COVID-19 is an emerging, rapidly evolving situation.


Get the latest information from CDC (Coronavirus.gov) | NIH Resources | NIDA Resources

Drugs & Health Blog

A Teen Scientist Is Helping To Solve Mysterious Drug-Related Deaths

L to R: Daphne Liu and Mia Yu. Image by Daphne Liu. 

The NIDA Blog Team

Every day in the United States, about 185 people die from a drug overdose. When that happens, the medical examiner or coroner records the death as accidental (unintentional) or intentional (suicide).

It’s important to know if the death was accidental or not, so we can develop better strategies to help prevent suicides. But often, officials can’t be sure if a death was intentional.

A person could intentionally overdose without telling anybody about that intention or leaving a note, making it difficult to determine if they meant to take their lives.  If a medical examiner can’t be sure if a death was intentional, they record it as “undetermined.” 

Psychological autopsy
In the United States, deaths are recorded by county officials, and the standards for recording deaths are different in each county.

In many cases, the only way to know whether a death was intentional is to conduct a “psychological autopsy”: family, friends, coworkers, and others are interviewed to learn the state of mind of the person who died. Unfortunately, that method is very expensive and time consuming.

But, thanks to teen scientist Daphne Liu of Utah and a team of other scientists, we may be getting closer to some solutions.

We first met Daphne when she and her science partner, Mia Yu, won NIDA’s first place Addiction Science Award at the International Science and Engineering Fair in 2018. Someone in Daphne’s local health department suggested that she study statewide overdose data, and because her brother had lost a friend to overdose, this idea felt right to her.

Machine learning
Daphne and Mia designed software that teaches computer systems how to review details about overdose deaths and determine how likely it is that a death was intentional. This is a form of artificial intelligence called “machine learning.”

Dr. Paul Nestadt at Johns Hopkins University in Baltimore read about the award and contacted Daphne’s science teacher to suggest she work with his team and refine her machine learning concept to improve its accuracy.

Daphne, now a junior in high school, says the scientists will continue to modify her computer algorithm to evaluate fatal overdoses from other states. They hope to build machine learning systems that could make psychological autopsies less expensive and time consuming, leading to more accurate outcomes.

So now Daphne's work could make a difference across the country. By more accurately accounting for how and why someone died, doctors and scientists can better help people who have problems with drugs—before it’s too late.

Learn more: How to get help if you or a friend is feeling suicidal.

Categories: 
Staying Healthy
Comments posted to the Drugs & Health Blog are from the general public and may contain inaccurate information. They do not represent the views of NIDA or any other federal government entity.

Add new comment

Current state: Approved
CAPTCHA
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.