Congress has passed and President Trump has signed new legislation designed to turn back the opioid epidemic. The motivation for new legislation is clear: The opioid crisis is now one of the worst epidemics in U.S. history, having claimed the lives of more Americans than both world wars combined. But everyone should be mindful that because epidemics are dynamic, the impact of any new policy can change dramatically over time.
My colleagues, Allison Pitt and Margaret Brandeau, and I present a predictive model of the effects of various opioid-related policies in this month’s issue of the American Journal of Public Health. Because the model projects impacts for the first five years, as well as years 6 to 10, that a given policy is in place, it reveals how the effectiveness of policies is time-dependent.
For example, more tightly scheduling opioids so that they are harder to prescribe causes an enormous loss in population life years (1.33 million) in its first five years because of how it affects individuals currently addicted to prescription opioids. Some of those individuals will stop using opioids (including, in some cases, with the aid of addiction treatment), but others are projected to switch to heroin (including heroin combined with fentanyl) and to die of overdoses as a result. However, in years 6 to 10, the same policy has a positive impact on population life years as the benefits of not addicting the next generation of patients to opioids starts outweighing the impact of leading some already pill-addicted individuals to switch to heroin.
Two other policies, expanding access to the overdose rescue drug naloxone and to medication-assisted treatment (e.g., methadone, naltrexone) increase life years in both five-year periods. Unlike changes in prescribing policies, these policies only affect people who are already addicted. They add to population life expectancy immediately and in the long term. Our predictive model found this to be the case for all other prevalent services for opioid-addicted individuals (e.g., psychosocial treatments, needle-exchange programs).
If services targeted to already opioid-addicted people harm no one and are consistently beneficial to population health over time, why not only employ such policies and never change how opioids are prescribed? Simply put, epidemics don’t go away if the policy response is limited to treating already-ill individuals. To turn around the HIV/AIDS epidemic, preventive measures such as encouraging safer sex were essential, just as campaigns promoting hand-washing and covering one’s mouth while sneezing are essential for arresting flu epidemics. The same lesson applies to the opioid epidemic: Waiting passively for people to become addicted and then trying to help them will ensure the epidemic goes on forever, particularly given that the United States prescribes more opioids per capita than any other nation in the world, by a huge margin.
Because the number of people who could become addicted is much larger than the number who are addicted at any given moment, the potential gains from preventive steps have a larger impact as the years go by. To appreciate the power of prevention, consider that in the second five years of implementation, the life years gained from tighter scheduling of opioids (410,000) is larger than that of expanded naloxone and medication-assisted treatment combined.
Because the net effect on life years of the first five and the second five years combined is still negative over a decade, rescheduling opioids more tightly is probably a bad idea. But other opioid prescription-reducing policies addressed in our modeling research (e.g., cutting down on opioids for people with sprained ankles) return net health benefits more quickly and thus deserve consideration even if they have some short-term costs.