R(eye)covery

Warning sober sponsors about a sponsee's opioid relapse

Social Impact: One of our team members is a pharmacy student and talk around the opioid epidemic is highly prevalent throughout her school's community. Synthesizing her knowledge from both classes and events such as naloxone awareness seminars, she realized how significant the opioid epidemic is and shared with the rest of the teammates. 4 in 5 heroin users start out with abusing prescription drugs and in 2016, 42,000+ people died from opioid addiction. Relapse rates are extremely high, with 91% of patients unable to stop usage and 59% relapsing within only one week.

Purpose: These high relapse rates are caused by the intensity of addiction. Drug addicts feel more pain than the average human because the body stops producing its own endorphins as the more extensive their drug usage gets. One of the major symptoms of opioid addiction is pinpointed pupils, which means the pupils are constricted and unchanging in response to light stimuli. Physiologically, this takes place due to the mu opioid peptide receptor and kappa opioid peptide receptor activation in the ocular nerve within the blood brain barrier. The stimulation of the parasympathetic nervous system results in constricted pupillary reflex.

Our Idea: We are choosing to focus on opioid addiction patients that are in recovery to support their efforts. Our app, R(eye)covery, encourages open communication between recovering addicts and their sober sponsors. The phone-based application passively monitors a recovering addict's pupils, using a Smartphone's location and camera features. The location can provide the sober sponsors with insights into a sponsee's commonly visited places, especially in high-risk areas of drug contact. The camera feature allows the application to monitor the sponsee for changes in pupil diameter, which can signal opioid use. The application will compare normal expected pupil sizes to potentially changing ones to detect pinpointed eyes. Depending on the user, the app will establish a normal pupil size baseline based on passive background imaging while the Smartphone is in use. The recovering addicts we are targeting are already getting help and are all participating voluntarily and willingly.

User Interface: The application offers log-ins and varying settings for both sponsees and sponsors. The sponsee can choose to share location and allow pupil detection via their camera. The sponsor has settings to choose to be notified when the sponsor either registers possible opioid use, turns off pupil detection, and stops sharing location.

Application Functionality: We have developed a Matlab image processing prototype to screen pupils for pinpointing based on image analysis. The market application would take series of images of patients' eyes to monitor pupil dilation. The application would also utilize facial recognition software to identify the patient. The images collected through retinal monitoring undergo filtering by deep learning methods through convolutional neural networks. 2D and 3D pupil ellipse sizes (measured in pixels) and angles (measured in degrees) are compiled via the CNN algorithmic structure to map changes in eye movement and to detect significant alterations in pupil diameter that reflect opioid abuse. “Pinpoint pupil” side effects are classified by measurement analysis of the major axis and change of dimensions of the pupil ellipse which characterize an episode.

Financial Motivation: Our application focuses on preventative care to help recovering drug addicts reduce the chance of opioid relapse, which is one of the biggest challenges to those recovering. The nationwide economic cost of the opioid crisis in 2015 was $504 billion. The opioid epidemic lead to increased costs because of drug treatment services, inpatient hospital services, medical examiner costs, criminal justice costs, law enforcement costs, costs of ER visits and of delivering naloxone.

Future Aims and Considerations: Pupil dilation analytics can be used as a warning mechanism for a sober sponsor, but is not a strong enough indicator to be used as a primary diagnostic tool for opioid use. In the future, pairing with other Smartphone analytic testing to create a combination opioid use detection method would move closer to point-of-care diagnostics. For example, a sensor can track respiratory breathing patterns to monitor for potential respiratory depression, another major symptom of opioid abuse.