WeCARE is a Machine Learning based health portal to promote value based care in the Healthcare industry.

The Project was made for Optum Stratethon Season 4

We aim to counter the problem of Reducing hospitalization (and re-hospitalization) risk through the early intervention program.

WeCARE uses various machine learning models to predict diseases based on data generated by Synthea.

Webapp Link - https://stratethon.web.app/

Machine learning Models - https://github.com/phoenix-aditya/UpCare-models

Youtube Link - https://youtu.be/SxTXiWJ3ikY

• Lack of data driven solutions to support healthcare professionals & patients therefore hindering them from making more informed decisions based on accumulated patient data
• Lack of a data pipeline to aggregate data scattered across various relational databases thus preventing development of Machine Learning & AI based solutions to existing problems
• Lack of a unified interface that provides analytics and predictions about patients in an effective manner
• The problem of severe costs & added complication faced by patients due to late diagnosis of diseases
• Lack of application to provide patients with tips & measures to enable them to make better health decisions

• UpCare is a unified portal that provides analytics and predictions on the basis of data accumulated by existing health records and input from patients
• UpCare has built in data pipelines for easier integration with existing health record databases which promotes easier code maintainability and further development
• UpCare uses multiple Machine Learning models to predict if the user has a tendency to develop
  • Heart diseases
  • Thyroid related problems
  • Location based allergies
• UpCare provides various analytics and graphs related to a patient’s existing medical records in an easy to use interface

• Our Solution empowers patients & doctors to get health predictions about a user based on Machine Learning models trained on vast amount of data
• The Analytic provided by the portal is data accumulated from various sources about a patient therefore providing complete information to healthcare professionals
• Predictions by our models enables users to seek medical attention before the problems aggravate therefore saving them money and reducing complications due to late diagnosis
• UpCare provides an easy to use portal that enables healthcare professionals to access medical data in a quick and efficient manner in situations where time is crucial

• Dedicated machine learning models for specific use-cases to provide accurate results
• Models Trained on synthetic data generated by Synthea which enables us to make powerful models over using real noisy datasets.
• Prebuilt data pipelines for data in format represented by Synthea for easier integration to existing solutions
• Codebase in MERN Stack & Python which promotes better code maintainability & easier further development
• Easy to use interface

We predict top 5 possible allergies that the user can develop on the basis of geolocation data generated by Synthea and therefore enable user to take measures before developing serious symptoms

The Allergies are predicted on the basis of the following data:

1. age
2. latitude
3. longitude

The geolocation based allergy model also enables users to know any allergies they might develop during their travels or if the change their home town.

We predict the possibility of Heart Diseases of a particular patient by obtaining data from the healthcare database and also some data inputed by the user.

We suggest doctor visits to the user based on the results of the machine learning model.

The model predicts data on the basis of the following data:

1. age
2. sex
3. severity of chest pain
4. cholestrol
5. blood pressure (systolic)
6. heart rate

We predict the possibility of Thyroid diseases on the basis of factors thus enabling user to take precautionary actions.

The model predicts the category of Thyroid that the user will be in on the basis of the following data:

1. Age
2. Sex (0-female, 1-male)
3. On_thyroxine (0 for false, 1 for true)
4. Query_on_thyroxine (0 for false, 1 for true)
5. On_antithyroid_medication (0 for false, 1 for true)
6. Sick (0 for false, 1 for true)
7. Pregnant (0 for false, 1 for true)
8. Thyroid surgery (0 for false, 1 for true)
9. I131_treatment (0 for false, 1 for true)
10. Query_hypothyroid (0 for false, 1 for true)
11. Query_hyperthyroid (0 for false, 1 for true)
12. Lithium
13. Goiter
14. Tumor
15. Hypopituitary
16. Psych
17. TSH
18. T3
19. TT4
20. T4U
21. FTI