Inspiration 🧠
How many times have you had to make a trip to the doctor's office, but didn't feel like your needs were truly being met? Did you feel like you were connected to the wrong professional? Did figuring out insurance leave you anxious and tired? Each year, hospitals across the United States receive 145.6 million patients per year. From confusing treatment plans to communicating with administration, many patients and families are bewildered by the many tasks they must figure out to receive proper healthcare services. Patients are often left not receiving proper treatment and struggling to fund their needed treatment. That's why the average readmission rate for US hospitals is roughly 14%. In fact, the readmission rate for individuals receiving Medicaid was almost twice of individuals using private insurance.
| We created AdvoHealth to lessen the burden on patients to receive proper healthcare. For those who want better experiences at the hospital, for those that want more options for their treatment, and for those who could use a helping hand, AdvoHealth was made for you!
What's the deal with this app? 🖋️
AdvoHealth helps users in multiple ways. It uses artificial intelligence to generate a list of hospitals that are not only accessible to patients, but also have better track records on readmission rate, mortality rate, and safety measures. In addition, we connect users with patients advocates to aid them in the processes they need to secure essential services. AdvoHealth gets patients into the highest quality clinics and hospitals in the lowest amounts of time.
Because the only thing that patients should have to worry about is recovering 😷
How we built it
We used python to create the k-cluster analysis and rank the hospitals. The libraries pandas, numpy, matplotlib, sklearn, urllib, and requests were all used in this process. React, Material UI, and Node JS were used on the front-end.
Challenges we ran into 🚧
We faced multiple hurdles along the way. Being separated by different time zones, we each completed our ends of the project at different times, making is harder to collaborate during the process. Also, during the earlier part of the 24-hour window, we focused quite a bit of time on getting a better idea of the project.
- Run Time Error - K-means cluster analysis was used instead of hierarchical agglomerative cluster analysis due to the large dataset we worked with. K-means cluster analysis had a better time complexity of O(n^2) while agglomerative analysis had a time complexity of O(n^3).
- Data of Hospitals - Due to the sheer number of hospitals, we could only use a limited number of hospitals. We decided to use a database that only accounted for hospitals that accept Medicaid insurance.
Research 🌐
- Hospital Dataset - Link
- Disease Prediction Dataset - Link
- Facts on the State of the Healthcare system - Link
Accomplishments that we're proud of
- We are very proud to have connected the front-end coding with the back-end
- We are proud that we designed and finished a viable business product within 24 hours amidst all the difficulties we faced.
- Training a model with 100% accuracy! Link
What we learned 🎓
Taking naps was our best friend to finishing our complicated and functional project on time
- We learned to use the sklearn python module for the cluster analysis, Logistic Regression, GBC, RandomForestClassifier, and DecisionTreeClassifier .
What's next for AdvoHealth 🤖
_ AdvoHealth has a lot of great capabilities to be used in the future. _ Healthcare is a very important field that a lot of people take very seriously. We will expand our impact by adding more functions to the app and include more hospitals in our analysis. This will allow more opportunities for our users and patient advocates to better understand the quality of their healthcare options.
_ A helping hand is exactly what people need in order to have a better experience in the healthcare system! _
We are participating in the 18+ division for the healthcare track
Log in or sign up for Devpost to join the conversation.