The Contemporary American Drug Overdose Epidemic in International Perspective
2019; Wiley; Volume: 45; Issue: 1 Linguagem: Inglês
10.1111/padr.12228
ISSN1728-4457
Autores Tópico(s)Cardiac Arrest and Resuscitation
ResumoPopulation and Development ReviewVolume 45, Issue 1 p. 7-40 ARTICLEOpen Access The Contemporary American Drug Overdose Epidemic in International Perspective Jessica Y. Ho, Jessica Y. HoSearch for more papers by this author Jessica Y. Ho, Jessica Y. HoSearch for more papers by this author First published: 20 February 2019 https://doi.org/10.1111/padr.12228Citations: 36AboutSectionsPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinkedInRedditWechat Introduction Drug overdose mortality has reached unprecedented levels in the United States. Over the past two decades, drug overdose has more than tripled to become the leading cause of injury deaths in the US, outnumbering deaths from motor vehicle accidents and homicides according to data from the Centers for Disease Control and Prevention (CDC) / National Center for Health Statistics (NCHS) Compressed Mortality File and Cause of Death Files (full information in the list of References). The epidemic shows no signs of leveling off: drug overdose mortality continued to rise through 2017, amounting to over 70,000 deaths in that year and increasing by 16 percent per year between 2014 and 2017 (Hedegaard, Warner, and Miniño 2018). Traditionally, drug overdose and other injury deaths have been regarded as "background mortality," the expectation being that mortality from these causes of death should continue to decrease and ultimately decline to negligible levels over time (Bongaarts 2006) with improved public health and safety measures, economic growth, etc. The substantial increases in drug overdose mortality that have occurred in the US are both unanticipated and alarming, and the need for a comprehensive examination of cross-national differences in drug overdose mortality is urgent. There is widespread agreement that there is a shortage of cross-national research on drug policy, drug use, and drug overdose mortality; particularly relative to the scope of the issue (Kilmer, Reuter, and Giommoni 2015; Martins et al. 2015; Zobel and Götz 2011). In addition, it is well known that American life expectancy lags far behind other high-income countries (Crimmins, Preston, and Cohen 2011; Ho 2013; Woolf and Aron 2013), and that the US's international life expectancy rankings have deteriorated in recent decades (Ho and Hendi 2018; Ho and Preston 2010; Palloni and Yonker 2016). However, the contribution of the contemporary drug overdose epidemic to life expectancy differentials between the US and other high-income countries has not been established. The main aims of this study are (1) to situate the contemporary American drug overdose epidemic in broader perspective by comparing levels of, and trends in, drug overdose mortality in the US to 17 other high-income Organization for Economic Co-operation and Development (OECD) countries,1 including examining whether drug overdose mortality is differentially patterned by age and sex across countries, and (2) to quantify the contribution of drug overdose to the magnitude and widening of life expectancy differences between the US and these comparison countries. Background The contemporary American drug overdose epidemic Since the mid-1990s, drug-related hospital emergency department visits, substance abuse treatment admissions, and drug overdose mortality have increased dramatically in the United States. Increases in drug overdose mortality have been particularly sharp since 2010 and have continued through the present (Hedegaard, Warner, and Miniño 2018). Initially, the epidemic was primarily driven by prescription opioids, particularly the prescription painkiller OxyContin (Paulozzi et al. 2011). Following the release of abuse-deterrent formulations of OxyContin in 2010 (Cicero, Ellis, and Surratt 2012; Evans, Lieber, and Powell 2018) and wider recognition of the prescription opioid epidemic, the burden of drug-related mortality shifted increasingly to heroin and other synthetic opiates like fentanyl (Paulozzi et al. 2011; Hedegaard, Warner, and Miniño 2018; Jones, Einstein, and Compton 2018). The subgroups that experienced the largest increases in drug overdose mortality include non-Hispanic whites and the less educated (Ho 2017). While the US has experienced prior drug epidemics, its current epidemic is distinctive in three key aspects. First, the magnitude of the contemporary epidemic in terms of the estimated number of users and deaths involved far exceeds that of prior epidemics. Second, the earlier epidemics were driven primarily by illicit substances (heroin in the 1970s and cocaine in the 1980s to early 1990s), while legal drugs (prescription opioids) played the main role in initiating and sustaining the contemporary epidemic until the most recent decade. Third, drug overdose mortality was previously concentrated in major cities like New York, Philadelphia, Baltimore, and San Francisco, while the contemporary epidemic has encompassed dramatic increases in drug overdose mortality in nontraditional locations, particularly midsize cities, suburbs, and rural areas (Paulozzi and Xi 2008; Rigg, Monnat, and Chavez 2018). This has led to a convergence in drug overdose mortality so that drug overdose death rates do not differ substantially between rural areas and metros at the national level, although a large amount of geographic heterogeneity exists in these patterns (Rigg, Monnat, and Chavez 2018). The events leading directly up to the contemporary epidemic can be briefly summarized as follows: prior to the 1980s, the prevailing belief in the medical community was that few safe and effective methods to manage pain existed, and that opioid painkillers were too dangerously addictive to be prescribed except to terminally-ill cancer patients. By the 1990s, however, a fundamental change had occurred in the American medical establishment (Chiarello 2018; Meier 2003; Wailoo 2014). A new narrative dominated: millions of Americans were suffering needlessly from untreated pain; freedom from pain should be considered a universal human right (Brennan, Carr, and Cousins 2016; Cousins, Brennan, and Carr 2004; International Association for the Study of Pain 2018; Lohman, Schleifer, and Amon 2010) and pain should be accorded the status of the "fifth vital sign"; safe, non-addictive, and effective painkillers had been developed to treat pain; and doctors had a moral obligation to treat pain using these painkillers. Not only did the assessment, management, and treatment of pain become areas of increased and intense focus for medical practitioners, but also, using prescription opioids to treat many different types of non-cancer pain became common. The new importance given to recognizing and treating pain was reflected in the establishment of pain medicine as a subspecialty and the proliferation of pain management specialists: the first certificates in pain management were issued in 1993, followed by a rapid expansion of pain medicine training programs (Conrad and Muñoz 2010; Rathmell and Brown 2002). These trends occurred alongside important structural changes in the health care system in the era of managed care, during which primary care physicians faced increased financial pressures, patient caseloads, and time constraints. With physicians' employment and pay increasingly tied to patient evaluations, physicians had strong incentives to prescribe painkillers (Quinones 2015; Van Zee 2009). Purdue Pharma, the manufacturer of OxyContin, played a pivotal role in developing and popularizing the narrative that not only was there a moral obligation to treat pain, but that there now existed a safe and effective means of doing so. It marketed OxyContin—a pain reliever consisting of the opioid oxycodone—aggressively for a wide range of conditions including headaches, back pain, sports injuries, and wisdom tooth extraction (Meier 2003; Van Zee 2009). Purdue spent hundreds of millions of dollars (an estimated $200 million in 2001 alone [Goldenheim 2002]) on encouraging prescribing and promotional activities—including sponsoring pain management conferences and continuing medical education seminars (which many states require physicians to take to maintain their licenses)—during which their representatives touted that the risks of addiction were "less than one percent" (Meier 2003; Van Zee 2009). In the United States, painkiller prescriptions rose rapidly to unprecedented levels. In 1996, the year following its initial approval in the US, sales and prescriptions of OxyContin amounted to roughly $45 million and 316,786 prescriptions, respectively. In 2002, these figures reached $1.5 billion and seven million prescriptions (GAO 2003), corresponding to a 34-fold increase in sales and a 22-fold increase in prescriptions. Sales of all opioid pain relievers quadrupled between 1999 and 2013 (Paulozzi et al. 2011). These trends reflect excessive prescribing on physicians' parts as well as patients' demands. The two intersected in "pill mills," clinics where doctors prescribed enormous amounts of painkillers without medical justification and where clients could obtain pills onsite for cash. Individuals outside the medical establishment were also involved: they owned and ran pill clinics, notably in Florida (Lawson 2015; Temple 2015), and they acted as "sponsors" who recruited groups of users, took them to pain clinics, and paid for their appointments in return for painkillers, which they then resold on the black market (Macy 2018; Quinones 2015; Rigg, March, and Inciardi 2010). The huge amounts of pills entering the population were also fueled by "doctor shopping"—patients obtaining prescriptions for opioid painkillers simultaneously from multiple (as many as five or more) physicians (Hall et al. 2008; McDonald and Carlson 2013). As awareness of the epidemic grew, measures to limit prescribing were instituted; however, drug overdose mortality continued to rise. The huge amounts of prescribed opioids had created a large population of addicts who switched from opioid painkillers to heroin, a cheaper and more easily accessible alternative (Cicero et al. 2012; Evans et al. 2018; Muhuri, Gfroerer, and Davies 2013; Quinones 2015). Since 2010, drug overdose mortality has continued to increase, largely due to heroin and illegally-synthesized fentanyl (Jones, Einstein, and Compton 2018; Rudd et al. 2016). China has emerged as a main supplier of fentanyl to both the United States and Europe (EMCDDA 2018; U.S. Congress 2018). Cross-national differences in drug overdose mortality An important open question concerns whether the contemporary American drug overdose epidemic is an isolated phenomenon: is drug overdose mortality higher in the US than other high-income countries?2 Have other high-income countries experienced similar increases in drug overdose mortality, and are they likely to going forward? Increases in opioid prescribing have been noted in several other high-income countries including Australia, Canada, Denmark, Finland, Germany, Sweden, and the United Kingdom (Häkkinen et al. 2012; Hamunen et al. 2008; Hauser, Schug, and Furlan 2017; Hider-Mlynarz, Cavalie, and Maison 2018; Karanges et al. 2016; Rintoul et al. 2011; Roxburgh et al. 2017; Schubert, Ihle, and Sabatowski 2013; Weisberg et al. 2014; Zin, Chen, and Knaggs 2014). However, these increases have occurred much more slowly, have often been observed for different opioids (e.g., morphine or weak opioids instead of strong opioids like oxycodone), and are generally much smaller in magnitude than the increases observed in the United States. Extant studies have documented significant increases in opioid-related mortality only in Australia and Canada (del Pozo et al. 2008; Fischer et al. 2006; Lenton, Dietze, and Jauncey 2016; Marschall et al. 2016; van Amsterdam and van den Brink 2015; Weisberg and Stannard 2013). To date, no systematic cross-national comparison of the magnitudes of, and trends in, drug overdose mortality has been performed, and many studies have highlighted the paucity of literature on opioid abuse in Europe (e.g., Kotecha and Sites 2013; see Casati, Sedefov, and Pfeiffer-Gerschel 2012 for one exception). This study aims to shed light on which of three potential scenarios is likely to hold: 1 American exceptionalism, where the drug overdose epidemic proves to be a uniquely American scourge and other high-income countries escape largely unscathed, 2 the US as vanguard nation, at the forefront of an epidemic that eventually spreads to other high-income countries, or 3 the US as a cautionary tale, where other high-income countries at risk of developing epidemics potentially avoid or mitigate them by learning from the American experience. Data and methods Data Data from the Human Mortality Database (HMD 2017) and the World Health Organization Mortality Database (WHO) are extracted for the set of 18 countries starting with the year in which each country first adopted ICD-10 coding3 and ending with the most recent year for which data are available (which ranges from 2013–2015) (see Appendix Table A-1).4 Because the earliest and latest years for which data are available for all 18 countries are 2003 and 2013, many of the over-time comparisons focus on these years. For Canada (Statistics Canada 2017) and the US (NCHS 2018), I draw on additional data from these countries' vital statistics agencies to supplement the WHO and HMD data.5 These data are used to produce country-, year-, sex-, and age-specific drug overdose death rates between 1994 and 2015. Drug overdose deaths are defined as deaths for which the underlying cause of death was ICD-10 codes X40–X44, X60–X64, X85, and Y10–Y14 following the standard definition used by the US National Center for Health Statistics (Warner et al. 2011; Hedegaard, Warner, and Miniño 2018). These include deaths from both legal and illegal drugs and from deaths of all intents (i.e., drug-related accidental poisonings, suicides, homicides, and deaths of undetermined intent), and they exclude alcohol-related deaths. They include deaths from all drugs (not limited to opioids) since the objective is to capture the full extent of variation in the burden of drug overdose mortality across countries, which may be driven by different substances across time and across countries. Methods Age-specific drug overdose death rates are obtained by combining all-cause life table death rates from the HMD with the corresponding fractions of total deaths due to drug overdose from the WHO6: (1) where m is the death rate, D is the number of deaths, a is age, s is sex, t is year, and i is country. These are used to calculate age-standardized drug overdose death rates using the US population in 2000 as the age standard (SEER). To assess the role of drug overdose in the US life expectancy shortfall, cause-deleted life tables that answer the question, "What would life expectancy in each country be in the absence of drug overdose mortality?" are calculated (Preston, Heuveline, and Guillot 2001). Chiang's assumption is used to specify mortality in this counterfactual scenario; this is the standard and appropriate assumption given that drug overdose mortality predominates at ages at which overall levels of mortality are quite low (Chiang 1968; Preston, Heuveline, and Guillot 2001). Years of life lost due to drug overdose are specified as the difference between observed life expectancy and life expectancy from the cause-deleted life table. The contribution of drug overdose to the life expectancy gap between the US and any other country in a given year is calculated as: (2) The contribution of drug overdose to the change in the life expectancy gap between the US and another country between two time periods is calculated as: (3)where in a given time period, and t1 and t2 refer to time periods 1 and 2, respectively. Results Levels of and trends in drug overdose mortality Figure 1 shows age-standardized death rates (ASDRs) from drug overdose for men (Panel A) and women (Panel B) in the 18 countries from 1994–2015. Corresponding numbers for selected years are presented in Appendix Table A-2. Prior to the early 2000s, the US was not an outlier in terms of drug overdose mortality. However, the US has posted the highest drug overdose death rates among this set of countries in each year for over a decade—since 2002 and 2005 for men and women, respectively. In recent years, the US has pulled far above its peer countries. Based on the most recent two years for which data are available, drug overdose mortality appears to be trending upward for men in most (11 out of 18) countries. In contrast, it appears to be trending downward for women. Figure 1Open in figure viewerPowerPoint Age-standardized drug overdose death rates (p. 100,000), men (a) and women (b), 18 high-income countries, 1994–2015 NOTES: The 10 countries with the highest drug overdose death rates in 2013 are highlighted. The remaining 8 countries (Austria, Italy, Germany, Japan, Netherlands, Portugal, Spain, and Switzerland) are represented by unmarked lines. Figure 2 shows ASDRs from drug overdose for men (Panel A) and women (Panel B) by four country groupings in 2013, the most recent year for which data are available for all 18 countries. The four groups are: Anglophone, Nordic, countries with moderate levels of drug overdose mortality ("Medium"), and countries with very low levels of drug overdose mortality ("Low"). Anglophone and Nordic countries have higher levels of drug overdose mortality, while the Medium and Low groups have lower levels. The ranking of individual countries within each group is fairly similar, but not identical, for men and women. The ASDRs ranged from 0.60 (Japan) to 16.97 (US) per 100,000 for men and from 0.39 (Japan) to 10.51 (US) per 100,000 for women. On average, drug overdose mortality was 3.5 times higher in the US than in its peer countries, although this figure ranged from 1.6 to 28 times higher. What is particularly alarming is that even compared to the countries with the next highest death rates—the Nordic countries and other Anglophone countries—drug overdose mortality in the US is now nearly twice as high as in those countries. Figure 2Open in figure viewerPowerPoint Age-standardized drug overdose death rates (p. 100,000) by country groupings, men (a) and women (b), 18 high-income countries, 2013 Figure 3 shows the years of life lost (YLL) from drug overdose in each country in 2003 and 2013 for men (Panel A) and women (Panel B). In all of the Anglophone and Nordic countries except for Norway, YLL from drug overdose increased between 2003 and 2013. In 2003, YLL from drug overdose ranged from 0 (Italy) to 0.28 (US) for men, and from 0.02 (Italy) to 0.17 (US) for women. In 2013, these figures were 0.02 (Portugal) to 0.45 (US) for men and 0.02 (Italy) to 0.30 (US) for women. In both years, the US lost the most years of life from drug overdose among this set of countries; however, the difference in YLL between the US and the comparison countries increased dramatically over this decade. Figure 3Open in figure viewerPowerPoint Years of life lost from drug overdose by country groupings, men (a) and women (b), 18 high-income countries, 2003 and 2013 Contribution of drug overdose to the US life expectancy shortfall Given that the US now has much higher drug overdose mortality than other high-income countries, and that this difference has widened over time, it seems likely that the contemporary American drug overdose epidemic is contributing to the US life expectancy shortfall relative to the comparison countries. Figure 4 shows the percent contribution of drug overdose to the gaps in life expectancy at birth between the US and each of the 17 other countries in 2013 plotted against these gaps for men (Panel A) and women (Panel B). The corresponding numbers are presented in Appendix Table A-3. In general, the association between these measures is negative—the larger the US life expectancy shortfall, the less of it that tends to be due to drug overdose. The percent contributions are generally larger for men (ranging from 4–48 percent) than women (ranging from 4–23 percent). On average, life expectancy was roughly 2.6 years lower in the US than in other high-income countries in 2013 for both men and women, and drug overdose accounted for 12 percent and 8 percent of these 2.6-year gaps, respectively. Figure 4Open in figure viewerPowerPoint Percent of the US life expectancy gap due to drug overdose, men (a) and women (b), 17 high-income countries and average, 2013 How much of the gap is due to drug overdose is a function of several factors including the size of the gap, how high drug overdose mortality is in the comparison country, and at what ages it predominates. For example, drug overdose accounts for more of the US's shortfall relative to countries like Portugal and Germany partly because both of these countries have very low drug overdose mortality. Drug overdose accounts for little of the US-Sweden life expectancy differential in part because Sweden, like the US, has high drug overdose mortality. However, there remains a considerable degree of variation. For example, Denmark also has fairly high drug overdose mortality, but the contribution of drug overdose to the US-Denmark gap is nontrivial (16 percent and 23 percent for men and women, respectively). Turning to how drug overdose contributes to changes over time in the US life expectancy shortfall (Table 1), we see that US life expectancy gaps widened by 0.72 and 0.33 years on average for men and women, respectively, between 2003 and 2013. For three countries—Japan for men and women, and Germany and Sweden for women only—these gaps decreased; however, the predominant trend was for the US life expectancy shortfall to widen over this period. In the absence of drug overdose, it would have widened to a lesser degree. On average, the widening would have been one-fifth and one-third smaller for men and women, respectively. Among men, drug overdose made the largest contribution to the widening gap between the US and Germany (40 percent of the 0.43-year increase in the gap) and the smallest contribution to the widening gap between the US and the UK (only 5 percent of the 0.89-year increase). Among women, drug overdose made the largest percent contribution to the widening gap between the US and Australia (61 percent of the 0.14-year increase) and the smallest contribution to the widening gap between the US and Portugal (9 percent of the 1.31-year increase). There were three cases—Belgium, Norway, and Switzerland—where drug overdose accounted for the entirety of the widening. In other words, in the absence of drug overdose, life expectancy differences between American women and women in these three countries would have narrowed instead of increasing. Table 1. Contribution of drug overdose to the widening of the US life expectancy gap between 2003 and 2013, 17 high-income countries and average A. Men B. Women Country Δ (2003–2013) % due to Drug Overdose Δ (2003–2013) % due to Drug Overdose Australia 0.56 20 0.14 61 Austria 0.50 33 0.37 39 Belgium 0.54 28 0.06 + Canada 0.83 9 0.57 15 Denmark 0.94 13 0.61 16 Finland 0.70 15 0.37 23 France 0.91 19 0.44 31 Germany 0.43 40 −0.15 * Italy 0.90 16 0.42 29 Japan −0.11 * −0.27 * Netherlands 1.14 14 0.47 21 Norway 0.58 34 0.04 + Portugal 1.26 14 1.31 9 Spain 1.43 12 0.80 15 Sweden 0.16 35 −0.34 * Switzerland 0.58 22 0.12 + United Kingdom 0.89 5 0.64 16 Average 0.72 19 0.33 34 *Gap narrowed between 2003 and 2013 and would have narrowed further in the absence of drug overdose. +Gap would have narrowed instead of widening between 2003 and 2013 in the absence of drug overdose. In general, drug overdose's contribution to the widening of US life expectancy gaps between 2003 and 2013 was larger than its contribution to the magnitudes of these gaps at a point in time (in 2013). The contribution of drug overdose to widening US life expectancy gaps is sizeable. Life expectancy gaps increased between American women and women in 14 of these high-income countries, and drug overdose accounted for 15 percent or more of this widening in 13 of those 14 countries. In the remaining three countries, life expectancy gaps narrowed, and would have narrowed even further in the absence of drug overdose. Among men, life expectancy gaps widened between the US and 16 countries, and the contribution of drug overdose exceeded 15 percent in 10 of these 16 countries. Age patterns of drug overdose mortality The next set of figures plots age-specific drug overdose death rates for the US (Panel A), other Anglophone countries (Panel B), and Nordic countries (Panel C) in three periods: 1999–2001, 2006–2008, and 2013–2015. Starting with men (Figure 5), it is apparent that the age profile of drug overdose mortality has changed tremendously for American men over the past 17 years. In 1999–2001, during the early stages of the epidemic, drug overdose death rates in the United States were highest at ages 35–49. Over time, the age profile rectangularized: death rates are now uniformly high between ages 25–59 in the most recent period. This rectangularization was partly driven by sharp increases in drug overdose mortality at younger ages. Although the age profile of drug overdose mortality became younger over time for American men, it became older for men in the other Anglophone countries. In Australia, Canada, and the United Kingdom, the peak ages of drug overdose mortality increased. One of the most striking findings is the similarities between the US and the other Anglophone countries—but with varying time lags. Although overall levels of drug overdose mortality remain lower in Canada than the US, in terms of its age profile, Canada appears to be following a similar trajectory to the US but lagging behind by about seven years (i.e., its 2013–2015 profile strongly resembles the US's 2006–2008 profile, and Canada's profile also appears to be rectangularizing due to increases at younger ages). Australia and the UK's profiles in the most recent period strongly resemble the US's in the earliest period. Trends in age profiles of mortality are less clear for the Nordic countries. In general, these countries had flatter age profiles in the first two periods; by the most recent period, there appears to have been a shift toward younger ages in Sweden, Finland, and Norway. Figure 5Open in figure viewerPowerPoint Age-specific drug overdose death rates (p. 100,000), United States (a), Anglophone (b), and Nordic (c) countries, men Overall, age profiles of drug overdose mortality are older for women (Figure 6) than men. For American women, the profile has shifted to progressively older ages over time. While substantial increases in drug overdose mortality have also occurred at younger ages, the profile for American women remains unimodal and did not rectangularize as it did for American men. In general, trends were similar in the US and other Anglophone countries: the age profile appears to be getting older in all of these countries, although levels of drug overdose mortality remain much lower in the other countries than in the United States. Canada bears the closest resemblance to the US in terms of the shape of its age profile and trends, again offset by several years. The Nordic countries started out having the oldest age profile compared to the other countries in 1999–2001 and did not undergo much change. Women in the other countries converged to their age profiles. Figure 6Open in figure viewerPowerPoint Age-specific drug overdose death rates (p. 100,000), United States (a), Anglophone (b), and Nordic (c) countries, women Limitations Consistent data on drug overdose mortality are available for most high-income countries starting in the late 1990s. Ideally, we would want to examine trends starting in the early 1990s to concretely establish how drug overdose mortality in the US compared prior to the contemporary epidemic. However, data limitations prevent us from precisely identifying drug overdose deaths in years before countries started using ICD-10 coding. The next best strategy is to examine trends in drug-related accidental poisoning deaths (ICD-9 codes E850–E858 and ICD-10 codes X40–X44), a subset of drug overdose deaths that can be consistently defined across ICD changes. Trends in ASDRs from drug-related accidental poisoning (Appendix Figure A-1) are highly consistent with the trends in total drug overdose: the US had fairly high death rates from drug-related accidental poisoning in the 1990s, with death rates increasing in the mid- to late-1990s and pulling away sharply from the other countries since the 2000s. Data comparability is an important consideration in any cross-national comparative work. Here, the primary concern is whether significant differences in the coding of drug overdose deaths exist across high-income countries. If drug overdose deaths are undercounted in the US relative to other countries, differences between the US and other high-income countries would be even more dramatic than those found in this study. The estimates of the contribution of drug overdose to the US life expectancy gaps and to the widening of these life expectancy gaps would be underestimates. If drug overdose deaths are undercounted in other countries relative to the US, the opposit
Referência(s)