Project Name : CleverCare
Reducing Hospital Readmissions
- One of the major contributors to healthcare costs is hospital readmissions. Research shows that 15-20% of people who are discharged from hospital are readmitted within 30 days or less. Many of these readmissions are preventable. There are some intervention techniques that can help address this issue. One of them is to use machine learning to predict the chances of readmission of a patient. The project facilitates out-patient care to take feedbacks at regular intervals from patient once he/she is discharged and calculating his/her chances of readmission using machine learning. The doctor can review these results and take appropriate action by providing some suggestions and educational content to the patient, and reschedule patient’s next feedback.
-
As a doctor, I can view feedback and prediction index of readmission for my patients, so that, I can decide measures to be taken to mitigate chances of readmission.
-
As an admin, I can view analysis of readmission data, so that, I can take preventive measures to reduce unnecessary expenses by avoidable readmissions.
-
As a Out Patient Department nurse, I collect feedback from recently discharged patients, so that received data can be fed to prediction engine and predict possibility of readmission for patients.
- As shown above, the different components of system are connected with each other.
- For example, when a user submits credentials in order to log into the system, the browser forwards this request to node.js client.
- Further, the node.js client communicates with node.js server using MOM-RabbitMQ.
- The node.js server validates the credentials against encrypted credentials stored in mongoDb database.
Sample Nurse Credentials:
username: bryony@clevercare.com
password: test
Sample Doctor Credentials:
username: shannon@clevercare.com
password: test
Sample Admin Credentials:
username: admin
password: admin
A detailed description of the project with proper screenshots and explaination can be found here.
-
Install mongodb on your machine, start mongodb service
-
Install node.js on your machine
-
Install rabbitmq on your machine, start rabbitmq service
-
Install python 2.7 on your machine
-
Download spark MLlib 2.0.2, extract, paste everything in ML directory into spark-..*/bin/ directory
-
Run spark-submit diabetes_prediction_program.py from command line
-
Start server.js file from project/clevercare-server/ directory
-
Start www file from project/clevercare-client/bin/ directory
-
your application will available on localhost:3000
This project is released under the MIT License.