ITcats developed a device for diagnosing diseases in the field of urinary disorders in patients. The project included the development of the software and hardware part, as well as training of the neural network to be able to make a diagnosis in automatic mode.
Customer
Industry
Healthcare
Region
Poland
Client since
2021
Our client, a company in the field of development and production of medical equipment, operates in Poland. The company specializes in the development of equipment for disease diagnostics and patient health monitoring. Detailed information about the client cannot be disclosed under the terms of the NDA.
Challenge
According to a study, urinary dysfunction occurs in 52% of women. Lower urinary tract symptoms occur in 15-60% of men, depending on age. Uroflowmetry is the most preferred method of examining the functional status of the lower urinary tract, as it is a relatively simple, non-invasive, yet informative and can be used as a screening method.
Our customer wanted to develop two variants of uroflowmeter: household and professional with additional equipment and advanced software features. It was assumed that both variants would be able to diagnose using a pre-trained neural network and be able to store data on the customer’s or medical institution’s server.
Solution
Our team developed the design of the device according to the specified requirements of the customer and current medical standards, designed and implemented the architecture of interaction with the device by means of remote wireless access on the part of both the patient and medical staff.
The developed solution allows the product owner to control unauthorized use and copying of devices by implementing a hardware part activation server. At the same time, all personal data can be stored on a dedicated server of a medical institution, as required by law. The developed architecture is universal for household and professional versions of uroflowmeters. The scenario of household use of the uroflowmeter provides for the use of the uroflowmeter by the patient at home independently, by means of interaction with it through the user account. The doctor in this scenario can analyze the results remotely through his account and adjust the therapy if necessary. The second scenario provides for the use of the uroflowmeter directly in a medical center, where the doctor fully controls the device and the patient is only required to take tests.
Also our team independently developed hardware part on the basis of the element base specified by the customer as recommended. The wireless and measuring modules were designed to meet high medical standards, the device was tested for durability through accelerated fatigue tests for devices of similar class.
Technologies
Languages
Python, JavaScript, C++, C
Frontend
React, Material UI
Backend
FastAPI, RabbitMQ
ML
TensorFlow, Keras, Transformers
DB
PostgreSQL, Redis
Hardware
Tensilica
Process
The waterfall software development process model was used in this project. Such a decision was due to the specifics of the project and strict requirements to the quality standards of medical equipment, as well as the need for documentation for the following stages of certification and registration in the register of medical equipment.
Team
2
Backend developers
1
Frontend developer
1
DevOps
1
Design Engineer
1
Project Manager
1
ML developer
1
Hardware engineer
Results
The developed device and system successfully passed clinical tests, received a certificate of compliance with the requirements for medical equipment and was registered in the register of medical equipment.