How do we decide priorities in medical research? Most Americans probably think scientists rely on abundant scientific data. They do, but just as often scientists follow hunches, interests and leads — without any guarantee of a result down the road.
This is surely part of the method — and romance — of scientific inquiry, but there may be more efficient ways to improve public health.
With Congress just having approved the 2005 budget for the National Institutes of Health prioritization of research is a timely subject. Geneticists Kathleen Merikangas of the National Institutes of Mental Health and Neil Risch of Stanford University have taken on this challenge by introducing an intriguing framework for prioritizing genetic research.
In their view, the best candidates for genetic research are disorders whose emergence and course cannot be derailed by change of personal habit or by environmental manipulation. Examples are autism, Type 1 diabetes and Alzheimer’s disease. While treatments may slow progress or provide relief, the long-term outlook for prevention and cure remains bleak. In contrast, lower priority on the genetic research hierarchy should go to conditions like Type 2 diabetes, alcohol or nicotine addiction. Type 2 diabetes, after all, can be avoided through exercise and weight loss and teens will buy less beer if taxes on alcohol are high enough. At the same time, a combination of social policy and tax levels has had a real impact on cigarette smoking.
Not surprisingly, the Merikangas and Risch proposal, which was published in Science last fall, continues to draw fire. Distinguished addiction researchers, led by the director of the National Institute on Drug Abuse, insist that drug and alcohol problems deserve a much higher ranking for genetic research funding. Substance abuse exerts a massive social burden of more than $500 billion annually, they point out. What’s more, they argued that discovering genes that determine receptor biology is critical to understanding addiction. “Those affected cannot afford to wait when advances in genetics will have a significant public health impact,” they wrote in rebuttal.
No one can dispute the high cost of substance abuse. But can genetic research make a sizeable dent in it? Probably not.
More relevant than cost, as Merikangas and Risch stress, is whether there are other mechanisms to reduce the damage even in the absence of genetic information. Imagine a front lawn dotted with several varieties of weeds. One species dies when lye is spread, a second diminishes when shade trees are planted, but the third is impervious to any change in soil or sunlight. If you want to use genetic research to facilitate weed elimination, starting with the third species makes the most sense because you don’t know any other way to get rid of it.
Environmental manipulations may not be as sexy as gene-based solutions, but we know they work. In the past 20 years, California has reduced smoking from 26% to 16% of adults through increased cigarette taxes, closer monitoring of sales outlets, restrictions of smoking in public places, endorsement of anti-smoking attitudes in the general public, and better decisions about health by current and prospective smokers.
Sweden’s alcohol policy provides an intriguing mirror image to the California case. In 1955, the country abandoned the limits on alcohol purchasing that had been in place for 35 years. Within one year, alcohol-related deaths increased by more than 25%, drinking-related criminal offenses by almost 100% and cases of delirium tremens by a stunning 438%. The Swedish government responded by sharply increases taxes on alcohol, and by 1958 drinking had dropped below its original level.
“Californizing” the country in a public health sense would reduce smoking to a much greater extent than a comparable investment of resources in genetics research. Within a generation’s time most of the people who continued to smoke despite every environmental barrier would be a group with a high genetic risk — the rest would be a small cohort of determined smokers who had no interest in quitting. At that point, investigating the genes of smokers might warrant a greater investment because those left would be a more highly genetically determined group. But for now, the lion’s share of resources for genetic research could be better directed toward diseases where society has no similarly potent environmental tools to wield.
Could genetic screening prevent addiction? Ideally, individuals of legal age could refuse cigarettes or alcohol if they knew their genes put them at higher risk for progressing from casual to compulsive use. (Though observing the habits of their first-degree relatives might be more convincing than a genetic test result). But screening can misfire, creating an unintended moral hazard; would fraternity members be more comfortable binging at the next party if they were told they did not have alcohol-risk genes?
In its defense, genetic research may one day improve addiction treatment. In their rebuttal to Merikangas and Risch the addiction researchers noted that therapeutic response of alcoholic patients to the medication naltrexone — an agent initially developed for heroin users– may be associated with a variant of a gene that codes for a specific brain receptor. If replicated, this finding might allow a clinician to use genetic information to predict whether he should offer naltrexone to a particular patient.
But future improvements in treatment from genetics research are unlikely to have much impact on the population because most recoveries happen without formal treatment. Several national surveys by the federal Substance Abuse and Mental Health Services Administration have found that less than 10% of the more than 20 million Americans with drug and alcohol problems receive treatment each year. For example, the National Institute of Alcohol Abuse and Alcoholism found that of a sample of over 3,300 once alcohol dependent adults who no longer had alcohol problems, three-quarters had never received treatment. What then, other than treatment, accounts for the recovery of so many addicts?
As Merikangas and Risch emphasize, substance abuse is malleable under the right circumstances. Recall that among U.S. soldiers who were addicted to heroin in Vietnam, only 12% maintained a heroin habit once back in the U.S. This is a striking example of a physiological process (drug dependence) interrupted by psychological and environmental processes — less need to manage intense anxiety or boredom associated with the war-zone, reduced availability of inexpensive heroin, and increased recognition that continued drug us posed too much of a personal cost.
Other examples abound. Government surveys show that binge drinking rates peak among 21year olds, and drop steeply by age 23, the period of life when college graduation, employment and marriage dramatically restructure the responsibilities and environments of daily life. Even among the addicted, particular environments can have powerful effects: When is the last time you saw a heavy-smoking parishioner light up during a religious service?
Finally and most critically, much of the public health harm of substance use has nothing to do with addiction. Basketball star Len Bias died in 1986 not because he was addicted, but because cocaine can induce sudden cardiovascular death. Improved treatments for alcoholism would not make our highways safe — most of the more than 32 million Americans who acknowledge drunk driving in the past year are non-addicted people who made a bad choice after drinking too much that night.
Improving treatment programs will little affect new users and non-addicted ones. In short, no matter what progress in genetics is made, addiction medicine will experience what every other specialty in the history of health care has learned: Treatment of individual cases never remotely approaches the benefits of environmental public health measures. One well-functioning city sewer system saves more lives than every available treatment for cholera ever could.
None of this should suggest that we abandon genetic research on addiction and substance abuse. In addition to the possibility of improving treatment, gene research on addictions may help us understand closely related traits, such as impulse-control and anxiety. We may also gain a better understanding of the basic reward structure of the human brain, which might be valuable in developing treatments for depression. But unlike the damage wrought by intractable diseases, major reductions in drug- and alcohol-related harm depend not on genes but on choices by policymakers and individual citizens.