Q & A on Latest Monthly Estimates of Drug Overdose Deaths

HOST: New data that came out on Wednesday of this week show that from June 2019 thru May of 2020, there were an estimated 81,230 deaths from drug overdoses in the United States, an increase of 18% from the 68,829 deaths that occurred the year before. Drug overdose deaths were up in all but four states – New Hampshire, Idaho, Nevada, and Utah. The biggest increases were observed in Louisiana, Iowa, Wyoming and Maine.   The increase in Washington, D.C. from last year was 60 percent – a larger increase than in any individual state. There were over 10,000 more deaths from opioids, particularly fentanyl, as well as increases in deaths from cocaine and from psychostimulants such as methamphetamine. Deaths from heroin have leveled off in recent years and were only slightly higher in this period ending in May 2020 than the previous year.

NCHS began publishing monthly estimates of drug overdose deaths in August of 2017 in response to the growing drug crisis in the United States. The lead analyst for this initiative, Farida Ahmad of the NCHS Mortality Statistics Branch, joined us to discuss this initiative as well as the latest findings.

Well let’s go into the latest numbers – so what do the latest numbers show?

FARIDA AHMAD: Drug overdose deaths went up — they had been increasing for many years, but they really shot up in 2017. But then they started plateauing in 2018 and 2019 but in the summer of 2019 we started seeing the number of deaths start to go up again and in this latest data which is through May 2020, we show not only a continuation of that increase but the beginning of a steep climb in April and May of 2020.

HOST: So this date is featured on a pretty elaborate dashboard on the NCHS website and on some of the graphs illustrating the trends talk about reported deaths and predicted deaths. Could you discuss briefly what the differences between reported and predicted deaths?

FARIDA AHMAD: Reported deaths are the actual number of deaths that NCHS has received to date – so at the time of analysis how many deaths we have received – but we know that data aren’t complete, this is provisional data and it can take months for some drug overdose deaths to be reported. So we want to make sure that we’re not reporting declines when it’s just, when we know it’s actually a factor of data lags or something in the data flow process that’s showing that we have fewer deaths. So you know we came up with these adjustment factors that take into account the typical delays in that data flow process so that we have a better idea of what the numbers should look like once they’re finalized. So we take provisional data and compare it to final data and come up with these adjustment factors to predict what the numbers would look like.

HOST: And the next question I guess would be what is the difference then between provisional data and final data?

FARIDA AHMAD: Provisional data, again, it’s based on incomplete data and we do apply adjustment factors but those aren’t perfect because sometimes states might submit data faster than we would have expected or that our adjustment factors had predicted or they might take longer than we had predicted. So the numbers could be a little different but with final data on the other hand you know that gives the states enough time to complete all the death investigations and complete their final death certificates so that we have the most complete information so it’s not numbers that we’re predicting but the actual numbers.

HOST: The last couple of months have shown substantial increases in overdose deaths – April 2020, May 2020,now this recent release, so people obviously will be pointing to the pandemic as driving those increases. Can you talk briefly about the most recent data we’ve seen and compare it to before the pandemic? Any notable trends there?

FARIDA AHMAD: Yeah you know it’s important to keep in mind that the increases had begun in the summer of 2019, so pre-pandemic we were already seeing steady increases in the number of drug overdose deaths. You know, in April and May we did some more stark increases, so still it’s a little early to tell exactly what the role the pandemic has played in those recent increases.

HOST: You talked about the plateau – two years ago there was actually a slight decline in the number of deaths and it was the first decline in 28 years, which is fairly amazing, and but now at the end of 2019 you saw this increase and now we’re seeing more of an increase. So in hindsight, was that just a pause or a blip in the trend there?

FARIDA AHMAD: It’s too early to tell what the trend will be going forward but you know one thing to keep in mind is that the drug overdose deaths had been increasing, as you said, for many years and so that not only was there that initial decline in 2018 but there was a sustained plateau for at least a year and that’s positive news. Now that we’re seeing increases again, again it’s early to tell but I’m hopeful that the declines in the plateau that we had seen in 2018… we can return to that state again.

HOST: So your monthly data looks at specific drugs responsible for overdose deaths and what is that data tell us about the types of drugs that are involved?

FARIDA AHMAD: What we’re seeing is that synthetic opioids like fentanyl are driving the increase in deaths nationwide, but we also do see that in parts of the country psychostimulants like methamphetamine have caused a lot of deaths too. Although in this most recent report you’ll see that even in some of those states the increases in recent months have been driven by synthetic opioids.

HOST: So overall, which states have seen the biggest increases in overdose deaths in this last release and which states are still making progress in reducing the deaths?

FARIDA AHMAD: A lot of states are seeing increases – some of the states with the biggest increases are Louisiana, Iowa, Mississippi, Maine, Florida and even DC. And some of the states that seem to be making progress are New Hampshire, Nevada, and Utah. But one thing I will urge is that this is provisional data and sometimes what looks like a decrease might just be not enough data.

HOST: OK well thanks for joining us Farida.

FARIDA AHMAD: My pleasure. Thank you for having me.