During the time of Covid-19, all the hospital workers are having a tough time getting account of all the patients. I started out this project with the simple idea of having a record of all the on bed records but it expanded to a real-time parameters finder and using rest API and jQuery along with MongoDB, excellent database, I was able to make a custom REST API using https request, so all common details like heartbeats, heart rate, blood pressure can be sent to the server using any capable kind of computation device.
This web-based tracker for hospital patients was made with situation like covid-19 in mind that would allow easy access to hospital professionals to interact with their particular cases. A level-4 encryption and decryption was used for security using both hash and decrypt. The future scope of the project would be to implement server to node communication for IoT implementation making managing much simpler. Later on, all the stored value of parameters were given to the ML model for training using github action on daily basis. The github action functionality of providing timed action make it easier to manage the whole model updating. This model is later used to predict if the certain patient is gonna have a predicted unstable days so they can be given a better observation. To make this reach from servers to mobile location, the balena-Fin provided to me by hackster challenge for fighting covid-19 came into action. However, once I am able to get my hands on the replacement of BalenaFin and some other hardware component this project will be expanded further.
Working Of Software Side:The web application allows for storing data of individual patient using any device to access a secure web page protected by bycrypt encryption. All the official information can be stored here as per the hospitals requirement. The user authentication details are stored in a local encrypted server so won't be accessed by anyone outside the circle of allowed authority. The data is shown in form of a table with all other important details in the home page. It has a live refresh block for retrieving the updating parameters, such as blood pressure, pulse rate etc.
The above page will later on redirect to personal page of each patient by clicking the see more info, this is achieved using the node.js and express functionality. This prevents having huge number of pages and traffic for similar data and provides much faster rate of transfer. The parameters, such as blood pressure are stored in form of an array and is updated every day recording the rate of recovery. This data is sent to the python script using github action. This allowed simple machine learning algorithms, in this case Autoregulation along with a few keras layers were used to predict the unstable days. The update was sent to the patient personal page as an alert.
The hardware sector of this project is still on hold due to the issues I am facing with the balenaFin but once that arrives I am aiming to get hold on a kit for measuring blood pressure and sending the data via REST API requests which can be accessed and worked upon by the server. the future scope of this system is still on hold. Other device I am planning to integrate include thunderboard Sense V2. This section of the project will be updated as for now I have been simulating the data on python script with ocr on an actual electronic blood pressure machine.
Comments
Please log in or sign up to comment.