Our team identified two intertwined health problems in developing countries:

1) Lack of easy-to-obtain medical advice due to economic, social and geographic problems, and

2) Difficulty of public health data collection in rural communities.

This weekend, we built SMS Doc, a single platform to help solve both of these problems at the same time. SMS Doc is an SMS-based healthcare information service for underserved populations around the globe.

Why text messages? Well, cell phones are extremely prevalent worldwide [1], but connection to the internet is not [2]. So, in many ways, SMS is the perfect platform for reaching our audience in the developing world: no data plan or smartphone necessary.

What it does

Our product:

1) Democratizes healthcare information for people without Internet access by providing a guided diagnosis of symptoms the user is experiencing, and

2) Has a web application component for charitable NGOs and health orgs, populated with symptom data combined with time and location data.

That 2nd point in particular is what takes SMS Doc's impact from personal to global: by allowing people in developing countries access to medical diagnoses, we gain self-reported information on their condition. This information is then directly accessible by national health organizations and NGOs to help distribute aid appropriately, and importantly allows for epidemiological study.

The big picture: we'll have the data and the foresight to stop big epidemics much earlier on, so we'll be less likely to repeat crises like 2014's Ebola outbreak.

Under the hood

  • Nexmo (Vonage) API allowed us to keep our diagnosis platform exclusively on SMS, simplifying communication with the client on the frontend so we could worry more about data processing on the backend. Sometimes the best UX comes with no UI
  • Some in-house natural language processing for making sense of user's replies
  • MongoDB allowed us to easily store and access data about symptoms, conditions, and patient metadata
  • Infermedica API for the symptoms and diagnosis pipeline: this API helps us figure out the right follow-up questions to ask the user, as well as the probability that the user has a certain condition.
  • Google Maps API for locating nearby hospitals and clinics for the user to consider visiting.

All of this hosted on a Digital Ocean cloud droplet. The results are hooked-through to a node.js webapp which can be searched for relevant keywords, symptoms and conditions and then displays heatmaps over the relevant world locations.

What's next for SMS Doc?

  • Medical reports as output: we can tell the clinic that, for example, a 30-year old male exhibiting certain symptoms was recently diagnosed with a given illness and referred to them. This can allow them to prepare treatment, understand the local health needs, etc.

  • Epidemiology data can be handed to national health boards as triggers for travel warnings.

  • Allow medical professionals to communicate with patients through our SMS platform. The diagnosis system can be continually improved in sensitivity and breadth.

  • More local language support



Share this project: