NABDA-bg
NABDA-logo

Nabda

Inspiration

The health-care system is always growing and discovering new methods to make everyone's life better and safer, but the alarming aspect is that when we look at the top ten deaths, medical errors come in second. More investigation reveals that many of them were caused by false or missing information. For example, in an emergency case, the person is female with blood vessel disease, and she is in a critical state. First, the ER gave her blood thinners, but she is not getting them, and she is actually getting worse, causing more bleeds and eventually her death.The inquiry revealed that she suffered from Himophilia, and providing her blood thinner was a deadly mistake.
Couldn't this be avoided? What is the use of having tons of archive hanging around hospitals if we can't use them efficiently at such a vital time? That's what prompted us to create Nabda and avoid similar tragedies.

What it does

An Electronic Health Record system that can create, gather and manage medical records, and make it accessible to authorized clinicians. A centralized system in which each patient's account is displayed with all of his medical information automatically sorted into different categories and medical fields.
Medecins, on the other hand, will have a dashboard for all of their patients and information, allowing them to make faster and more precise judgments. It also provides a new approach to submit descriptions, which can extract all relevant medical information.

How we built it

Nabda was separated into three major sections. The first was our front end, which was designed to be as user-friendly as possible in order to facilitate both physicians' and patients' work. This was accomplished through the usage of React as the front framework. We chose Flask for a Restless API that will supply the essential information for the app and maintain its data. The final section is Models, because there are no benefits to maintaining such a large amount of data in digital form only for the sake of keeping them, we made sure to use them as effectively as possible, thus we created two models. The first was used for an organization purpose, to classify all papers based on their nature and the medical area, while the second was used to make the doctor's work easier by auto completing forms from his text using entity recognition.
Docker was used to containerize all of them for simple deployment and administration.

Challenges we ran into

The hurdles were substantial, both in terms of technical issues and the essence of the notion.
Making models that can categorize such diverse materials, in addition to medical domains, appears to be an insurmountable challenge. In addition to that, when seen in the context of the notion we are attempting to implement, changing an entire system opens the door to a slew of new obstacles, including legal issues, privacy concerns, how to manage permissions, and how to deal with all of the old records.

Accomplishments that we're proud of

Reaching the stage of fixing all aspects of the application and creating an architecture that not only replaces the present health care system with something that is far better, faster, simpler, and more efficient on paper is already a significant accomplishment.
Another accomplishment is the use of entity recognition with medical terminology to automatically extract all relevant information.

What we learned

The learning value is actually quite high, because we are dealing with models that we do not normally use, as well as dealing for the first time with such a large target with enormous complexity, which in our case is electronic health records.

What's next for Nabda

Although we have made headway with Nabda, we are only at the beginning, the concepts that can be implemented are far from exhausted.
Our ultimate objective is to create a peer-to-peer system in which all hospitals are linked to one another and exchange profile data in order to enhance the health-care system.
The idea is to leverage the data by developing models that analyze it and anticipate future prospective diseases in order to warn the user for early detection.