THE NATURAL HISTORY OF PRESCRIPTION-OPIOID AND HEROIN EPIDEMIC IN THE US

Purpose: Non-medical use of prescription opioid is a major public health problem, with nearly 11 million reporting using prescription opioids non-medically in 2014, and the mortality rate quadrupling from 1.5 to 6 deaths per 100,000 persons in the last decade.  The purpose of this study is to describe a mathematical model that explains the natural history and the relationship between prescription opioid and heroin use in the US.

Methods: We used a dynamic disease transmission (compartmental) modeling approach.  We defined the compartments by the level of dependency on opioid and heroin using the DSM-IV drug-dependence classification system.  In addition, we included compartments for individuals receiving medication assisted therapy (MAT) and death.  The model consists of a set of differential equations that describes flow among the compartments.  These flows are determined by a set of transition rates and the number of people in each compartment.  We used annual opioid and heroin use prevalence data from the drug-related mortality data from the National Survey on Drug Use and Health (NSDUH), and drug-related deaths data from the Mortality and Population Data System (MPDS).  The transition rates between the compartments were calibrated using a constrained optimization algorithm in MATLAB.

Results: The figure compares the observed annual drug-use prevalences and drug-related deaths to those predicted from the calibrated model.  Opioid- and heroin- use and dependence all continue to rise. Furthermore, we used this calibrated natural disease model to forecast these outcomes to 2016, which showed that 200 thousand individuals are estimated to become opioid users, and of the current opioid users, 175 thousand are expected to be dependent on or abuse opioids.  In addition, 70 thousand of those who currently abuse opioids are expected to become heroin users while 68 thousand to be dependent on or abuse heroin. Furthermore, nearly 39 thousand individuals are expected to die in 2016 due to opioid or heroin overdose.

Conclusions: Using a compartmental model that describes the natural history of opioid and heroin use, we were able to reproduce the observe annual prevalences and mortality data.  In addition, we were able to project these outcomes to future years.  This effort is required to evaluate the relative efficacy of various health policies and mitigation strategies to address this epidemic.