On Friday, JHPIEGO informed us that women in low- to middle-income countries were 50-100x more likely to die from post-operative complications. The reason was simple: there were not enough trained healthcare professionals to monitor patients during the critical window after surgery, when they are most likely to develop infections or begin hemmorhaging. By the time complications are diagnosed, we were informed, it is often too late to save the patient. We at EMP challenged ourselves to develop a better way to support these over-worked healthcare professionals and reduce the overall post-operative mortality rate, particularly for women recovering from cesarean delivery. From Friday night's pitch until Sunday morning, we worked. We researched. We brainstormed. We questioned. We spat out crazy ideas and tested and failed and succeeded. And finally, at the conclusion of 36 hours of collaborative effort, we believe we have a system to address the need at hand.
What it does
Early Morning Prioritization System couples an inexpensive wearable with a centralized data monitoring system. A set of temperature-sensitive and PPG-enabled wristbands will be provided to surgical facilities in low-income countries, with the assumption that the clinic has an iPad to support the dashboard software. When a patient leaves the operating room, she is fitted with one of the wristbands and her name associated with the device identifier. The wearable records key vitals such as her heart rate, blood oxygen saturation, and body temperature and transmits them via class 1 Bluetooth to the clinic's iPad. Our algorithm then ranks the patients with wristbands according to their vitals - those with elevated blood pressure, elevated body temperature, or abnormal blood oxygen saturation levels are ranked to the "red zone" tier, while those with slightly abnormal readings are ranked to the "yellow zone", and the stable patients to the "white zone." Health care providers can click on the patient name to view the current readings on the patient and see if action is required. This allows providers to constantly monitor and prioritize their treatment. Moreover, if a patient is 'upgraded' to the red zone tier, an SMS message is sent to the nurse on duty alerting him or her to the change.
How we built it
We leaned heavily on our coding team's extensive knowledge to develop our algorithm and dashboard for physicians to use. Our goal was to make the dashboard as intuitive and graphic-based as possible so as to limit the need for translation or potential miscommunications. While we lacked the time and technical components necessary to build a functional prototype of our wearable, we were able to develop a rough model of what the device will look like and how all of the necessary components would be arranged within the wristband cell. Furthermore, we read and vetted numerous studies affirming both that - and how - our modeled device would work and communicate with our dashboard system once assembled.
Challenges we ran into
Navigating the intricacies of Bluetooth was more difficult than expected, and our next step would be to coordinate more than 7 devices to a single dashboard hub as well as try to extend the range. While we initially wanted to transmit our data over WiFi, we recognized that many clinics and hospitals in low-income nations do not have consistent and reliable access to the internet. Incorporating a Bluetooth chip with a longer range does increase our per-device cost, but allows the wearable to be used in far more clinics than would a WiFi-based system. Moreover, we attempted to 'hack' a FitBit and an Apple Watch to show real time data measurement and the associated display on our dashboard, but we were informed we could not open the FitBit to extract the heart rate sensor (the Apple Watch was a team member's).
What we learned
As a team, we went from limited knowledge on how to measure and interpret patient vitals (or even which vitals we wanted to measure) to a robust and applicable understanding of PPG, temperature sensing, motion artifacts, filter noises, and the potential applications of our device. We researched different placements and accuracies of existing wearables, postulated on what would be most reasonable for our patients, and engaged in great debate and deliberation on how we could minimize the overall cost of our system. We also learned a great deal about the current clinical situation in developing nations, and differences in healthcare practices worldwide. We also postulated a potential use for our wearable to reduce bottlenecking in overcrowded clinics, by developing outpatient facilities through community partnerships. The EMP would allow physicians to remotely monitor patients for complications while the patients wait for sutures to be removed (or wait for any other follow-up that may be needed). We hope that increasing community involvement in post-operative care, particularly of new mothers, will also increase awareness of the signs and symptoms of complications and the need for quality surgical care.
What's next for earlymorning
Our next step would be to build a physical prototype of the wristband and transmit the vitals recorded from the device to our dashboard. While we have modeled and determined how we would make our data useable, we are eager to begin work on a real device should we be given the chance to do so. We also would investigate how we can coordinate large numbers of Bluetooth-enabled devices and their associated patients, as our current capabilities are limited to seven.