There's a communication problem in medicine. COVID-19 has brought it to the forefront.

My mother is a labor and delivery nurse at Brigham and Women’s Hospital, one of the most respected medical institutions in the world. Even there, PPE is scarce among staff and providers, and must be rationed accordingly.

On many wards, like Labor & Delivery, patients must be attended to immediately, before the results of a COVID-19 test are able to come out. One day, my mother came home and said that she had taken care of a patient who within hours had tested positive for COVID-19. Because she had neither a precise way to assess the risk of this patient before the test result came out nor proper PPE, she was incredibly anxious that she was at risk of contracting the virus. Her colleagues, many of whom live with people or themselves deal with autoimmune conditions or other factors that put them at high risk of complications from COVID-19, were even more petrified.

We set out to make it so that nurses and providers will never face this issue again.

Even when COVID-19 test results do come out, nurses, staff, and providers may not be informed until they log into a centralized hospital system such as EPIC. This puts them at risk of not knowing the most up-to-date COVID-19 status of the patient and thus implement necessary precautions. Outbreaks of COVID-19 in hospitals ( and often a lack of widespread testing among providers and nurses mean that providers are potentially also putting their other patients at risk of COVID-19.

With the Code Blue Informatics GreenLight Wearable, we solve the COVID-19 communication problem.

We built a platform that warns Nurses and providers in real time of the COVID-19 risk of the rooms that they enter. We use a published algorithm developed by the Cleveland Clinic to approximate the percent chance of a patient testing positive for COVID-19 based off of demographic and health factors — those that would be typical of a verbal screening done as the patient enters the hospital. We remove identifying information and send it to a database of room numbers and calculated risk levels (between 0 and 100).

Nurses and providers wear a hand-held buzzer and light system which alerts them of the patient’s COVID-19 probability or status as they approach the door to the patient’s room. The buzzing sound and bright blue or red led associated with medium and high risk patients is hard to ignore in even the loudest hospital setting. The buzzer reminds the provider to take further precautions — for both their safety and the safety of the patient — before entering the room.

To build the Code Blue Wearable Alert System, we interfaced Python with Arduino. Arduino relays information to Python about how close the provider is to an ultrasonic sensor attached to the room. The ultrasonic sensor is calibrated so that it won’t be triggered unless the provider is very close to the room, and so this reduces the risk of accidentally setting off the sensor. Python then finds the COVID-19 risk number associated with this room and tells the wearable room risk. The wearable then triggers a distinctive low, medium, or high risk state and the provider is informed of the room’s COVID-19 risk. In our database, patients who have tested negative receive a low risk rating. Patients who have tested positive receive a high risk rating. Patients who have pending test results receive a risk rating determined by the Cleveland Clinic Algorithm.

We also worked extensively on rolling out an input and encryption system for screening data using Flask. As of now, we have a static website for this input. The team that worked on this aspect knew nothing about SQL and Flask before BorderHacks. They learned a great amount about Flask and SQL during the hackathon, and they are proud of this.

Our foremost concern is patient privacy as outlined by HIPAA, and for this reason our Python file has information that confirms the identity of single patients. For ease of use, we will, in the future, integrate a system that uses end-to-end encryption to protect patient personal information beginning when providers enter screening information into their computer. Our final system will encrypt this information and send it to EPIC and our wearable database, mitigating the risk of non-providers accessing this information and meaning that providers only have to enter screening information once.

We also greatly acknowledge the deleterious effects of algorithmic racial bias in North America’s current healthcare system, and the disproportional toll that COVID-19 has taken on communities of color. Thus, we strive to implement solutions for reducing racial bias in our assessment of room risk through an algorithm which already takes demographic factors into account.

We have gotten positive feedback from doctors about the potential of this system, and so we aim to collaborate with them to implement this hack in a real hospital setting.

Built With

Share this project: