Orchid

Inspiration

The existing healthcare systems often struggle to provide timely & cost-effective solutions, especially in resource-constrained areas. The lack of accessible and affordable healthcare services poses a significant barrier to addressing the global health burden accompanied by following challenges:

Non-communicable diseases (NCD) including CVD, diabetes etc. continue to be a significant health concern globally. NCDs account for approximately 71% of all deaths worldwide, with CVD being the leading cause. Access to early detection of NCD remains limited in resource-constrained areas.
Mental health disorders affect people of all ages and have a substantial impact on individuals and societies. Globally, around 1 in 4 people will experience a mental health disorder in their lifetime. However, the treatment gap for mental health is significant, with approximately 50% of people not receiving the mental health services they need.
Malnutrition, including both undernutrition and micronutrient deficiencies, contributes to increased morbidity and mortality rates, impaired cognitive development, and reduced productivity.
Limited healthcare facilities, scarcity of specialized expertise, and inadequate diagnostic tools hinder the early identification, precise diagnosis, and effective treatment of these health conditions.

Detailed Analysis of the Problem: Link

What it does

Orchid is an AI-powered healthcare web & mobile application, aims to provide accessible and affordable healthcare services to people worldwide. The application includes NCD's diagnosis tools, AI diet planner, mental health diagnosis, ASD diagnosis for age groups 1 to 3 and AI self-diagnosis, utilizing state-of-the-art technologies like TensorFlow & Scikit-learn. The mobile app is built with Flutter, while the web application is built with Django. Orchid's loosely coupled architecture allows for easy integration, making it a versatile solution to meet the specific needs of healthcare providers. The application also includes features where users can maintain their medical profile, analyse their diagnosis history, book appointments, read blogs, report bugs, and much more.

Features

Non Communicable Diseases (NCD) Diagnosis: The burden of NCDs affects individuals, families, and healthcare systems, emphasizing the need for accessible and cost-effective solutions. Orchid incorporates advanced machine learning algorithms to predict the occurrence of NCDs like cardiovascular diseases, diabetes, liver disease, and brain tumors. By offering precise diagnosis and personalized treatment plans, Orchid enables early detection & accurate diagnosis, leading to improved health outcomes.
Personalized Nutrition for Optimal Health : Orchid's diet planner is an innovative feature designed to provide personalized and AI-driven meal plans. It leverages advanced algorithms and machine learning techniques to create customized diet plans tailored to individual needs and goals. The diet planner takes into account various factors, including an individual's age, gender, weight, height, activity level, dietary preferences, and any specific dietary requirements or restrictions. By considering these factors, Orchid's diet planner ensures that the generated meal plans are well-suited to each user's unique nutritional needs.
Mental Health : Orchid's mental health tool, powered by TensorFlow, goes beyond diagnosis and prediction by providing users with a comprehensive mental health score and valuable suggestions for support. By leveraging TensorFlow's capabilities, Orchid analyzes input data related to behavioral patterns, symptoms, and other factors to generate an individual's mental health score. In addition to the mental health score, Orchid's tool utilizes Google Cloud SQL to fetch a curated set of suggestions and helpline information. These suggestions encompass self-help techniques, coping strategies, and resources tailored to the individual's mental health needs.
Autism Spectrum Disorder for Toddlers: The ASD diagnosis tool for age groups 1 to 3 serves as a valuable resource for parents who suspect their child may be on the autism spectrum. Our ASD detection tool is a cutting-edge application designed to aid in the early identification of autism spectrum disorder (ASD) in children aged 1 to 3. Powered by the Random Forest algorithm and built with Scikit-learn, this tool leverages machine learning to predict ASD traits based on a comprehensive set of behavioral features and individual characteristics. Benefits:
Beyond Diagnostics: The inclusion of features like maintaining medical profiles, analyzing diagnosis history, booking appointments, and accessing informative blogs adds further value to the application. These features in Orchid brings significant benefits to the overall healthcare experience:
Medical Profile Maintenance: By allowing users to maintain their comprehensive medical profiles, Orchid ensures that essential health information is easily accessible and securely stored. This feature addresses the problem of limited accessibility to medical records, especially in resource-constrained areas.
Analysis of Diagnosis History: Orchid's capability to analyze diagnosis history provides valuable insights into an individual's health journey. This feature allows users and healthcare professionals to track progress, identify patterns, and make informed decisions regarding treatment and interventions. It overcomes the challenge of limited accessibility to diagnosis history, contributing to improved continuity of care and better health outcomes.
Appointment Booking: By offering a convenient and user-friendly platform for booking appointments, Orchid ensures that individuals can access the care they need in a timely manner, reducing delays and improving overall healthcare access. Leveraging the power of Gmail SMTP in the backend, Orchid automates the appointment booking process by sending automated emails to both the user and the expert.
Access to Informative Blogs: Orchid's provision of informative blogs offers a valuable resource for individuals seeking information and guidance on various health topics. This feature helps address the problem of limited access to reliable health information, empowering users to make informed decisions about their health and well-being.
Accessibility & Affordability: This comprehensive approach facilitates precise diagnoses and offers a one-step solution for both healthcare experts and common users, enabling efficient and accurate healthcare services. The problem statement highlights the challenges of accessibility and affordability in healthcare. The mobile app is built with Flutter, while the web application is built with Django, providing a seamless user experience across platforms

How we built it

Backend
Web Framework: Django Framework,Python
Mobile Application: Flutter, Dart, WebView
Database: SQLite (Local), MySQL (Production)
Frontend
Languages: HTML, CSS, Javascript
API’s : Gmail , Bootstrap, Google Fonts
AI and Machine Learning
Tensorflow, Scikit-Learn, Random Forest Classifer, Random Forest Regressor, Gradient Boosting Classifer, Gradient Boosting Regressor, Support Vector Machine, Pandas, Numpy, Seaborn, Matplotlib, Joblib
Google Cloud Platform
Cloud SQL for Database
Google App Engine: Deployment Server
Cloud Storage: Hosting Static & Media Files, Machine Learning Models
Others
Anaconda for Package Management
Powersell for command line operations
Software Requirements: Atom IDE,Visual Studio Code , Android Studio, Jupyter Notebook, Google Colab

Detailed Design Document & Architecture: Document Link

Challenges we ran into

Deployment Challenge:
During deployment, we encountered a challenge with the Cloud SQL Database. Django default SQLite database was not compatible with GCP. Thus we switchd to Google Cloud SQL database for our production. However, after a day of deployment, we noticed that our cloud credits were being used at an alarming rate, primarily due to the default configurations of Cloud SQL. After researching alternative solutions, we managed to configure the database with the minimal required settings to host our website, thus avoiding additional cost overheads. This experience helped us to learn how to manage Cloud SQL more efficiently in the future.

AI Integration Challenge:
One of the major challenges we faced was integrating a variety of AI models into our application. As using heavy models in the backend would have resulted in longer response times, we came up with a solution to serialize each model as an object. This allowed us to use the trained model in the backend directly without having to retrain it repeatedly for making decisions. We used the joblib library to serialize the models and stored them in Google Cloud Storage. This approach enabled us to seamlessly integrate all the models into our application while maintaining optimal performance.

What's next for Orchid

Partnerships with Healthcare Providers: Collaborate with healthcare providers such as hospitals and medical organizations to integrate Orchid's platform into their existing systems. This partnership would enable healthcare professionals to recommend and use Orchid as a supplementary tool for diagnosis, treatment and patient engagement.
Collaboration with Government and NGOs working in the healthcare sector to leverage their networks, resources, and expertise. Engage in public-private partnerships to facilitate the adoption of Orchid in public healthcare systems and expand its impact on a larger scale.
Multi-language Support: This would involve providing language options for users to access Orchid's features and content in their preferred language, thereby enhancing inclusivity and accessibility.
Data Privacy and Security: Implement robust encryption techniques, comply with relevant data protection regulations, and regularly audit the platform's security infrastructure to safeguard user data.
Blockchain-Based Identity Verification: Incorporating blockchain-based identity verification mechanisms can enhance user authentication and eliminate the risk of identity fraud. By utilizing blockchain for identity management, Orchid can provide a secure and trusted environment for users to access the platform's features and services.
Immutable Health Records using blockchain: Storing health records on a blockchain can ensure their immutability and accessibility across healthcare providers. Orchid can explore partnerships with healthcare organizations and leverage blockchain technology to create a unified and secure health record system. This would enable users to have control over their health data and easily share it with healthcare providers, leading to improved continuity of care.