In an ideal world, data on illicit drug use, theoretical understanding of the causes of illicit drug use and policies to reduce the health and social problems associated with illicit drug use would be strongly integrated. There would be: (i) multiple sources of data that converged, (ii) strong (falsifiable) theories for interpreting the data and (iii) evidence-based policies. The paper by Caulkins et al. in this issue 1 illustrates how far we currently are from such a utopia. Caulkins and colleagues present a convincing case that, during the last decade, cocaine use has decreased dramatically (approximately a 50% decrease) and marijuana use has increased almost as dramatically in the United States (approximately a 30% increase). Caulkins and colleagues were able to develop a plausible explanation for the decrease in cocaine use. They believe that a decrease in the supply of cocaine to the United States was important, although they are not certain what supply factors, perhaps eradication efforts, perhaps interorganizational warfare among drugtrafficking organizations in Mexico, or perhaps in criminal organizations in Columbia switching from smuggling cocaine to smuggling gasoline, are causing this decrease in supply. With respect to the increase in marijuana use, Caulkins and colleagues do not select a most plausible explanation: ‘We do not knowwhy therewas such a large increase in the number of daily/near daily users’. It is interesting to note what Caulkins and colleagues did not consider as possible explanations. Our current dominant policy with respect to limiting the use of marijuana and cocaine is criminalization of the sale and use of these drugs. In the most recently available Federal Bureau of Investigation (FBI) arrest data there were 1 552 432 drug-related arrests (12.7% of all arrests in 2012); of the drug-related arrests, 6.1% were for heroin and/or cocaine possession and 42.4%were for marijuana possession. An additional 5.9% were for the sale or manufacture of marijuana. Caulkins and colleagues do not consider the possibility that cocaine use is decreased because arresting people for cocaine possession has been effective in reducing use, nor do they consider the possibility that marijuana use is increased because we are not arresting enough people for marijuana possession. Enough said and not said? The situation inwhichwe have imperfect but convincing data on important drug use trends has, at best, partial understanding of the trends, and where the trends are clearly not related to our dominant public policy with respect to illicit drug use sounds pretty dire. Could it be worse? Yes! Caulkins and colleagues note that the Arrestee Drug Abuse Monitoring Program (ADAM) (that conducts urine screens of arrestees and may be our best source of data on heroin, cocaine and methamphetamine use) 2 is about to be defunded. The lack of ADAM data would make it still more difficult to base public policy on drug use data, evenwhen therewas the political will to do so. The current situation suggests the need for a ‘doover’ at both the data collection and policy levels for psychoactive drug use in the United States. While ADAM should clearly be continued, it would be important to further develop data sources related to the public health aspects of psychoactive drug use. There are multiple examples of using health-related data to develop effective policies to reduce individual and societal problems associated with psychoactive drug use. Studies of the health effects of cigarettes, including second-hand smoke, led to restrictions on tobacco sales and use and to large reductions in cigarette smoking 3. Studies of HIV transmission among people who inject drugs have led to the implementation of state and local government-funded syringe exchange programs which have been associated with large reductions in HIV transmission 4. Data on the increase in opioid overdose deaths have led to the implementation of naloxone distribution in many areas and efforts to remove prescription requirements for naloxone 5. There are also theoretical concepts (social norms, self-efficacy for protecting one's own health, altruism for protecting the health of peers) for interpreting the data/linking the data to the effective policy changes. These policy changes were not easy; they often faced considerable resistance, and they have not been adopted uniformly across the United States. Nevertheless, they offer clear examples of cases where health-related drug use data have led to effective evidence-based policy. While the United States is currently far from a situation of (i) high-quality data on psychoactive drug use, (ii) falsifiable theories to interpret the data and (iii) evidence-based policies for addressing individual and societal problems associated with psychoactive drug use, shifting our emphasis to collecting more health-related data may provide the best route for moving in the direction of consistencies between data, conceptual understanding and evidence-based policies. None.