Harms of prescription opioid use in the United States

Pain has been long regarded as a stepchild area of medicine until recently. Prescription opioids (POs) are mainly used as analgesics in the treatment of mild to severe pain, including cancer and chronic non-cancer pain. In the context of expanded pain care, their consumption levels have tripled globally since 1990; however, this expansion has occurred almost exclusively in high-income countries, particularly the United States, which has ranked highest in per capita consumption of POs (based on standardized doses) for the past decade [1].

Coinciding with this expansion in consumption levels of POs in the United States have been several PO-related harms, namely non-medical prescription opioid use (NMPOU), PO-related morbidity and PO-related mortality. For example, prevalence of NMPOU during the past year has risen to almost 5% over the past decade [2]. Similarly, substance abuse treatment admissions for POs have increased from 28,326 in 2000 to 157,171 in 2010 [3], whereas there has been a fourfold increase between 1999 and 2010 in the number of drug poisoning deaths involving POs [4]. Therefore, it is hardly surprising that PO-related deaths now surpass deaths related to heroin and cocaine use combined [5].

Given previous experiences with other psychoactive substances, one perspective on this area of research, supported by emerging evidence from the United States [6‐8] and Canada [6, 9, 10], implicates increasing consumption levels of POs as the primary system-level driving factor behind the surging opioid epidemic. This short report aims to further this perspective by using published data for the United States from 2001 – 2010.

Figures 1 and 2 visualize the relationship between consumption levels of POs and each of NMPOU, PO-related morbidity and PO-related mortality measures included in the present study. Table 2 presents the results of the correlational analyses. Based on these data, moderate correlations were observed between consumption levels of POs and prevalence of NMPOU in the past month (r =0.74), past year (r =0.64) or lifetime (r =0.75). However, consumption levels of POs were strongly correlated with average number of days per year per person of NMPOU among the general population and NMPOU users (r =0.90 and 0.72 respectively). Furthermore, consumption levels of POs were very strongly correlated with morbidity and mortality measures, including substance abuse treatment admissions for POs per 10,000 people (r =0.95) and PO overdose deaths per 100,000 people (r =0.99).

This short report documented significant correlations between consumption levels of POs and PO-related harms, including NMPOU, PO-related morbidity and PO-related mortality. These correlations are corroborated by evidence from several other American [8, 9] and Canadian studies [7, 10, 11]. For instance, based on nationally representative data from the United States dating from 1995 to 2004, Wisniewski et al. also documented significant correlations between dispensing levels of two POs and NMPOU prevalence and emergency department visits [9].

However, the magnitude of three correlation coefficients observed in the present study were noteworthy, as near perfect correlations (r >0.90) were observed with consumption levels of POs: average number of days of NMPOU among general population, substance abuse treatment admissions for POs per 10,000 people and PO overdose deaths per 100,000 people. The strength of these correlations indicate linear increases in PO-related harms with increases in consumption levels of POs. Such near perfect correlations have been demonstrated previously for PO-related morbidity in the United States [6], but to our knowledge not for NMPOU or PO-related mortality. Importantly, these findings suggest that in comparison to prevalence estimates of NMPOU, average number of days of NMPOU may be better indicators of NMPOU in the United States. This may partly be due to the stabilization of NMPOU prevalence estimates over the past decade (p >0.05), and the parallel increases (p <0.05) in average number of days of NMPOU over the same time period.

Based on recent data, there seems to be a convergence of evidence implicating increasing consumption levels of POs as the primary system-level driving factor in the surging opioid epidemic. Interestingly, consumption levels of POs have continued to increase throughout the past decade, despite limitations in evidence regarding the effectiveness of pharmacotherapeutic treatment of cancer and chronic non-cancer pain with POs [14, 15]. For example, a review on the effectiveness of long-term opioid management for chronic non-cancer pain concluded that there was only weak evidence to suggest clinically significant reductions in pain [15]. Similarly, the evidence for effectiveness of POs for cancer pain is mixed [14]. Given this current state of the evidence, health policy must weigh benefits of increased consumption levels of POs against the PO-related harms, i.e. potential gains in addressing indications of pain vs. NMPOU, PO-related morbidity and/or PO-related mortality. There exists a need for rigorous reviews of clinical indications that legitimately warrant treatment with POs in comparison to other alternatives. Experts must debate whether there is a genuine need to treat as many individuals as currently treated with POs in the United States.

Irrespectively, the link between consumption levels and public health relevant harms of POs is similar to that observed with other psychoactive substances, including alcohol and tobacco [16, 17]. This link persists in the opposite direction as well when the availability of POs is decreased, i.e. reductions in availability coincide with reductions in PO-related harms [18‐20]. These findings emphasize that increased availability of psychoactive substances, including medications, should always be considered in a broader context, encompassing both a clinical and public health perspective.

A key limitation of the present study pertains to the inability to infer causation from the correlations presented. For example, closer surveillance due to increasing public health interest may be responsible for the observed increases in consumption levels of POs or PO-related harms, which would inevitably result in positive correlations. In summary, using data from the United States dating from 2001 – 2010, correlations between consumption levels of POs and PO-related harms were documented. These findings suggest that curbing consumption levels of POs at the population level may be an effective strategy towards limiting PO-related harms.