As a team, we were inspired to solve a global issue by mitigating the impacts of the aging population. It is estimated that by 2030, over 1 billion people will be aged 65 years or older (National Institute of Health). Older population is also more prone to developing chronic illness. At least 80% of Americans aged more than 60 have at least one chronic illness which is also the leading cause of death among people 65 and older (CDC, 2017). In addition, we are also experiencing global shortage in trained health care workers and the imbalance in supply and demand will only get worse over time. It is projected that the global deficit of healthcare workers will worsen from 7.2 million in 2013 to 12.9 million by 2035 (WHO). By now we can all see the potentially disastrous outcome in the overall health of the elderly population in the near future. We need a tool that can address the issues of growing healthcare demand from the aging population and the limited availability of healthcare workers. In order to alleviate this problem, we created the Guardian Angel.

What it does

The Guardian Angel is an AI powered patient monitoring system that does not require patients to use any advanced devices like smartphones. All that patients need to do is pick up a phone call and start conversing with the Guardian Angel.

How it works: Guardian Angel will initiate a call to patient A at a scheduled time Guardian Angel follows up on the patient of their overall health and asks for any new or changing symptoms. It may also ask for the most recent BP and body weight readings. While the patient is answering, the Guardian Angel is constantly analyzing the patient's speech and picks up on any symptoms that require further investigation by the healthcare professionals If the patient mentions any alarming symptoms, the Guardian Angel will notify the nurse and the nurses will take a look and triage them if they need an appointment with a physician for further investigation.

Guardian Angel enables three things that the traditional telemedicine could not: Any landline phone can provide service to the elderly without access to advanced technology Enhanced automation enables extraction of medically relevant information provided by patient Conversion of speech to text facilitates electronic health record documentation

How I built it

We used the following tools to build Guardian Angel:

  • Flutter
  • Flutter packages :
  • url_launcher, firebase_auth, provider, highlight_text, avatar_glow, and more
  • Google Cloud Speech-To-Text API
  • Google Cloud Firestore
  • HTML, CSS, JavaScript (WebFlow)

Challenges I ran into

We were first planning to use Twilio Programmable Voice to implement the automation of phone calling, until we found out that flutter_twilio_voice has limited support across OS. That was when we came up with an alternative for Twilio, the Flutter url_launcher package. Another challenge we encountered was enabling voice recognition in android emulators: while there was no problem with our code, the emulator was not detecting, thus pouring a lot of effort in making it functionable. Building an algorithm to detect words that may hint illness was another challenge.

Accomplishments that I'm proud of

We believed that the most efficient way of showing a patient’s health status was by presenting patient information and data extracted from Guardian Angel on a doctor’s tablet. We never had an experience building an application with flutter but we tried to learn the skills as much as we could in a limited amount of time. At the end, we felt very proud that we successfully created a proof-of-concept solution to a serious problem that needs to be addressed.

What I learned

From this experience, we learned a lot about challenges that the current healthcare system encounters, and what difficulties elderly patients face with. We also learned that with small changes to the healthcare system, it could bring much more valuable experience between elderly patients and doctors.

What's next for Guardian Angel

  1. Refining machine learning to improve accuracy in important data extraction
  2. Support for multiple languages
  3. Introduction of “small talk” function to alleviate feelings of social isolation among elderly population

Built With

Share this project: