The United States is experiencing an epidemic of drug overdose deaths: every year, nearly 44,000 lives are tragically lost, with more than half of the fatalities caused by prescription drugs.

We’ve analyzed the latest government data to find out why medications designed to improve lives are instead prematurely ending more than at any other time in history. What can be done to curb the rate of prescription drug deaths in America?


Our data come from the Centers for Disease Control (CDC), which analyzes the death certificates of more than 2.5 million residents every year. Alongside the cause(s) of death, the CDC records key demographic facts about decedents, such as their gender, age, race, and where they lived and died. By filtering CDC’s mortality data using International Statistical Classification of Diseases codes (ICD)1, we are able to see precisely how many deaths were caused by the accidental or deliberate misuse of prescription drugs.

The graph above summarizes the big-picture story over a 16-year period (up to 2014 – the most recent year for which data are available). Transport accidents were once the leading cause of unintentional deaths in America, but since 2010, drug poisoning deaths have taken over, killing 14.7 out of every 100,000 people in 2014. “Drug poisoning deaths” is a parent category used by the CDC, and it includes illegal drugs like cocaine and heroin, and legal drugs, such as prescription and OTC medications.

Prescription opioid analgesics, such as hydrocodone (e.g., Vicodin, Lortab), oxycodone (e.g., OxyContin, Percocet), morphine, and codeine, are used to control pain and are highly addictive. In fact, 73% of prescription drug deaths in 2014 were attributed to opioid analgesics.

The graph above shows that prescription drug deaths in America, fueled mostly by opioids, have been climbing for years. The animated map below shows how they increased in each state, beginning in 1999 in New Mexico, which at the time had the highest rate in the country, at 11.4 per 100,000 residents.

003 Over the course of 16 years, the map becomes darker and darker, signifying a country-wide increase in prescription drug deaths. From beginning to end, no state remained unaffected by the epidemic; nationally the death rate nearly tripled, from 2.7 per 100,000 people in 1999 to 8 per 100,000 in 2014.

Using a heat table, in which low rates are shown in green and high rates in red, we can visualize the progression of prescription drug deaths, and pinpoint when the problem first began to escalate. We can also see which states – if any – have managed to curb the issue in recent years.


The farther right a state’s green cells extend along the table, the longer it was able to keep its prescription drug overdose rate relatively low. Unfortunately, by 2013 and 2014, very few states have green cells and relatively low rates. North Dakota had the lowest median rate over the 16 years, at 2.5 prescription drug deaths per 100,000 residents (the national median was 6.4), but its rate nearly tripled between 2013 and 2014, from 1.9 to 5.5. Nebraska had the second-lowest median, but – like almost every state – still saw a huge difference between 1999 and 2014 (it more than quadrupled).

At the other end of the scale is West Virginia, which had the highest median rate over the 16-year period, at 17.6 (nearly three times higher than the national median). Perhaps most strikingly, the rate of overdoses in West Virginia caused by prescription drugs was about 8.6 times higher in 2014 than in 1999, and annually rose an average of 26.6% between those years, which was well above the national year-on-year change of 7.8%. Dr. Rahul Gupta, the state’s health officer, has given a partial explanation for why West Virginia has such a problem with prescription drugs, which includes a lack of substance abuse programs, and poor education in communities who live in isolated, mountainous terrain.

Certain states stand out in the table for other reasons. Alaska resisted the dramatic increase seen in most other states for a relatively long time (as the green bar of cells extending to 2007 shows), but then saw a massive spike in deaths in 2008. In fact, there were nearly 6.6 times more deaths than the year before (101, up from 15). New Mexico is also interesting, in that it had one of the lowest average year-on-year percentage increases of all states (5.2%) but only because its prescription drug overdose rate was much higher than most other states in the early 2000s.


The graphic above shows the five highest and lowest states for prescription drug overdose rates. West Virginia is easily the highest, as its rate of 30.2 deaths per 100,000 residents was almost 1.5 times higher than New Hampshire’s, which was second highest. West Virginia was also 3.8 times higher than the national average of 8 per 100,000 people.

Louisiana featured the lowest rate, which – at 3.9 per 100,000 – was almost eight times lower than West Virginia and half the national average. West Virginia’s problem with prescription drugs cannot be overemphasized, as the map below shows. It combines all 16 years of CDC’s prescription drug mortality data into a single county-level map.

006 Knowing how many people die in each county because of prescription drugs isn’t easy – the CDC data suppress numbers that fall below a certain threshold to protect decedents’ privacy. However, if we combine all 16 years (1999–2014) of available data into one data set and map it, we are able to produce a map that does tell the story of prescription drug deaths at the level of individual counties; in many cases, this means individual communities.

West Virginia plays a big part in the story; 10 of the 30 counties (and 6 of the top 10) with the highest rate of prescription drug deaths were in West Virginia (which only has 55 counties in total). Kentucky is also present throughout the top 30, with 10 of its 120 counties making the list as well. In fact, just seven states account for the 30 counties with the highest prescription overdose rates, and they are West Virginia (10), Kentucky (10), Virginia (5), Tennessee (2), Utah (1), New Mexico (1), and North Carolina (1).

Perhaps the most striking way of expressing the country’s current problem with prescription drug deaths is by comparison to the average population of a U.S. county. In 2014, the average county was home to 25,714 people.2 In the same year, 25,760 people across the country died from prescription drug overdoses. In other words, in 2014, the equivalent population of an average U.S. county disappeared because of the misuse of prescription drugs.

With the personal toll of prescription drugs in mind, let’s shift our attention to some of the demographic characteristics of the people whose lives were prematurely ended by them.


Men are more likely to die from drug overdoses than women, but the gender gap differs depending on the type of drug involved. Heroin is most gender-biased, with 77.1% of overdoses involving men, followed by cocaine (72.3% men), and then prescription drugs (56.6% men). Opioid analgesic overdoses, which is a subcategory within prescription drugs, is 58.1% men.

This gender gap exists despite women being more likely to visit the emergency department because of opioids3 (a CDC report from 2013 showed that for every woman who died from an opioid overdose, 30 visited the emergency department)4 and being prescribed abuse-potential medications more often than men. However, the overdose gap is closing. Between 1999 and 2014, deaths from prescription opioids among men rose 245%, compared to 444% among women.


Prescription drug deaths also differ by age and race. The table above combines both factors and shows that overdoses in 2014 were highest among white people aged 45 to 54 (16.3 deaths per 100,000 persons). Prescription drug overdoses among black or African American people in the same age category were nearly two times lower, at 8.5 per 100,000.

Another metric recorded by the CDC in its data on prescription drug overdoses is where deaths occur.

009 The CDC data show that most drug overdoses happen at home, but opioid analgesic overdoses are more likely to occur at home than those involving heroin or cocaine. This is a worrisome finding. The chance of surviving a drug overdose decreases with every passing minute medical assistance is unavailable (as it usually is in the home), and even when an overdose is witnessed by another person, it has been estimated that only 10 to 56% of people will call emergency services.5 Of course, not all overdoses are accidental; sometimes, individuals deliberately ingest painkillers in an effort to take their own lives. We can track the proportion of intentional versus unintentional prescription medication deaths using the CDC data.


The gap between intentional and unintentional overdoses involving prescription drugs has gradually been widening. In 1999, almost a third of deaths involving prescription medications were suicides; by 2014, that number had dropped to 16%.

Heroin (not included above) has seen a similar change: In 1999, 1.5% of heroin deaths were suicides, compared to 0.91% in 2014. So while far fewer people kill themselves using heroin than prescription drugs, the proportional shift has been similar for both categories: 40% fewer suicides for heroin, and 44.9% fewer for prescription drugs.

This parallel is not a coincidence: Heroin and opioid analgesics are closely related. In recent years, heroin abuse and overdoses have increased to epidemic proportions alongside prescription pain relievers, largely because they both have similar effects on users and extremely high abuse potentials. Now more than ever, people are switching between heroin and prescription opioids to keep up their addictions.

This fact was recently highlighted by the National Institute on Drug Abuse (NIDA), which showed that in the 1960s more than 80% of people who entered treatment for heroin addiction first started their addiction directly with heroin. However, in the 2000s, more than 75% of heroin addiction began with prescription opioid abuse.6 NIDA has postulated that this dramatic increase in the poly-abuse of heroin and prescription opioids is largely due to the increased availability of the latter. In 2012, health care providers dispensed 259 million prescriptions for painkillers – enough for every adult in the U.S. to have a bottle of pills.7

We searched CDC’s data for records that showed heroin and an opioid analgesic in decedents’ systems at the time of their death and found the following dramatic increase between 1999 and 2014.


In 1999, only 557 people who died from a drug overdose had both heroin and an opioid analgesic in their systems. This figure dropped for the next few years and then climbed significantly in the late 2000s. The most dramatic spike occurred between 2013 and 2014, when there was a 75% increase in deaths involving both heroin and opioid analgesics. This figure is potentially the clearest reminder of the increasingly deadly relationship between illegal opiates like heroin and legal opioids prescribed by doctors. While their origin and official purpose may differ, the potentially deadly outcomes of abuse are identical.


The CDC has issued a new set of guidelines for physicians across the country that recommend reducing opioid prescriptions in favor of non-opioid painkillers, as well as increasing physical therapy and urine tests to ensure that patients are not abusing their medications.8 This is an important first step towards combatting the prescription overdose epidemic, but it won’t cure the problem overnight. In fact, according to researchers at Columbia University, the epidemic may not peak until 2017. They used Farr’s Law, which posits that an epidemic’s rise is mirrored in its decline, to predict that after reaching an all-time high of 50,000 annual deaths next year, drug overdoses will gradually decrease until 2035, when they will finally return to the level they were at in 1980, before the epidemic began.9

Time (and new CDC data) will tell whether their predictions are correct, but in the meantime, it’s up to doctors and patients to do all they can to ensure powerful prescription medications are used safely and sensibly.

Sources and Notes

The majority of the data for our analysis were exported from CDC WONDER’s Multiple Cause of Death database (http://wonder.cdc.gov/)

  1. ICD-10 codes were taken from previous CDC analyses.
    • Drug Poisoning Deaths: X40-X44, X60-X64, X85, Y10-Y14
    • Prescription Drugs: T36-T39, T40.2-T40.4, T41-T43.5, and T43.7 -T50.8
    • Opioid Analgesics: T40.2-T40.4
    • Heroin: T40.0 and T40.1
    • Cocaine: T40.5
  2. https://www.census.gov/popest/data/counties/totals/2014/
  3. http://sgwhc.org/resources/professional-education/case-studies/gender-prescription-opioid-abuse/#sthash.mQKVcgdd.LmWjWP42.dpuf
  4. http://www.cdc.gov/vitalsigns/prescriptionpainkilleroverdoses/
  5. Tracy, Melissa, et. al. “Circumstances of witnessed drug overdose in New York City: implications for intervention,” Drug and Alcohol Dependence 79 (2005): 181-182.
  6. http://newyork.cbslocal.com/2016/03/03/heroin-gateway-drugs/
  7. http://www.cdc.gov/vitalsigns/opioid-prescribing/
  8. http://www.cdc.gov/drugoverdose/prescribing/guideline.html
  9. https://www.sciencedaily.com/releases/2015/01/150106161940.htm