Reminder Care

Conectando Salud: Recordatorios Médicos y Citas para el Bienestar de la Tercera Edad, Todo a través de una Llamada.

Comment

Inspiration

Noncommunicable diseases (NCDs) kill 41 million people annually, representing 74% of all deaths worldwide. These diseases affect all age groups and regions but are often linked to older age.

Medication adherence is crucial to avoid complications such as increased adverse effects, drug resistance, and morbidity. According to the World Health Organization (WHO), non-adherence is a major barrier to realizing the full benefits of medication, with adherence in chronic disease patients in developed countries at only 50%.

Key barriers to medication adherence include:

  • Patient-Specific Barriers: Lack of information, motivation, and understanding of medication regimens.
  • Illness-Specific Barriers: Misunderstandings and poor awareness about medication necessity.
  • Healthcare and System-Specific Barriers: Issues related to healthcare delivery and patient support.

Patient-focused solutions like reminder apps can help, but not all users, especially older adults, have access to or are comfortable with smartphones.

What it does

It allows a doctor to register patients and add data such as treatments, upcoming appointments, age, and contact information. This can be added manually or automatically by uploading a photo of the prescription; a computer vision algorithm extracts the text from the image and uses keywords to automatically obtain the information.

This information is then transferred to a database, after which another program uses the information to schedule automated calls with personalized messages for each patient. These messages help remind patients of important aspects of their treatment, such as medical recommendations, medication schedules, or upcoming medical appointments.

How we built it

Aquí tienes la traducción al inglés del texto proporcionado:

First, a database was defined in MongoDB using Python and JavaScript code. Subsequently, a web page was created using Flask, React, and Node.js. This page handles user registration and login via Auth0 for authentication. Once credentials are validated, a menu is displayed that allows the selection of the data entry type. Data can be entered manually using a form implemented in JavaScript or through text extraction from an image (medical prescription). This extraction is carried out using a text recognition model implemented in MATLAB, which removes illegible characters and retrieves important information using keywords such as patient name, medication, age, etc.

This information is stored in the database. Separately, we created a program in Python that retrieves data from the database and generates personalized messages for each patient. We use a scheduler to calendarize the calls according to the need and Twilio to make these calls automatically. The messages are generated using a date system to obtain reminders. The Chat GPT API can be used (tested, but currently not operational due to the costs involved) to obtain personalized health care recommendations.

Challenges we ran into

Throughout this project, we faced various challenges, such as integrating MATLAB with Python, and making calls from a script in Python, which was something we had never done before. However, there were two parts that were particularly challenging for us:

Firstly, identifying a problem and choosing how to solve it. Implementing the text extraction algorithm in MATLAB, as it required preprocessing using dynamic thresholding.

Accomplishments that we're proud of

We feel proud of the progress we made with a project that we believe could have a significant social impact, and as a team, we were able to overcome challenges

What we learned

Many of the technologies used here actually required us to learn how to use them. However, we would like to highlight the use of APIs. We learned to utilize them for process automation, for interfacing MATLAB with Python, and for user authentication.

What's next for Reminder Care

The system is intended to be improved by implementing a finer system in image processing, IA. a stronger network with a lot of training that allows the most efficient and reliable conversion to text possible, On the other hand, the AI that recommends a healthier lifestyle could be changed by one that is free of cost and also helps older adults to have a healthier old age with these recommendations and without stopping taking the indications that their doctor indicates, for its part, in the academic area these databases could be used to make a relationship between the ailments that adults have and the type of medications they consume and their vital data every time they have appointments.

Share this project:

Updates