-
Solution Diagram
Inspiration
Nurse rounds give doctors the most up-to-date information to interpret a patient's condition and make clinical decisions. However, due to the intensity of manual monitoring, medical errors and stress on the caregivers can take place. Automated, continuous monitoring reduces the likelihood of medical errors and the data it collects can be used to reduce the time-consuming/stressful work such as scheduling medical and personal obligations.
What it does
The main purpose of the Miniso Virtual Room is for automated continuous patient monitoring. It will collect a variety of data (vitals, eating/sleeping habits, exercise activity, etc.) and send them to a data server to store them as medical tracking history. Then, the server can process the data, utilizing AI/ML models trained to predict potential risks and critical issues based on factors like abnormal heartrate and vitals. If a patient is flagged for a potential non-urgent risk, then additional preventative treatment can be offered to a patient's medical treatment plan, such as rehabilitation or setting up with a specialist such as a speech pathologist. If a patient is flagged for a critical urgent issue, then the system will immediately alert caretakers of the problem so that they can attend to the patient and stabilize the patient's condition. In addition to automated monitoring, the Miniso Virtual Room aims to use the data it collects to provide patients (such as those who have recently had a stroke) with a sense of control of their life. Specifically, the data it sends can help determine the medical needs of a patient such as medication times and medical appointments as well as personal needs such as personal events like weddings or birthdays. Depending on these needs and other considerations, Miniso can generate a daily, detailed schedule for the patient, taking the burden of logistical planning away from the patient and loved ones and letting them focus on enjoying their life.
How we built it
Our initial pitch is an abstracted model of how the Miniso Virtual Room will be used to collect data and how that data is processed to give the necessary updates to medical tracking history and record as well as services for the patient such as a personalized daily schedule. The idea is conceptualized in Figma and Microsoft Powerpoint.
Challenges we ran into
There was great difficulty in implementing the AI/ML aspects of our solutions. For example, the Miniso Virtual Assistant would be used for a daily routine checkup of facial, emotional, and speech-related changes. One aspect that was difficult to implement was an image classification model for facial expressions. The idea was to have a standardized facial expression (ex. "Give me a smile!") in order to track physical changes that could be used to predict if a patient was at risk for a future stroke. Additionally, there was difficulty in using a sentiment analysis model against speech input (ex. asking "Tell me about your day.") in order to measure the positivity/negativity of a response for emotional monitoring. There was also great concerns over what the virtual room could monitor on its own. Noninvasive and simple procedures to collect data are useful and likely ethical, but invasive procedures such as blood tests would raise ethical concerns on many accounts. However, collecting certain data may give more insight into a patient's condition, so there is a balance of collecting data and the ethics behind it.
Accomplishments that we're proud of
Integrating a variety of unique views and ideas into a cohesive, realistic solution. Solving a problem that many families could experience.
What we learned
On the ethical side, AI/ML has its perks, but it also has its limits. It cannot do everything, but it can make simple tasks automated and convenient. Rather than a tool to replace everything a human does, it can be used as a cooperative tool to make humans more efficient, letting machines do the simple, manual work and letting humans do work that requires expertise and careful attention. On the medical side, we learned about the latest procedures and technologies that goes into monitoring during nursing rounds and innovated new solutions to make the process more efficient. On the technical side, we learned about different AI/ML tools and techniques for image classification and speech sentiment analysis as well as possible solutions for scheduling with many factors.
What's next for Miniso Virtual Room
Currently, the Miniso Virtual Room is a conceptual idea. However, with current technology, much of the ideas mentioned can be implemented (i.e. the data collection and using AI/ML models to process the data). Additionally, the Miniso Virtual Room is scalable in the sense that emerging medical technologies that collect data can be integrated into the room.
Log in or sign up for Devpost to join the conversation.