Both funding and access to experts will enable us to overcome this barrier.