Drugs kill at least 30% of Americans.
You never hear this from drug policy reform groups because it shows that drugs are dangerous. And you never hear this from the government because it shows that our drug policies are a complete and utter failure.
Nevertheless, drug use and drug-related causes are the single largest underlying cause of death each year in the United States, as the government's own statistics show. Even if you don't consider yourself to be a drug user, all else being equal it's more likely that you'll die from drug use than from any other cause.
Deaths from drug use
There are at least 480,000 annual deaths from cigarettes and secondhand smoke. This is about 19.1%. And when considering products like cigars, pipes, and hookahs, the 2014 Surgeon General's report states that total mortality from any combusted tobacco product is likely approaching 500,000 deaths (19.9%) per year. The report concludes that "Due to the slow decline in the prevalence of current smoking, the annual burden of smoking-attributable mortality can be expected to remain at high levels for decades into the future, with 5.6 million youth currently 0 to 17 years of age projected to die prematurely from a smoking-related illness."
Alcohol kills an additional 88,000 Americans per year, about 3.6%. This includes chronic conditions, like alcoholic liver disease, and also acute causes, like motor-vehicle crashes. While many fewer Americans are killed from alcohol than from tobacco, the absolute percentages are somewhat deceiving; whereas the lifespan of current smokers is about 12 years shorter compared to those who have never smoked, the average death from alcohol comes a whopping 29 years prematurely.
For prescription drugs, there are an estimated 106,000 annual deaths (4.7%) from Adverse Drug Reactions (ADRs). This excludes fatalities that occur in nursing homes, despite the fact that the elderly have the greatest risk due to using the most drugs in the most combinations. In citing this statistic as the best estimate of ADR deaths, the FDA notes that, "The exact number of ADRs is not certain and is limited by methodological considerations. However, whatever the true number is, ADRs represent a significant public health problem that is, for the most part, preventable."
There is some evidence that fatalities from ADRs may actually be increasing; The percentage of Americans taking three or more prescription drugs increased from 12% to 21% over the past two decades, and the percentage of Americans taking five or more prescription drugs increased from 4% to 10% over the same time period. And according to the FDA, "The rate of ADRs increases exponentially after a patient is on 4 or more medications." The percentage of reported deaths due to ADRs increased by 50% between 1999 and 2006, and while reported ADR deaths don't necessarily correlate well with actual ADR deaths, the trend isn't exactly encouraging.
Since ADRs don't include overdose (poisoning) deaths, these need to be counted separately. According to the CDC, there were 22,134 overdose deaths from pharmaceutical drugs in 2010, and another 9,429 overdose deaths where none of the drugs involved were specified. This accounts for an additional 1.3% of deaths. And neither the ADR rate nor the overdose rate accounts for deaths from the long-term consequences of previous prescription drug use, e.g. someone who takes Accutane and dies of liver failure 20 years later.
Lastly, illicit drugs account for an estimated 17,000 deaths per year, or 0.7%. This includes not only overdose deaths, but also deaths indirectly caused by drug use such as those from motor-vehicle accidents, HIV infection, and hepatitis. This mortality rate is perhaps surprisingly modest given that by age fifty, approximately 90% of Americans are either current or former illegal drug users.
And yes, there really are zero overdose deaths from marijuana.
Deaths from non-use and non-adherence
In addition to deaths from drug use, there are also an enormous number of deaths due to people not taking drugs they should be taking. While superficially this may appear to be a completely separate problem, the underlying causes are largely the same: poor personal decision making with regards to drug use, combined with the lack of a high-quality and accessible public health infrastructure.
As an example, the American Heart Association recommends that patients without contraindications should take aspirin at the first sign of heart attack symptoms. Citing a 23% reduced mortality risk, the AHA estimates that if all patients not currently receiving aspirin were to do so, this would prevent 5,000 to 10,000 premature deaths per year. And although there is actually a fantastic product on the market that makes it easy to store aspirin in your wallet so that you're always prepared in the event a heart attack, it's sadly failed to gain traction in the U.S.
In addition to deaths due to non-use of the limited number of potentially lifesaving OTC drugs, there is the much larger issue of medical nonadherence. According to one highly-cited meta-analysis, the average nonadherence rate for drugs prescribed by a nonpsychiatrist physician is 21%. The author notes that while different conditions have different nonadherence rates, adherence "often is lowest among patients with the very conditions for which adherence would be most effective."
For hypertension alone, one recent study estimated if all untreated patients received effective medication and took it as prescribed, and if those patients achieved normal blood pressures, then 89,000 premature deaths per year could be prevented. Other chronic conditions such as diabetes, asthma, and heart disease have similar levels of nonadherence, each with their own associated mortality. In addition to chronic conditions, even acute conditions that can be readily treated have high levels of nonadherence. Fun fact: of patients who go to their doctor and get diagnosed with an STD, 23% don't bother to actually follow the recommended treatment.
It should be noted that unlike with deaths from drug use, the percentage of deaths from each source of drug non-use and nonadherence can't simply be added together; doing so would result in significant levels of double counting. This is true since, for example, many of the folks who die from not taking aspirin during a heart attack are only having a heart attack in the first place due to smoking.
All in all it's impossible to say exactly what percentage of Americans die prematurely from drug use, drug misuse, and drug-related causes. What we can say is that it's at least 30%.
What we can also say is that, in aggregate, drug-related deaths are by far and away the #1 underlying cause of death in the United States. If we look at the top two proximate causes of death, heart disease and cancer, and then subtract from each the cases where the underlying cause was alcohol or tobacco, then neither even comes close.
So why look at drug deaths through this lens, when the issues surrounding illicit drugs, licit recreational drugs, and prescription drugs are so different? Because our public policy in each area is completely broken. And while some of these issues do seem to be slowly getting better, progress is slow enough and inconsistent enough that relying on the government for salvation probably isn't a great bet.
So what to do?
While ostensibly exercise and proper diet are the best two things for your health, these are both difficult; In addition to having to learn what to do, you then have to actually spend time doing this stuff on a regular basis in order to benefit.
While I'm not suggesting by any means that one shouldn't bother with proper diet or exercise, or that they're a waste of time, what I would suggest is that hour-for-hour, the ROI from learning about drug use is probably much greater. This is because if you sit down and spend a few weeks reading books about drugs then that's it — from a risk-prevention perspective you're basically set for life.
For prescription drugs, books worth reading that focus mostly on public policy include The Truth About Drug Companies, by the former editor of the New England Journal of Medicine, and also Ben Goldacre's new book Bad Pharma. Another great book about this, Overdosed America, is written from the perspective of a practicing doctor trying to give sound medical advice to his patients.
For illicit drugs, there are books about harm reduction, like Marijuana is Safer: So Why Are We Driving People to Drink? There are also books on the potential health benefits of various illegal drugs. A good example of this is Neal Goldsmith's book Psychedelic Healing: The Promise of Entheogens for Psychotherapy and Spiritual Development.
And for medical nonadherence, Robin DiMatteo has a new textbook available called The Oxford Handbook of Health Communication, Behavior Change, and Treatment Adherence.
These books cover both not only the best practices for using various drugs, but also how our public policy shapes — and misshapes — the ways we use drugs. And of course these aren't the only books available; these days you'd be hard pressed to find a major class of drugs or medical condition without a highly rated book from one of the world's foremost experts.
Drugs are important; We're literally partly made of them. But what the statistics show is that most of us could be made of better ones. And while learning about drugs won't fix our public policy on its own, having a quorum of well-educated citizens is a necessary precondition.
We owe it to ourselves to become well-educated about these issues, and we owe it to society to take them seriously.
 The Health Consequences of Smoking — 50 Years of Progress, p.12, 649 - 679
 The percentage of deaths from each cause, except smoking, was calculated based on the total deaths in the year(s) when the absolute number of deaths from that cause was calculated. For tobacco deaths, the Surgeon General projects the same absolute number of deaths each year from 2010 - 2014, so the percentage was calculated based on the provisional data for 2011 deaths, the most-recent year available.
 CDC: Average for United States 2006-2010 Alcohol-Attributable Deaths Due to Excessive Alcohol Use
 21st-Century Hazards of Smoking and Benefits of Cessation in the United States
 Average for United States 2006-2010 Years of Potential Life Lost Due to Excessive Alcohol Use
 Incidence of Adverse Drug Reactions in Hospitalized Patients: A Meta-analysis of Prospective Studies.
 Preventable Adverse Drug Reactions: A Focus on Drug Interactions
 One of the most alarming aspects about our medical system is that we're missing large amounts of data on the safety and efficacy of even some of the most basic parts of our medical system. To quote a 1998 report on the quality of healthcare in the United States, "Perhaps the most striking revelation to emerge from this review is the surprisingly small amount of systematic knowledge available on the quality of health care delivered in the United States. Even though health care is a huge industry that affects the lives of most Americans, we have only snapshots of information about particular conditions, types of surgery, and locations of care."
 Health, United States, 2012, table 91
 Adverse drug reaction deaths reported in United States vital statistics, 1999-2006.
 Under-reporting of adverse drug reactions : a systematic review.
 Pharmaceutical Overdose Deaths, United States, 2010, via Drug Overdose in the United States: Fact Sheet
 Actual Causes of Death in the United States, 2000
 How Do They Measure Up? Examining Drug Use Surveys and Statistics, Part I: The Sources and also Part II: The Problems
 Monitoring the Future: National Survey Results on Drug Use 1975-2012 - Volume 2. Cf. p. 108 for lifetime prevalence of the sample population, chapter 3 for the methodology, and Appendex A in volume 1 for estimates of drug use prevalence across the total population. See also p. 101 of the 2009 report.
 The Validity of Self-Reported Drug Use: Improving the Accuracy of Survey Estimates
 Aspirin and Heart Disease
 Aspirin as a Therapeutic Agent in Cardiovascular Disease: A Statement for Healthcare Professionals From the American Heart Association
 The benefits of aspirin in acute myocardial infarction. Still a well-kept secret in the United States.
 Variations in patients' adherence to medical recommendations: a quantitative review of 50 years of research.
 Improving patient adherence: a three-factor model to guide practice
 The Value Of Antihypertensive Drugs: A Perspective On Medical Innovation
 Deaths: Leading Causes for 2010