For Massachusetts General Hospital (MGH) Neurologist Dr. Mark Albers, research into the loss of smell as an early warning sign for the onset of Alzheimer’s had yielded powerful evidence.
Dr. Albers’s collaboration with ADK on developing a smartphone-based Alzheimer’s smell testing application coincided with the onset of the COVID-19 pandemic in early 2020.
Emerging research into COVID-19 had also revealed a correlation between loss of smell as an early symptom.
So Dr. Albers approached ADK about whether their ongoing collaboration on Alzheimer’s smell test research could apply to developing a COVID-19 smell test app. It was clear the world needed low cost, noninvasive, and scalable applications for COVID-19 detection, and the partnership with ADK could help to realize a viable approach.
Further collaboration between Dr. Albers and ADK could revolutionize COVID-19 early detection through a smell testing application. But the limited funding and grant application windows would create major challenges to creating an application that:
Essentially, the challenge was designing, building, and testing the application with limited funding and within two weeks to support a pilot study.
Dr. Albers and the ADK team began working in April 2020 via Zoom and phone calls to develop the project plan and define the MVP. The immediate goal focused on app usage for research with a longer-term aim of a clinical application and eventual commercialization.
With a pivotal research study as the primary goal, the short timeline meant that the teams would need to work quickly but carefully.
ADK and Dr. Albers determined that much like their approach to the Alzheimer’s smell test, the COVID-19 app detection process would comprise:
Initial meetings covered types of smell cards, eConsent flows, and defining the branching logic that would determine test subject research study routing.
Dr. Albers’s original Alzheimer’s research had been built using ADK’s mHealth App Kit as a base. The Clinical Data Collection Platform included the core features needed for a research study while enabling the development of customized smell test support.
The fact that ADK was responsible for the design, development, and strategy enabled us to build the app quickly. Third-party component developers IFF, MFR Samplings, and Arcade Beauty would enter the project at the right juncture to provide scent development, pilot study scent cards, and odor labels.
The ADK team and engineers partnered with a consulting data scientist from MGH to develop the machine learning (ML) detection algorithm and integrate it via API.
This algorithm was the key to aggregating the smell test answers that would provide user results as determinate or indeterminate of smell loss related to COVID-19 infection. The app sends all results to a HIPAA-compliant AWS cloud instance connected to a clinical research dashboard for review.
ADK simultaneously ran design and development sprints with the first version of the app completed by April 13, less than two weeks after kickoff. The following week, a Spanish-language version was completed and made ready by April 20th. By the end of the month, the teams had launched:
The completed application then went into the R&D pilot study testing phase with limited and controlled use among study groups. Study participants and others could view the smell test process on the COVID smell test website.
Up to 400 patients at MGH, Brigham Women’s Hospital (BWH) and the Spaulding Rehabilitation Hospital have already taken part in the pilot study.
The study was a phenomenal success, demonstrating that the app was >90% effective in accurately identifying the presence of COVID-19 in participants.
This project’s timeline would not have been possible without the mHealth App Kit, which also fueled the original Alzheimer's study thanks to its:
While ADK worked predominantly with Dr. Albers, his clinical support team included 20+ medical professionals and researchers. Once the pilot study phase began, ADK took on additional roles of supporting research grant development and studying business opportunities. This included market strategy and exploring production possibilities to move from research to scalable mass market manufacturing and delivery, which was led by ADK and mHealth App Kit CEO, Dan Tatar.
Fast timelines such as this are often part of meeting potential research study windows of opportunity. But the unprecedented nature of the COVID-19 pandemic made these timelines even tighter. ADK and the MGH team shared a sense of urgency to meet changing parameters while developing a system and structure that is solid enough for clinical use, but flexible enough for commercial use.
As longtime collaborators with Dr Albers, MGH, BWH, and other hospitals within the Harvard Medical school affiliate orbit, ADK’s defining approach was as a full life cycle partner. The entire team sees the COVID-19 smell test app's potential for future growth while helping to push the world closer to a day when the pandemic is under control.
Partnerships are vital to digital transformation in an unpredictable age of healthcare where solutions must embody speed, accuracy, simplicity, low costs and maximum efficacy. MGH and ADK have poised the smell test app to provide expansive and broad data sets for research and testing possibilities that pave the way to that future.
Dr. Mark Albers created a novel, non-invasive, and cheap COVID-19 test with >90% accuracy using a smell test.
Developing a low cost, noninvasive, and scalable application for COVID-19 detection within two weeks.
An accelerated Agile process to fully customize the mHealth App Kit, while partnering with exceptional organizations like IFF.
Launching the app within two weeks meant that up to 400 patients at MGH, Brigham Women’s Hospital (BWH) and the Spaulding Rehabilitation Hospital took part in the pilot study. The results have shown a >90% accuracy in successfully identifying COVID-19.
