Inspiration

WNBA player Asia Durr continued to experience severe symptoms more than 8 months after she first tested positive for the virus, and hasn’t been able to shoot a free throw, let alone train. She had to drop out of the 2020 WNBA season as a “medical high-risk player’’ A 32-year-old from Louisville, Kentucky had the first symptoms on March 12, 2020. For the next five months, she had 16 emergency department visits and 3 hospitalizations. Furthermore, she feels as if she’s in a brain fog These 2 patient stories are just a tip of an iceberg and as per CDC’s interim guidelines (published June, 2021), the frequency of long-term symptoms and conditions following COVID-19 varies widely in the literature, ranging upto 80%. Majority of the COVID-19 long haul symptoms like fatigue, Brain fog etc. are seldom considered serious by physicians, and as a result, these COVID-19 long haulers continue to be underserved and undertreated, both medically and otherwise. The long-haul symptoms are spread across a wide spectrum of indications, some of which can lead to severe chronic complications. Hence, we propose “Post COVID Care”, a mobile application that can help the long haulers self-assess their condition at regular intervals and guide them regarding the need of medical intervention.

What it does

“Post COVID Care” app helps the users to self-assess their current medical condition, including vitals, symptoms and impact on daily functionality due to the symptoms. Based on this information, the app suggests the users if they should consult a doctor and guides them to the nearest physicians.

How we built it

The backend was built using Flask and SQLite database. RESTful APIs were created to login, register and add details for a given patient, while SQLAlchemy was used to facilitate create, read, update and delete operations on the database. Marshmallow was used to serialize the objects to and from the database. The frontend part of the project was built using Kotlin in Android Studio. In Android Studio we used Retrofit client to consume and parse endpoints provided by the API. Clinical inputs were used to streamline the building of this app.

Challenges we ran into

Definition of long haulers is still vague and is very subjective. This created challenges in defining the parameters for identification of the user as potential COVID-19 long-hauler. Arriving at the precise thresholds for each of the parameters and then building the function to classify the patients based on these parameters, was challenging yet fun.

Accomplishments that we're proud of

We are glad that we could come up with a comprehensive criterion to identify COVID long haulers, based on their symptoms, comorbidities and their functional status using exhaustive secondary research and in-house clinical expertise. What makes our app unique is its ability to quantify the impact of the disease on the everyday life of the patients, beyond the binary outcome of mortality or survival.

What we learned

COVID-19 long haulers can present in myriad ways and we need to have an open mind to identify these patients as these conditions are new to the medical community. On the technical front, we learnt that microframeworks such as Flask, and database agnostic tools like SQLAlchemy are flexible enough to accommodate a plethora of use cases. They provide us the leeway to collect disparate datapoints, process them in the background and return a meaningful response to the user. The use of Kotlin greatly reduced our burden as it is faster to compile, lightweight, and prevents applications from increasing size. The code being in a concise language means fewer chances of both runtime and compile time errors.

What's next for Post Covid Care

The app can further evolve to provide a detailed ‘self-management regimen‘ to patients who do not require expert medical intervention. It can suggest a weeklong routine to the user, based on diet, exercise, sleep, rest and relaxation, or cross reference to other wellness apps and initiatives. The demographic and clinical data collected by the app can be used to build datasets for IQVIA, in addition to being leveraged by the larger research community to advance our understanding of the pathophysiological trajectory of COVID.

Share this project:

Updates