Timeline of various medicines with potential hazards
Potentially hazardous combination
Pharmacy locator and automated ordering
Have you been sick in the last year? Can you tell which drugs you have taken? Chances are you won't be able to recall all medicines you have taken. And why does it matter?
In the US, 1 in 12 teenagers who take several medications together experience severe side effects, usually because those drugs where incompatible with each other. Women who take birth control pills can actually get pregnant if they take an incompatible medication which happens to decrease effectiveness of the pill. Unintended interactions between medicines is both an individual health concern and a public health problem.
What it does
While technology has streamlined many aspects of our lives, healthcare is lagging behind when it comes to digitalization. Our app removes friction from the moment your doctor hands you a prescription. But more importantly, it keeps track of medicines you are taking and warns you about potentially problematic drug combinations. You can order medicines online, or check drug compatibility before buying them over the counter. Finally, by showing your medicine history to medical professionals they can better understand your condition, prevent accidental mistakes and provide you with better care.
How I built it
Using advanced image recognition technology and natural language processing, we digitalize and process hand-written doctor prescriptions on the fly. Our sophisticated algorithms pull in various data sources to predict potentially hazardous combinations of drugs.
Challenges I ran into
On the technical side, applying NLP on handwritten notes and medicine packages was not a trivial task as well as visualizing and presenting our findings in a simple yet intuitive manner. Additionally, narrowing down our pool of ideas into a coherent storyline was not a simple task.
Accomplishments that I'm proud of
Forming a cross functional team and delivering a proof of concept that provides customer value.
What I learned
VanillaJS is not that much more simple to spin up a web app :)
What's next for pharmHouse
More sophisticated training of our models for image recognition, combining more data sources like medical products from aboard and eventually branching out into disease and symptom associations for a holistic medical approach.