死亡人数与类罂粟碱一项研究显示，在过去的二十年里，美国的失业率可能比之前认为的要高30%。

研究人员查看了美国国家健康统计中心保存的数据库中1999年至2016年期间死于药物过量的人的数据，其中包括632，331例病例。这使得研究小组能够在没有具体原因的情况下将死亡记录信息与药物过量进行匹配。

其中，78.2%的病例有相关药物的信息，而21.8%的病例没有。研究小组发现，在研究过程中，71.8%的未分类药物过量与阿片类药物有关，比之前报道的多约28%。与处方类阿片、海洛因或芬太尼相关的死亡人数增加了99，160人。

该小组发现包括阿拉巴马州、密西西比州、宾夕法尼亚州、路易斯安那州和印第安纳州在内的各州差异最大。根据研究人员所在的罗切斯特大学的一份声明，例如，宾夕法尼亚州报告的与阿片类药物相关的死亡人数为12374人，但该研究将这一数字定为26586人。

这项研究是在一个类鸦片导致130名美国人死亡的药物过量流行病s每平均每天，根据美国疾病控制和预防中心预防。在2017年死于药物过量的70，200多名患者中，大约68%的人服用了一种药物类鸦片。这包括处方药和非法药物，如海洛因及其更强的合成药物芬太尼。这场危机始于1990年的一次上涨类鸦片20世纪90年代的处方。

安德鲁·杰博斯利特,徐子凝 Denham和伊莱恩·希尔共同撰写了这篇发表在杂志上的论文上瘾。他们说新闻周刊:“我们怀疑的司机少报关于类鸦片在美国，药物过量可能是由于缺乏医疗检查的资源，以及对验尸官的持续依赖(尽管越来越少)，验尸官是选举出来的官员，在医疗检查方面的经验可能有限。"

在解释他们为什么进行这项研究时，他们说:“许多研究人员推测社会-过去二十年来，该国许多地区的经济不景气是导致吸毒过量和死亡率上升的原因之一。

“在过去的几年里，我们一直在研究页岩繁荣、煤炭萧条和其他经济冲击以一种有意义的方式影响了药物过量的流行。在我们开展这项研究的过程中，我们通过查看数据和阅读文献注意到，大约20%到25%的药物过量没有在死亡记录中列出原因药物。

“我们意识到这是一个问题，不仅仅是对我们的研究而言，它依赖于对当地药物的高质量评估类鸦片跨时空的用药过量率——但也是为了让国家了解类鸦片过量流行。"

该小组继续说道:“我们怀疑美国阿片类药物过量报告不足的原因可能是缺乏医疗检查资源，以及对验尸官的持续依赖(尽管越来越少)，验尸官是选举产生的官员，在医疗检查方面可能只有有限的经验。”

他们本可以在机器学习中应用更先进的工具，这凸显了他们研究的局限性。

​

OPIOID-RELATED DEATHS IN THE U.S. COULD BE FAR HIGHER THAN PREVIOUSLY THOUGHT, STUDY SUGGESTS

The number of deaths linked to opioids in the U.S. over the past two decades could be almost 30 percent higher than previously thought, a study has revealed.

Researchers looked at data on people who died of drug overdoses between 1999 to 2016 from a database kept by the National Center for Health Statistics in the U.S., which included a total of 632,331 cases. This enabled the team to match up information on death records with drug overdoses without a specific cause.

Of those, 78.2 percent of cases had information on the drug involved, while 21.8 percent didn't. The team found that 71.8 percent of unclassified drug overdoses over the course of the study involved opioids, or approximately 28 percent more than previously reported. That amounts to 99,160 additional deaths linked to prescription opioids, heroin, or fentanyl.

The team found states including Alabama, Mississippi, Pennsylvania, Louisiana, and Indiana, had the highest discrepancies. According to a statement by the University of Rochester where the researchers are based, the number of reported opioid-related deaths in Pennsylvania, for instance, was 12,374, but the research puts the figure at 26,586.

The research comes amid an opioid overdose epidemic which kills 130 Americans every day on average, according to the U.S. Centers for Disease Control and Prevention. Of the over 70,200 individuals who died of a drug overdose in 2017, around 68 percent had taken an opioid. That included prescription and illegal drugs like heroin and its stronger synthetic counterpart fentanyl. The crisis started after a rise in opioid prescriptions in the 1990s.

Andrew J. Boslett, Alina Denham, and Elaine L. Hill co-authored the paper published in the journal Addiction. They told Newsweek: "We suspect that the driver of underreporting of opioid overdoses in the United States may be due to a lack of resources for medical examination, as well as continued (though decreasing) reliance on coroners, who are elected officials with potentially only limited experience in medical examination."

Explaining why they carried out the study, they said: "A number of researchers have speculated that socio-economic malaise in many areas of the country over the last twenty years has played a role in increasing drug overdose-related abuse and mortality.

"Over the last few years, we have been developing research on whether the shale boom, coal bust, and other economic shocks have influenced the drug overdose epidemic, in a meaningful way. As we developed this study, we noticed by looking at the data and reading through the literature that around 20 to 25 percent of drug overdoses did not have a drug-of-cause listed in the death record.

"We realized that this was an issue, not only for our studies—which would rely on high-quality estimates of local drug and opioid overdose rates across space and time—but also for the country's understanding of the toll of the opioid overdose epidemic."

The team went on: "We suspect that the driver of underreporting of opioid overdoses in the United States may be due to a lack of resources for medical examination, as well as continued (though decreasing) reliance on coroners, who are elected officials with potentially only limited experience in medical examination."

Highlighting the limitations of their study, the team they could have applied more advanced tools in machine learning.