Exponential growth of drug overdose poisoning and opportunities for intervention

Abstract

Overdose deaths in the United States (US) have been tracking along a dismayingly predictable exponential growth trajectory for more than four decades [1]. Yet, successes and failures of public health interventions are all too often measured in the short term, as annual swings in overdose deaths, with scant attention paid to this multi-year trajectory. Proper elucidations of the underlying dynamics of the long-term as well as the short-term trends are crucial to inform effective control and prevention policies. Figure 1 below shows the exponential trend of drug overdose deaths since 1979 and opioid overdose deaths since 1999—the earliest years of available data. Both are shown as linear trends on a log-scale. Over the years, there have been some deviations around this basic trajectory, with short-term year-to-year accelerations and slowings of the basic exponential rate. One recent example of a transient period of hyper-exponential growth followed by a temporary period of slowing with return to basic trajectory was the 2016-2017-2018 rise and fall in overdose (OD) deaths, due almost entirely to a transient surge in the availability of carfentanil. What was deemed as a political success in 2018 was in fact a return to the long-term exponential trend line [2]. The rapid increase in OD deaths in 2020 is another period of hyper-exponential growth, this one probably associated with the social isolation and restrictions caused by the COVID epidemic [3]. Before the epidemic, OD deaths returned to the exponential growth in 2018 and 2019. There is no certainty about how long this latest COVID-associated surge in the OD death rate curve will last. However, it is reasonable to expect that there will be a decline back down to the basic exponential curve sometime soon, regardless of what immediate specific control responses are put into effect. A decline in year 2021 to 86 000 OD deaths would put us back exactly on the long-term exponential growth curve and should not be taken as proof for the effectiveness of any specific policies. Although the total US curve of OD deaths has been on a long-term exponential trajectory, this overall national curve subsumes a shifting set of sub-epidemics composed of varying drugs, demographics, and locations [1]. One relatively stable set of sub-epidemics is based on birth year [4], perhaps because of social experiences and economic opportunities experienced by various birth-year cohorts [5]. These patterns can provide opportunities for targeted interventions. At this point, the exponential growth curve of US overdose deaths is a purely statistical observation, without clear evidence for how and why the annualized growth of total OD deaths has been maintained at 9% per year for over 40 years. The drivers of this growth remain largely unknown. Some research points to demand-side factors, such as increased economic inequalities, deteriorating employment opportunities, heightened sense of despair and alienation [6, 7] whereas other evidence implicates supply-side factors involving aggressive marketing of prescription opioids, substitutions, and increased availability of illicit opioids [8, 9]. The steady exponential growth of OD deaths could be the result of an auto-catalyzing economic cycle, because of increasing efficiencies of production and delivery, lower costs per effective dose, and increasing markets [7]. The forces driving this overall exponential growth may involve feedback loops of both demand and supply side factors. Lives are undoubtedly saved by increasing the availability of naloxone to revive persons who have overdosed, and by expansion of treatment slots with buprenorphine, but medical interventions like these are still severely limited and unlikely to alter the overall pattern of exponential epidemic growth. Moreover, the illicit market that plays an increasingly important role is understudied because of restricted access to law enforcement data. Therefore, it may be necessary to understand, at a mechanistic level, what sustains the exponential growth, and to target interventions to the critical feedback loops that drive these processes. Such critical interventions are more likely to be social and economic than medical. Quantitative understanding of what drives the US OD death exponential growth curve should be a high priority [10]. Any effective set of national goals for OD deaths should measure, track, forecast, and evaluate the impact of interventions on the trajectories of these drug, demographic, and location specific sub-epidemic patterns, in addition to the overall national aggregate trajectory. A federal level program of substance use modeling and forecasting should be established, similar to the new Center for Forecasting and Outbreak Analytics at the Centers for Disease Control and Prevention. This modeling and forecasting program should declare the integration of health data with social, economic, and law enforcement data as a top priority. H.J. is supported by National Institutes of Health/NIH/ National Institute on Drug Abuse, grant K01DA048985. None. Hawre Jalal: Conceptualization; data curation; formal analysis; funding acquisition; software; visualization. Donald Burke: Conceptualization.

Topics & Concepts