A new software system for smartphones can quickly and unobtrusively detect early signs of opioid overdoses, according to a new study. The software successfully identified respiratory depression (a hallmark symptom of overdose) in opioid users in an injection facility and during simulated overdoses (mimicked by anesthesia) in an operating room (OR). Although further optimization is needed, the smartphone system holds potential as a cost-effective tool for identifying early-stage overdoses, which require prompt medical attention. Every day in the U.S., 115 people die from opioid overdoses, and the annual death toll continues to climb. High doses of fentanyl or other drugs contribute to the opioid crisis because they cause apnea (a temporary cessation in breathing) and death due to respiratory failure. In search of an earlier intervention strategy, Rajalakshmi Nandakumar and colleagues created algorithms that run on commercially available smartphones and detect precursors to opioid overdose. Their software converts the smartphone's speaker and microphone into a short-range sonar device that continuously sends out acoustic signals, which bounce off surfaces such as a moving chest and return to the phone. The system then analyzes these signals and detects changes in breathing and movement patterns that indicate the beginning of an overdose. Nandakumar et al. first tested the platform in a supervised, approved injection facility with 194 participants who injected heroin, fentanyl or morphine. The software - installed on a Galaxy S4 smartphone placed within one meter of the participants - identified 97.7% of post-injection apnea events and 89.3% of post-injection respiratory depression events (overdosed participants were resuscitated by staff without issue). It also correctly detected 19 out of 20 simulated overdose events in an OR, where patients underwent anesthesia to mirror overdose symptoms such as loss of consciousness. The authors envision that an optimized version of the platform could be linked with emergency services or family members and alert them of developing opioid overdoses in real time.
Science Translational Medicine