The Feelix Diagnostic Stethoscope

I am a current PhD candidate at Johns Hopkins University and Chief Technology Officer at Sonavi Labs. I have a background in electrical and biomedical engineering, focusing on low-cost hardware, machine learning algorithms, and novel materials for solutions to problems facing non-traditional clinical settings.

I began working on what would become Feelix while an undergraduate at Johns Hopkins University studying Electrical Engineering and Romance Languages. Under Professor James West, I began working on a solution that could reduce pediatric mortality as a result of pneumonia. After many years and through my graduate work, I developed the Feelix smart stethoscope, which uses patented adaptive noise suppression and classification algorithms to provide diagnostic support to clinical and patients. 

As CTO, I have been motivated by the ability to build the technology to market that could directly impact lives and ensure that the technology reaches the people who need it the most.

Respiratory diseases impact over 600 million lives globally. Pneumonia is the leading cause of death in children under five; asthma is the most common chronic condition of childhood; COPD is one of the leading causes of death in adults. Poorly managed and undiagnosed conditions affect individuals and drain economies of countless resources, and often disproportionately affect minority and underserved communities. These problems have only been exacerbated by the likes of COVID-19. 

Feelix listens and analyzes body sounds with the same accuracy as a trained pulmonologist. The device provides clinicians and community health workers with real-time diagnostic support. It can also provide patients with a simple and easy-to-use tool to track and monitor their chronic conditions, and telemeter data to their providers when needed. Feelix is adaptable and can be used in any environment due to a proprietary adaptive noise suppression that has been proven to collect the highest quality recordings.

Health systems and frontline workers lack the tools necessary to adequately and efficiently coordinate and manage healthcare for large populations. The technology needed to comprehensively assess and manage patient health, especially in underserved populations, has not existed due to limitations on resources, complexity, and clinical applicability. Our current dramatic shortage of trained physicians also leads entire populations to lack access to quality health care or struggle to maintain consistent communication with care teams. These issues and deficiencies result is hundreds of millions of people left without healthcare and trillions of dollars in already limited resources wasted.

Common chronic conditions like asthma and COPD are far too often uncontrolled, resulting in numerous exacerbations, excessive hospital visits and even death. The misdiagnosis of acute conditions like pneumonia and tuberculosis is resulting in millions of lives lost annually, and a rise in antimicrobial resistance, further draining resources and making it more difficult to combat these diseases.

All of these challenges could be mitigated with AI as an on-the-spot clinical decision support tool that could quickly and accurately screen patients, and that would allow patients to screen themselves wherever they are to manage their chronic diseases, gain insights, and optimize their care.

The Feelix devices have embedded AI that detects abnormal lung sounds to help more accurately classify respiratory infections and mange the health of chronic disease patients. In addition, Feelix gives patients the ability to communicate with primary care providers remotely to manage chronic conditions from the comfort of home without sacrificing the integrity of in-person visits.

Feelix makes it possible to obtain high fidelity lung sounds in any environment and incorporates three advantages compared to competitor digital stethoscopes: a larger sensitivity area, active noise suppression, and onboard detection of abnormal lung sounds.

Adaptive noise suppression removes ambient noise from screenings, which are recorded and immediately analyzed by an embedded classification algorithm. The process mimics a trained physician identifying abnormalities by listening to a patient's lungs through a stethoscope and making a determination based on the conditions present. Feelix can differentiate wheezes from crackles, quantify and grade severity, and provides immediate feedback to facilitate more informed decisions using real-time, objective data. A wider sensitivity area makes it easy for anyone to use Feelix to collect high-quality recordings.

We integrated supportive apps and a secured cloud to allow patient screenings to be recorded, analyzed, shared and stored for further review or comparison. 

The Feelix solution is intended to support patients in every community around the world. We know that respiratory diseases do not discriminate based on geographic location so we developed a solution that is universal in design and user-friendly so that we could make it as ubiquitous as the thermometer.

Feelix can support respiratory disease patients no matter where they live, and provide a digital health assessment tool that allows for rapid screening to identify potentially fatal abnormalities. We have ongoing studies in the US, Peru, Bangladesh, Belgium, and are working to expand our sites to ensure that our data is unbiased and that we can support wider patient populations and combat more diseases.

This project began in an effort to provide better resources for community health workers (CHWs) to screen pediatric patients for pneumonia in extremely remote and rural regions. To that end, highly trained clinicians, frontline CHWs and supportive organizations have been integral to our development and have provided critical feedback to enhance the usability and functionality of our devices and software.

We also have studies that deploy devices directly into patient homes, allowing them to engage with care teams remotely, while maintaining the integrity of an in-person visit.

Feelix is designed to address disparities in health care delivery, specifically in respiratory diseases. By providing easy-to-use and low-cost screening and monitoring tools, we aim to ensure that everyone has equal access and opportunity to quality healthcare and timely interventions.

This project began because too many children globally were losing their lives every year to a treatable and detectable condition. In the US, minority children are disproportionately affected by respiratory diseases. There are severe gaps in our society's ability to provide adequate care to every individual and Feelix can be the tool to elevate the health of our global community. 

This project began in 2014 when Dr. James West was approached by global health specialists to help address pediatric mortality as a result of pneumonia. As an undergraduate student at Johns Hopkins University, motivated by the promise of free food, I attended a lecture that resulted in a long conversation with Dr. West, who convinced me to join the research team, along with Dimitra Emmanouilidou and Mounya Elhilali. As a team, we discussed the issues addressing global health workers and the limits of current digital auscultation. As a result of those discussions, we developed the foundational technology at the core of Feelix.

I would take the helm of the research and continue to work on Feelix throughout my graduate work at JHU, refining the device and algorithms to run in real-time. I eventually met Ellington West, Dr. West's daughter, who at the time was a Sales Director for a healthcare organization. Ellington and I began to formulate a strategy to found Sonavi Labs and work to deploy the technology to every corner of the globe, to increase access to the healthcare for everyone, especially those who have traditionally been left behind.

The fact that one child dies every 39 seconds from pneumonia, a detectable and treatable disease, should compel everyone to action. This work is important to me and my team equally because we all understand the simple premise that no human being should lose their life because they did not have access to care, especially due to racial disparities in healthcare delivery.

Dr. West, my mentor, is an African-American octogenarian from Farmville, VA, who overcame segregation and racism to invented the electret microphone, the cornerstone of modern acoustics. My mother is an immigrant from India, who lost her sister to pneumonia when she was young. My mother went on to be a Director at one of the largest healthcare organizations in the US. Their accomplishments highlight the severe issues that we face globally and the continued need for equal access to healthcare. 

I have always wanted to find a way to make meaningful change in the world, and even contemplated going to medical school following my undergraduate studies. Dr. West and my mother made me realize I could make an impactful in the healthcare space globally through my work, and hopefully spark a new era of modern diagnostics.

My background is in computerized lung sound analysis, biomedical engineering, and solutions for non-traditional clinical settings. Feelix is the culmination of six years of research with some of the most accomplished and qualified scientists and researchers in the field. Sonavi Labs is the result of a collaboration between those experts and successful and talented business leaders.

Being at Hopkins has allowed access to the worlds leading medical system and public health institution, and our technology would not be what it is today without our close working relationship. Our proximity to this renowned healthcare institution has enabled additional research collaborations both domestically and internationally.

The work at Hopkins culminated in four patents ranging from classification algorithms to novel sensing materials. This year, I was granted five additional patents through Sonavi Labs based on the work to develop and realize Feelix into a remote monitoring platform for all patients. We designed Feelix to be continuously upgraded and to evolve as we continue our research.

I have been recognized as a leading voice in modern digital diagnostics: I have received several NIH grants, I was asked to present at a WHO sub-committee, and I was invited to participate in an industry roundtable for digital diagnostics with officers from organizations like Philips, Siemens, and Roche.

Feelix has offered me the opportunity to travel the world and meet users from every walk of life. Because I speak several languages, I have been able to interact directly with the patients for which Feelix was ultimately designed. 

In 2017, after years of research, we had our algorithms on data gathered in 9 different countries and they worked well in controlled settings on laboratory computers. We had committed to delivering 25 units to Malawi and Bangladesh when we discovered that when deployed on the device, the 'real-time' classification algorithm would not run in under 30 minutes on the device. After countless hours of debugging, I realized that I had to completely redesign the algorithm from the ground up. After weeks of rapid coding and quick clinical validation, I was able to design and deploy an entirely new algorithm to the device that would run in 22 seconds. I also took the opportunity to attempt new methods, which subsequently resulted in a higher overall accuracy. We successfully deployed the devices that resulted in the formative clinical validation of Feelix.

After successfully overcoming the technological hurdle, we have faced many logistical hurdles in widespread adoption of the technology. Primarily, the project is consistently met with skepticism by the clinical community. We have met these challenges with rigorous clinical designs and continue to pilot the device to prove the value of the device in clinical settings. 

I firmly believe that leadership ability is a learned trait that needs to be honed properly, not some innate skill. As the cofounder of Sonavi Labs, I am often reminded of the first experience that required me to be a leader. As an undergraduate, I joined Engineers Without Borders working on water projects for rural communities in Guatemala. As one of the few Spanish speakers, I assisted the Community Liaison and went on implementation trips. On my very first implementation trip, the Community Liaison had a last minute emergency, so we were stranded in the country without them and with a tight deadline. The engineering team looked to me to for leadership for the remainder of the trip. Though overwhelmingly nervous, we successfully implemented Phase I and were able to meet our future deadlines.

I was forced into a leadership position; it was not self-selected. It was also in a time of crisis, during an international trip, with a lot at stake. As an officer of Sonavi Labs, I frequently find myself in similarly difficult situations and am grateful for the learning opportunities that projects like EWB gave to me so that I may become a better and stronger leader. 

The Feelix smart stethoscope is an innovative solution that seeks to empower every clinician, and lay person with the ability to identify a potentially harmful disease. The devices are simple and effective, and we have studies that show implementation leads to higher accuracy and more appropriate treatments.

Our competitors are making strides to add electronic features to traditional stethoscopes, however, they lack the attention to acoustic quality, and the sophistication to deploy analytics directly onto the hardware. This allowed Sonavi Labs to create a device that anyone can use anywhere, empowering clinicians of all skill levels, while other systems find themselves limited by environment or usability.

Feelix can bridge the gaps in access to timely diagnosis during a pandemic, by giving minimally trained community health workers an advantage in supporting the most vulnerable and traditionally under-resourced communities. Slowing the spread of a pandemic requires rapid deployment of healthcare workers and they need to be armed with advanced tools that can support rapid patient screening. If we are able to collect data now on COVID-19, we may be in a position the next time it spreads to catch it more quickly.

We are working to drive the cost of our hardware down so that we can afford to give them away. Feelix has the ability to impact millions of lives and if we can make the device more accessible, then we can not only make the technology better, we can make whole communities healthier.

With a quick, easy and rapid solution that deploys real-time analytics we can multiply the capacity of the frontline, saving time and lives.

Our theory is very simple, in that if we empower frontline healthcare workers with advanced tools that rapidly and accurately detect diseases, then patients will be more likely to receive appropriate treatments, and thus better outcomes. Even further, we are developing population level analytics capabilities into our smart cloud to identify outbreaks before they spread. 

The organization and utilization of quality data has stifled researchers and frontline workers from optimizing practices and standardizing patient screenings. Our theory is that we can change the course of patient treatment by starting with the very first interaction then use data to intervene at the community level.

We know that one of the major challenges caused by the shortage of healthcare professionals is an over-reliance on World Health Organization (WHO) guidelines, which are notorious for leading to the over-administration of antibiotics. Anti-microbial resistance is on the rise and will cost health systems countless resources and is a detriment to patient outcomes. Early data from some of our recent studies show that by implementing the Feelix platform, the accuracy of community health workers to identify pneumonia increased by 22%. If we model this out across diseases like Tuberculosis and even COVID-19, not only will we save countless resources by limiting the need for traditional wet diagnostics or x-rays, but we can also save millions of lives.

Sonavi’s approach is automated and removes the need for human interpretation of respiratory sounds. Our plan is to evolve the way diseases are detected and managed by using simple, universally accessible tools. Patients will have access to tools that are just as intelligent and powerful as the clinicians, but simple enough for them to understand.

We believe that every choice in healthcare has a consequence, and AI will be the key to ensuring the critical choices of frontline healthcare workers are more informed, objective and accurate. Further, we hope our population level analytics will help health systems and governments make more informed choices to protect citizens.

Since founding the company in 2017, we have launched over 150 devices into 5 different countries, including the US, Belgium, Nigeria, Peru, and Bangladesh. Community health workers have screened over 10,000 patients in total and we are excited about the opportunity to scale our platform to touch millions of lives in the next five years.

It is difficult to pinpoint the exact amount of patients screened by Feelix devices, but if we assume the 150 devices are with community health workers who screen an average of 2,000 patients per year, we can estimate that this year, we will screen approximately 300,000 patients. Studies show that even confirmed diagnosis using a radiographic chest x-ray range in accuracy of 47% - 75%, while the Feelix devices consistently benchmark above 90%.

If we continue to scale our distribution at the current rate, which is at a rate of 3x increase in the number of devices being distributed each year, in the next year, we will distribute between 3,000 and 5,000 devices, screening between 6 and 10 million patients.

In 5 years, we hope that we will have made the hardware extremely affordable and accessible, such that we will have distributed between 500,000 and 1,000,000 devices through distribution and research partnerships. In the event of the next pandemic, we believe that Feelix devices can screen billions of potential patients.

Completing the regulatory process is a top priority for us, and we have already begun the process with the FDA. Within the next year we plan to complete the regulatory process for our hardware and classification algorithms. We have decoupled our hardware from the software for the FDA submission in order to reduce the complexity of the review and minimize the likelihood of complications. The classification algorithm will be submitted following the approval of the Class II 510(k) hardware, and we will simultaneously submit the combined hardware and software application to receive CE marking from the appropriate notifying bodies. Completing both regulatory processes in the next year will enable broad distribution of our platform.

In the next five years, we plan to complete the regulatory processes for other major markets to scale the Feelix platform even further. Countries like Brazil do not abide by FDA or CE marking but offer significant opportunities for our solution. We also plan to validate our classification algorithm on the broad spectrum of respiratory diseases beyond our current validation for pneumonia and asthma. With the launch of an @Home version to support patients, even those in remote communities, and a wearable device, we plan to dramatically strengthen the lines of communication between patients and their healthcare providers. Easing the burden on disease management will have a drastic impact on the communities most devastated by chronic respiratory conditions.

Disparate regulatory processes are certainly going to present a challenge as we aim to engage ministries of health around the world. We expect to encounter widely varied reactions from government agencies and consumers as we understand different markets have different priorities and motivations.

We are currently building relationships with the US Commercial service agency as well as distributors on the ground, who will help guide us through regulatory bodies. The US Commercial Service has also provided due diligence reporting at a very nominal fee, to ensure that our interests are being protected as we engage international agencies.

The major driving force behind our optimization efforts will be resources and data collection partners. While we have been able to secure a substantial amount of grant funding, each clinical trial is limited to a single disease. We need the support of private and public health systems to fund validation studies and pilot programs, which will help to eliminate bias from our data as well as providing distribution opportunities international markets.

Our current price per unit would create a barrier for low- and middle-income markets as well as low-resource areas in high-income markets. However, we have identified engineering, logistical, and funding pathways to increase the capabilities of our device while reducing the cost. Current clinical studies are providing data that will enable our researchers to enhance the capabilities of the classification algorithm. 

In parallel, we intend to conduct engineering work to optimize the device for cost and power through hardware redesign and software optimization to decrease the manufacturing costs of the device by half while improving algorithm design.

Countries like India present significant opportunities for data collection but the average consumer in the market cannot bear the cost at current. This is why we are also actively seeking partnerships with global health organizations, impact funds, and community health initiatives who are seeking diagnostic equipment to support frontline workers. Organizations like the Gates Foundation and our partners at Johnson and Johnson, among others will be able to provide funding to supply devices in areas of high need and low resource.

Sonavi Labs is currently a part of the Johnson and Johnson Innovation, JLABS program. This opportunity is exposing Sonavi Labs to researchers within the J&J network, in addition to providing resources to help us grow and scale the company. J&J is working to support our research as a leader in the Tuberculosis space, we have been connected with leaders of the Global Public Health initiative.

We have also partnered with FIND, a leader in the diagnostic space for global health and a WHO affiliate, to deploy the device and validate for Tuberculosis use. 

We have begun a pilot program with the INOVA Fairfax Lung Transplant team. The Feelix device is being used to support patients in their post-operative care, which requires numerous follow-up visits. Many of these patients travel several hours each way to visit with care teams for a 15 to 20 minute visit. Physicians have the ability to customize notifications and device parameters, recognizing that everyone's base line is different. Their goal is to more deeply understand patient trends while avoiding excessive alerts.

Sonavi Labs has also partnered with LifeBridge Health system in Baltimore, MD to develop additional clinical trials and pilot programs that prove the impact of remote monitoring solutions in managing chronic diseases for minority patients.

Additionally, we have partnered with the mHealth Lab at the Bangladesh University of Technology to expand our algorithms to identify cardiovascular abnormalities. This partnership is being supported by the TIGER IT Foundation following our win in the MIT-SOLVE TIGER Challenge.

We believe that solving global health challenges will require additional partnerships and funding from various sources. Our mission is to do well as a company, through strategic business model and a holistic approach that offers solutions to the entire healthcare ecosystem. We have identified B2G, B2B and B2C opportunities for device sales, subscription services, and data analytics contracts.

Feelix, FeelixPro and Feelix@Home devices will be distributed at varied price points, while maintaining healthy margins to support wide sweeping data collection initiatives, and further enhancement of the classification algorithms. Bulk contracts will be sought with non-profits and global health  organizations. We estimate significant profits through reimbursement plans in high-income markets where the FeelixPro and Feelix@Home platforms will support patients with chronic conditions like asthma and COPD.

Subscription services will be offered to these patients to benefit from longitudinal analysis. We have predictive analytics algorithms and a health-privacy compliant cloud system that aggregates patient data and provides a useful tool to clinicians.

The cloud also offers an opportunity to capitalize on data analytics services provided to health systems, payers and government organizations. We plan to use population level data to help better evaluate the spread of diseases and where to best allocate resources.

We plan on pursuing every possible avenue toward achieving financial stability. Our business model includes multiple revenue streams, but as a research company, we will continue to pursue grant funding in addition to venture funding to help scale our solution.

We will continue applying for grant funding to support research and further enhancements of our classification algorithm and patient monitoring platform. We have so far been very successful in applying for NIH funding, and hope that our research will yield results that warrant continued funding. We have also identified grant opportunities with USAID, the US military and other institutions that are seeking to invest in and deploy novel innovations.

We will pursue venture funding to help us scale the company. VC's will provide the institutional support we will need as entrepreneurs to ensure that we are maximizing capacity and filling in operational gaps where needed. Leveraging VC networks will enable us to build a sustainable infrastructure and acquire top talent.

We will also generate revenue through device sales, subscription services and data analytics contracts. Our model includes multiple streams of revenue which include government contracts, bulk device orders for health systems, community organizations and insurance organizations.

At the inception of this project, over $3 million in grant funding was provided by NIH, NSF, NASA and the Gates Foundation. This funding supported the early research, which yielded the proprietary adaptive noise suppression algorithm and classification algorithm.

The founding team contributed approximately $500,000 in funding to develop a commercially viable product, which is currently in review with the FDA.

Approximately $2 million has been raised in a seed round, which offers convertible notes.

We have raised approximately $1.6 million in grant funding from the NIH to support research in Asthma and COPD. The Offices of Minority Health and Health Disparities as well as the National Heart, Lung and Blood Institute have contributed this funding, through Phase 1 and 2 funding vehicles.

In 2019 approximately $35,000 in revenue was generated from device sales to research partners. In 2020, we anticipate approximately $300,000 in revenue will be generated from device sales in Bangladesh, Turkey, and the US.

Sonavi Labs has a strategic plan to maximize all funds raised in order optimize the hardware and software, and scale the business operation into multiple markets. This will include seeking grant funds, equity investments and partnerships to support inventory production and deployment, especially across low and middle-income countries.

We will be closing our open seed round upon meeting our current goal of $2 million or the completion of the FDA review of the Feelix hardware, which is expected by the end of Q3 2020. This round offers convertible notes to qualified investors.

Our current funding provides enough runway through Q3 of 2021 and will support the first round of production of hardware. We anticipate opening a Series A round of approximately $5 - $7 million in Q2 of 2021, which will be an equity round.

In addition to seeking institutional funding from venture firms, we plan to continue applying for grant funding, prized innovation, social impact, and minority-focused competitions to propel the research of the company and our reach in traditionally underserved communities.

We estimate that expenses in fiscal year 2020 will be approximately $2.7 million.

Funding has been allocated to support additional clinical trials that will serve to enhance the features and function of the Feelix devices. Funds will be used to support device production in these research initiatives. 

Resources will also be used to support subsequent engineering updates to the software and further development of the Feelix apps. We have designed apps to support the many users of the Feelix, FeelixPro and Feelix@Home devices, and will need to further develop these applications as we launch the hardware.

Additional funding has been allocated to legal expenses and consulting fees to protect intellectual property. As we engage with international markets, this has been a strategic priority and one that we hope will protect the company in the long run.

Funding has also been allocated for marketing and sales efforts to update the Sonavi Labs website in preparation for the commercial launch of the products.

I am applying to the Elevate prize because I believe that this opportunity will tremendously accelerate our ability to scale the technology and get it into the hands of the people who need it the most.

We created the Feelix device to save the lives of children fighting pneumonia. As we developed the technology and built a platform around it, we realized there were so many more applications for this device and software than we had previously imagined.

What we need now is for the device to collect more data so that we can train our algorithm to detect more diseases and support broader populations. The Elevate prize will help us do that by elevating our ability to scale, not only through additional financial resources, but also through the vast MIT-SOLVE network.

We hope to continue building partnerships with research institutions, pharmaceutical companies, health systems and public health agencies.

Research collaborations are always warranted as we believe that by feeding our classification algorithm more data, we can train it to detect and differentiate even more diseases in more body systems. 

We are seeking partners to deploy devices into research trials as well as on the frontline with community health workers. Deploying our platform is not going to be an easy task.  Despite the universal design of our device, we need people on the ground with familiarity in the regions to help with training and ongoing support.

Not only is the Feelix hardware capable of impacting millions of lives on the ground, but as we collect more and more data, we will be able to provide useful population level data to help health systems and government agencies allocate resources and capital more efficiently.

The TATA Center - Richard Fletcher and Christian Infante

- The TATA Center at MIT seeks to create affordable pulmonary disease diagnostic tools. We would be happy to collaborate on research and implementation projects with the TATA center.

Access Afya - Melissa Menke

- This mobile screening solution would be a great partner to support pediatric patients in rural and under-resourced communities. We hope to build this functionality into our supportive app.

Blue Sky Analytics - Abhilasha Purwar 

- Uses AI to provide air quality and emissions data, which would be important for supporting predictive analytics for chronic respiratory diseases patients. We aim to incorporate this data into our base station in order to provide predictive support to chronic disease patients.

Ada Health - Hila Azadzoy

- AI powered digital decision support platform for community health workers would be a powerful tool to incorporate into the Feelix community health worker program.

E-Heza - Wendy Leonard

- A low-cost, quality driven health record for families in low-income countries would be a great partner to ensure patients in remote and rural communities have access to longitudinal trend data to better manage chronic respiratory diseases.