Scattered around the Web, mostly at sites dealing with drug addiction (here and here, for example), is a "Cocaine Timeline" that includes this 1912 milestone: "U.S. government reports 5,000 cocaine-related fatalities in one year." The number seems awfully high, given that the Drug Abuse Warning Network (DAWN) counted about 4,000 "drug misuse deaths" involving cocaine in 2004, when the U.S. population was about 294 million, three times what it was in 1912.
Although the DAWN data, based on reports from medical examiners in 31 metropolitan areas, do not cover the entire country, they include almost all of the big cities, and they tend to overstate the number of deaths caused by cocaine in each area. As a recent DAWN publication puts it, "It is important...to remember that not every reported substance is, by itself, necessarily a cause of the death or even a contributor to the death." A 2008 article in the journal Forensic Science International reinforces skepticism about deaths tagged as "cocaine-related," concluding that it is "very difficult to attribute a death to cocaine," that "isolated blood cocaine levels are not enough to assess lethality," and that "we can affirm that cocaine can be responsible for the cause of death only when there is a reasonably complete understanding of the circumstances or facts surrounding the death." Leaving such concerns aside, even if the actual number of cocaine-related deaths in 2004 was double the DAWN total, it would still be smaller as a share of the population than 5,000 deaths in 1912.
Such a disparity, of course, would be consistent with the view that the legal availability of cocaine, which was banned (except for medical use) by the Harrison Narcotics Act of 1914, led to widespread abuse and addiction, greater (in relative terms) than what we see today. But it turns out that the figure for cocaine-related deaths circa 1912 is fictitious. Surprisingly, the cocaine timeline was produced by Erowid, a site that is known for calm, rational advice about intoxicants, as opposed to anti-drug hyperbole. The original version of Erowid's timeline cited Steven Karch's A Brief History of Cocaine as the source for the estimate of cocaine-related deaths. In the 2006 edition of that book, Karch writes:
According to a U.S. government report, the total number of deaths from heroin and cocaine in the United States constitutes a smaller proportion of the total population now than it did in 1912, when the number exceeded 5000.
Karch then confusingly compares "the total number of deaths from heroin and cocaine" in 1912 to the number of "cocaine-related deaths" today. Given this juxtaposition, it's not hard to see how readers might conflate the two categories. But where did Karch get the first number? Although he does not identify which "U.S. government report" he has in mind, Gabriel Nahas' 1989 book Cocaine: The Great White Plague seems to provide the answer:
According to a 1912 official publication by M.J. Wilbert and M.G. Motter of the United States Treasury Department, cases of fatal poisoning, excluding those due to alcohol, numbered 5,000 in one year and the majority were related to opium or cocaine.
Note that we have gone from "cocaine-related deaths" to "deaths from heroin and cocaine" to "cases of fatal poisoning." Nahas cites Martin I. Wilbert and Murray G. Motter's Digest of Laws and Regulations in Force in the United States Relating to the Possession, Use, Sale, and Manufacture of Poisons and Habit-Forming Drugs, a report issued by the U.S. Public Health Service. A table on page 17 of that report provides mortality figures, based on Census Bureau data, for deaths by poisoning in the years 1900 through 1910. The highest total, including suicides but leaving out alcohol (2,578 deaths), "inhalation of poisonous gases" (1,837), lead (86), and "other occupational poisonings" (five), is for 1909, when it was 4,503. If you round that up, you get 5,000, which presumably is how Nahas arrived at his number. But the conclusion that most of these deaths "were related to opium or cocaine" appears to be Nahas'; the table does not say anything about either drug.
Since DAWN data indicate that cocaine is very rarely used for suicide, it's not clear that the 2,462 suicides by poison in 1909 should be included; leaving them out reduces the total number of fatal poisonings to 2,041. Of those, 1,779 (87 percent) were described as "acute poisonings," a category in which cocaine probably played a small role as well. The Forensic Science International report notes that "cocaine-related deaths occur for the major part after prolonged drug use." Fatal overdoses are more likely to involve opiates (often in combination with other depressants). And then we have to consider the role of every other poison (aside from alcohol, lead, and gases). It seems safe to say, based on the very report that Karch and Nahas cite, that the true annual number of cocaine-related deaths in the years leading up to the Harrison Narcotics Act was much smaller than 5,000. It might have been closer to 500.
After I emailed the folks at Erowid about this factoid, they corrected their timeline. But the number is still floating out there in many different places, including the website of the U.N.'s International Narcotics Control Board. No doubt drug warriors will continue to cite it as evidence of prohibition's success. But the persistence of this fake fact should not be seen as an indictment of accuracy in the Internet age. The original misinterpretations of the poisoning data, after all, appeared in books, and I never would have tracked down the error without an emailed tip from a psychopharmacology researcher in Norway and the help of Google, Google Books, and Amazon. Finding the 1912 report did require visits to brick-and-mortar libraries in other cities, which my colleagues Jesse Walker and Jeff Winkler were kind enough to help me with.