Inspiration

Health conditions like heart disease, asthma, diabetes affect large parts of the US population. The nature of such conditions requires patients to lead very disciplined and careful lives, practicing vigilance over their diet and medication. Thus, an emergency related to such conditions is highly critical and is the cause of a lot of preventable deaths in the United States and the world. Such patients also have a higher than average likelihood of suffering depression and suicidal tendencies. An overwhelmingly large fraction of such cases are older adults (65+), economically underprivileged people and people of color. Asthma alone kills around 4000 people a year. Diabetes was the cause of over 100,000 deaths in the US in 2020. 10.3% of all emergency room visits in the United States in 2018 reported to be suffering from chronic asthma and 12.8% reported to be suffering from diabetes.

In the US, the driving time to the nearest hospital can be as long as half an hour, with rural areas being affected the worst. With our healthcare system pushed to its limits, even a short delay can spell the difference between life and death. This inspired us to design Apollo, a solution that can provide faster urgent care to people suffering from such diseases.

What it does

Apollo consists of various interconnected systems which operate in harmony to ensure you get the quickest emergency care.

Infrastructure: Apollo's main hardware infrastructure is a fleet of high-tech drones. These drones are equipped with high resolution, high optical zoom cameras. Modern delivery drones are capable of flight times as long as 30 minutes and are easily capable of transporting lightwight medical supplies (e.g. Epipen, inhaler etc.). These drones are distributed in various locations across the deployment zones. Each location has an inventory of emergency medical supplies. When Apollo detects an emergency, a drone is dispatched to the affected user's location tracked by GPS. Upon arrival, the drone delivers the package and settles into surveillance mode, acting as an audio-visual portal for first responders to monitor and attempt communication with the patient. The drones can also alert nearby people to the emergency so that they may be able to help.
Prediction and Preparedness: Apollo is always learning. Machine learning on anonymized statistics lets Apollo predict when and where emergencies are more likely to occur. The system redistributes drones and resources to meet demand more efficiently. Apollo learns from previous records to learn hotspots (e.g. elderly retirement homes). The idea here is that ideally all accidents are equally likely, but there may be underlying causes (neglect) which make some areas more prone to accidents than others. With the development of quantum computing beyond the NISQ era, we are excited to explore what speedup and efficiency boost we can derive from quantum graph algorithms in this regard.
Connectivity: Apollo's effectiveness comes partially due to it's inter-connectivity with several health monitoring ecosystems. Patients with special equipment like smart blood-glucose pumps can connect it to Apollo. Any critical readings in such devices leads to an emergency detection for Apollo. While Apollo is primarily accessible to the end-user as a mobile app, we plan to make Apollo capable of running on wearable devices like smartwatches independently as well, making the system more robust.
Hands-free: Apollo doesn't require any user interaction to operate. Once the initial configuration is complete, user interaction with the system is optional. Detection of a medical emergency is handled by wearable monitoring devices, which can identify several emergency situations. A summary of the detected emergency is sent to emergency services dispatch, and first-aid measures are automatically chosen.
Privacy focused: Apollo's data travels over secure channels only. All event detection is performed on device and data is event data is end-to-end encrypted in motion and at rest.
Accessible: The hands-free design of Apollo was considered with accessibility in mind. Upon detection of an emergency, the app transitions to an alerted state and will automatically contact 911 unless the user (by button press, touch or voice) stops the process. Allowing different modes of interaction means that patients with disabilities are afforded equitable access to emergency care.

What makes Apollo unique is its multi-pronged approach to emergency care. By employing drones, Apollo can achieve fast response times delivering crucial first-aid resources, and help first responders and EMT gauge the situation even before they arrive.

How we built it

Given that the development of such a system is a very nontrivial task, we focused on designing the system and the end-user experience for Apollo. The primary mode of interaction with Apollo is the mobile app and wearable app. We designed mock-ups for the basic flow of interaction with this app using Figma.

Challenges we ran into

There are many obvious challenges to the development of an emergency response system and we ran into a few of them.

Dealing with False positives: Any such system is prone to false positives. In this case, a false positive leading to an EMT dispatch wastes both time and resources. To overcome this, we allow Apollo to run on multiple devices. Currently, Apollo will live on the mobile app and the smartwatch app, but we can see this expanding to other devices in the future. Having redundant data for the same individual means we have higher signal-to-noise ratios for detecting an actual emergency.
Reaching digital non-natives: Some of the most vulnerable members of our society are those who require special considering in terms of their interactions with technology. To that end, the hands-free nature of Apollo acts to our advantage. Minimal interaction is required for the system to work.

Accomplishments that we're proud of

We are proud that we have been able to come up with a technology solution that will afford people living with chronic diseases and health conditions to live a more peaceful life, knowing that they will get the care they need in the case of an emergency.

What we learned

Our biggest takeaway from this project is the vast complexity of tacking problems in healthcare. Emergency care is a delicate system that requires balancing demand with scarce resources in a highly time-sensitive manner. Although our solution shows great promise to vastly improve the prognosis of unreported emergencies leading to preventable deaths, it is but a small step towards designing a more effective healthcare system.

What's next for Apollo

Apollo is still early in the development stage.

A lot more research needs to be done on the optimal way to collect and process data on the users devices for true event detection while maintaining privacy.
We must take advantage of modern machine learning and quantum computing techniques to ensure that we operate as efficiently as possible.
With an increase in drone fleet size, Apollo may be able to assist elderly and disabled members with prescription delivery, with appropriate authentication mechanisms.
Deep integration with upcoming smart health accessories to be able to automatically detect more conditions such as seizures, anaphylaxis etc.