An analysis of drug deaths using machine learning algorithms suggests that the number of people dying from opioids in the United States over the past 20 years is likely to be far higher than previously reported. Elaine Hill of the University of Rochester in New York and her colleagues, when examining data on drug overdose deaths, realized that 22 percent of such cases reported between 1999 and 2016 were classified as overdoses on death certificates without specifying the drugs involved, New Scientist reported.
“Given the scale of the problem, we find this noteworthy. Team member Andrew Boslett said.
The team tried to estimate that opioid-related deaths were caused by opioids by analyzing the coroner’s medical reports on opioid overdoses and non-class overdoses.
First, the researchers used machine learning algorithms to analyze deaths recorded as opioid overdoses. They were able to identify common factors involved in the treatment of opioids, such as long-term pain and arthritis.
Using this information, the team estimated that 72 percent of unclassified overdose deaths were related to opioids. The findings show that the number of people dying from opioid overdoses in the United States is 99,160 more than previously thought, an understated 28 percent. According to these new results, more than 450,000 people have died from opioid overdoses in the United States since 1999.
“We were initially surprised by the data, but then when we saw what was being done across the country, we thought it seemed reasonable. Hill said.
According to the analysis, the number of opioid overdose deaths reported in some states is much higher than in others. Pennsylvania and Delaware have the most serious problems in this regard.
The team is currently studying why so many deaths from overdoses in the United States have not been classified, and the situation continues.
“It is well known that the number of deaths associated with opioid overdoses is often underestimated. Jennifer McNeely of New York University says that only knowing where the situation has the greatest impact can people intervene most specifically.