Living through the COVID-19 pandemic, we've all gone through the minor annoyances associated with transitioning our everyday lives into an online format. However, there are certain experiences which translate extremely poorly into the virtual space. One of our team members volunteers at a Philadelphia clinic to help patients there gain access to social services, including therapy and support group sessions for recovering victims of alcoholism, drug abuse, and mental health disorders. Working with these patients in the past few months, he's heard firsthand how these services have severely deteriorated in quality after their attempts to transition online. Events such a zoom-bombings have led to concerns over privacy, and video conference calls simply cannot offer the same level of intimacy as in-person sessions. This means that these individuals are cut off from one of the few safe spaces they had to speak about their experiences. Combined with the emotional and financial stress being felt worldwide due to the pandemic, it's no surprise that both mental illness-related suicides and mortality related to drug overdoses have significantly increased over the past few months. We created Telesafe as a way to bring back these safe spaces to these individuals, enabling them to reconnect with their network, which is a vital part of their treatment and recovery. With our domain, we want to provide a safe space for those in quarantine to continue getting the support they need and promote mental wellness.

What it does

Telesafe is a closed-network medium to connect with other members of an individual's support group. Members will be invited to join a given group by the group facilitator, and any identifying information remains hidden to all other members unless both the individual and the group facilitator wishes to make that information available for each of the members. Members will be able to chat, share updates, share photos, and video-call under the purview of the facilitator, and the facilitator can also place members into smaller groups to create a more intimate experience. Of particular interest to facilitators, especially clinicians, is the use of the Google Cloud Natural Language API to analyze each participant's general sentiment and mood over time, as determined from the language of their posts. The mood level of each participant is displayed to the facilitator in a simple, easy to understand manner in a dashboard that also allows the facilitator to keep track of and schedule appointments.

How we built it

Given the complexity of the features required for such a solution (e.g. video conferencing), our main goal given the time constraints of MedHacks was to provide a visualization of what Telesafe would look like and a functional implementation of the Google Cloud APIs. Website mock-ups were designed in Figma, a popular UI design tool, after considering the various features and information that we researched that would benefit both members and facilitators. We were also able to create a working site that demonstrated the usage of sentiment analysis in the facilitator's dashboard.

Stretch goals for this project include creating a functional site to better demonstrate the experience of Telesafe's users including a full dashboard and profile with metrics for facilitators and clinicians in understanding their patients at a glance.

Regarding the implementation of the sentiment analysis, we are utilizing Flask to connect the Python script which passes the relevant text data to Google Cloud Natural Language API via an HTTP request and retrieves the asynchronous response back (via JS promise) with the JavaScript logic in our site to display the correct visual representation of each patient’s sentiment in the doctor dashboard. This visual representation was implemented using a color scale with red, yellow and green, to convey patient mental welfare at a glance.

Challenges we ran into

Given that none of us has extensive experience working on web development, many of the technologies utilized in this project represented our first time using them. As a result, we inevitably faced several challenges getting our project into the current state. We had some trouble setting up API access to Google Cloud as the tutorial we followed unfortunately was out of date, but we managed to pull through to get an end-to-end implementation of the sentiment analysis.

Another issue that we faced is that, since we are utilizing Flask, we need to deploy our web app to a service like Google Cloud App Engine, AWS Elastic Beanstalk, or Heroku. We experienced some difficulties in deployment and unfortunately we were not able to deploy the project in time. However, we believe that our demo will successfully demonstrate the plan for the project and address many concerns regarding its feasibility.

Accomplishments that we're proud of

While we were unable to fully flesh out a working implementation of the website given the time constraints of the hackathon, we believe that we succeeded in developing an solid plan and feature set for Telesafe that addresses many of the problems we identified with current solutions as well as several of the concerns associated with a solution like ours.

We were able to further cement this through detailed mock-ups of the user interface of Telesafe that reflect the careful design choices that were a result of our research into support groups and medical privacy.

Finally, we are proud to have demonstrated a use case for Google Cloud Natural Language in our project by using sentiment analysis as a way for doctors to easily glean information of each of their patients in a simple manner.

What we learned

From the technical perspective, we learned a great deal on web development as a whole in addition to experience in several technologies that we previously never worked with such as Google Cloud.

But more importantly, we learned a great deal about the medical field, particularly regarding privacy of patient information and support groups in general. As a platform for support groups, Telesafe must ensure that patient information remains secure and do its best to keep data (posts, video chats, etc.) to stay within the site for the sake of patient privacy and trust.

What's next for Telesafe

Moving forward, we'd like to finalize all the various features of Telesafe, such as the secure video conferencing. Additionally, we'd like to test Telesafe with a focus group of patients and doctors to receive feedback on how we can improve the user experience. We'd also like further research to be done on using the data from the Google Cloud Natural Language API for diagnoses, as we believe this holds a lot of potential.

Built With

Share this project: