Mexico's Methamphetamine Precursor Chemical Interventions: Impacts On

RESEARCH REPORT doi:10.1111/j.1360-0443.2010.03068.x

Mexico’s methamphetamine precursor chemical

interventions: impacts on drug treatment admissionsadd_3068 1973..1983

James K. Cunningham1, Ietza Bojorquez2, Octavio Campollo3, Lon-Mu Liu4 & Jane Carlisle Maxwell5 Department of Family and Community Medicine,The University of Arizona,Tucson,AZ, USA1, Directorate General of Epidemiology, Ministry of Health, Mexico City, Mexico2, Center of Studies on Alcohol and Addictions, Universidad de Guadalajara, Guadalajara, Jalisco, Mexico3, Public Economics Research Center, College of Social Sciences, National Taiwan University,Taipei,Taiwan4 and Center for Social and Behavioral Research,The University of Texas,Austin,TX, USA5

ABSTRACT

Aims To help counter problems related to methamphetamine, Mexico has implemented interventions targeting pseudoephedrine and ephedrine, the precursor chemicals commonly used in the drug’s synthesis. This study examines whether the interventions impacted methamphetamine treatment admissions—an indicator of methamphetamine consequences. Design Quasi-experiment: autoregressive integrated moving average (ARIMA)-based intervention time–series analysis. Interventions: precursor chemical restrictions implemented beginning November 2005; major rogue precursor chemical company closed (including possibly the largest single drug-cash seizure in history) March 2007; precursor chemicals banned from Mexico (North America’s first precursor ban) August 2008. Settings Mexico and Texas (1996–2008). Measurements Monthly treatment admissions for methamphetamine (intervention series) and cocaine, heroin and alcohol (quasi-control series). Findings The precursor restriction was associated with temporary methamphetamine admissions decreases of 12% in Mexico and 11% in Texas. The company closure was associated with decreases of 56% in Mexico and 48% in Texas; these decreases generally remained to the end of the study period. Neither intervention was associated with significant changes in the Mexico or Texas quasi-control series. The analysis of Mexico’s ban was indeterminate due largely to a short post-ban series. Conclusions This study, one of the first quasi-experimental analyses of an illicit-drug policy in Mexico, indicates that the country’s precursor interventions were associated with positive impacts domestically and in one of the Unites States’ most populous states—Texas. These interventions, coupled with previous US and Canadian interventions, amount to a new, relatively cohesive level of methamphetamine precursor control across North America’s largest nations, raising the possibility that the impacts found here could continue for an extended period.

Keywords Methamphetamine, Mexico, Texas, precursor chemicals, intervention time series analysis, treatment admissions.

Correspondence to: James K. Cunningham, Department of Family and Community Medicine, The University of Arizona, 1450 N. Cherry Avenue, Tucson, AZ 85719, USA. E-mail: [email protected] Submitted 21 April 2009; initial review completed 29 June 2009; ﬁnal version accepted 15 March 2010

INTRODUCTION suppressing production of and thus problems attendant to the drug. This study examines whether these interven- Mexico has become a principal producer of methamphet- tions impacted methamphetamine treatment admissions, amine in North America [1]. Possibly in association, an indicator of methamphetamine consequences [9,10]. methamphetamine abuse in Mexico has increased in Following concerns about methamphetamine, Mexico recent years, a striking development given that the coun- conducted a study comparing legitimate need for meth- try’s drug consumption historically has been low and amphetamine precursor chemicals to actual import relatively stable [2–7]. In light of this, Mexico’s federal levels, and concluded that imports were excessive government has targeted the precursor chemicals [11–13]. In response, Mexico reduced its pseudoephe- used commonly in methamphetamine synthesis— drine imports from 132 826 kg in 2005 to 45 524 kg pseudoephedrine and ephedrine [8]—with the goal of in 2006 [12]. Monthly import data show that this

© 2010 The Authors, Addiction © 2010 Society for the Study of Addiction Addiction, 105, 1973–1983 1974 James K. Cunningham et al. reduction, which was thought to have placed imports methamphetamine-related hospitalizations in Califor- roughly in line with legitimate need [11], was achieved by nia, Arizona and Nevada [25]; and (v) methamphet- consistently limiting imports to a range of 150–7000 kg amine treatment admissions in California [10]. This last per month, a pattern which began in November 2005 finding suggests that if precursor controls in Mexico [14]. In comparison, during the preceding period of should prove effective, the country’s methamphetamine January–October 2005, imports ranged from 475– treatment admissions would also decrease. That said, 31 000 kg per month. the generalizability of US precursor control findings to In February 2006, Mexico expanded its precursor other countries is an open question, as recent research restrictions by requiring that (i) pseudoephedrine dis- indicates that precursor controls in Canada were tributors register customers in a federal computerized not associated with decreases in that country’s data system; (ii) pharmacies limit pseudoephedrine methamphetamine-related hospitalizations [26]. product sales to Յ60 mg per product per day; (iii) diver- This study uses autoregressive integrated moving sion, loss, theft, wastage or extraordinary sale of pseu- average (ARIMA)-based intervention time–series analy- doephedrine products be reported immediately; (iv) sis (a quasi-experimental procedure) to examine the monthly electronic input/output records of pseudoephe- possible impact of Mexico’s precursor chemical interven- drine sales be submitted to the government; and (v) tions on monthly methamphetamine treatment admis- federal officials, upon request, be given immediate access sions for the nation as a whole (1996–2008). Cocaine, to records created in the past year [15]. In an update, heroin and alcohol treatment admissions in Mexico are prescriptions were required to obtain any pseudoephe- used as quasi-control series. Methamphetamine use in drine product in a pharmacy beginning September 2007 Mexico is concentrated primarily in states within the [16,17]. northwest region of the country, and in states within the In March 2007, the Mexican government searched central region ([3] cf. [27,28]). Time–series (without the Mexico City home of Zhenli Ye Gon, head of Unimed ARIMA analyses) are presented for these two groups of Pharm Chem de Mexico—a commercial chemical states. company—and seized more than $200 million cash in Drugs produced/transported in Mexico are often des- US currency plus additional cash in other currencies: tined for US consumption [29], suggesting that Mexico’s possibly the largest single drug-cash seizure in history controls could also impact the United States. To consider [18,19]. Officials accusedYe Gon of using the company to this issue, methamphetamine, cocaine, heroin and illicitly import 60+ tons of pseudoephedrine [18]. Follow- alcohol treatment admissions in Texas are examined ing an international manhunt, Ye Gon was arrested in using ARIMA analysis. Texas is one of the most populous July 2007 [20]. The ability of Ye Gon and Unimed Pharm US states; its Mexico border is the longest of any US state Chem de Mexico to continue massive pseudoephedrine (1248 miles), and the bulk of the methamphetamine in smuggling (assuming the charges are accurate) was Texas reportedly originates in Mexico [30–33]. probably compromised significantly, if not ended, when US drug treatment admissions are often coerced by Ye Gon’s home was searched and the cash was seized. government legal/enforcement systems (e.g. courts), and Subsequent to the Ye Gon investigation, in September such admissions can reflect government administrative 2007 the Mexican government filed 41 criminal com- priorities more than drug-market dynamics [10]. In con- plaints involving multiple companies for the falsification trast, voluntary admissions (barring changes in treat- of 161 import permits related to pseudoephedrine and ment availability) are thought to better indicate actual other health supplies [21]. treatment demand, and thus drug-market dynamics In August 2008, Mexico banned pseudoephedrine [10,34]. Consistent with this, research has found that and ephedrine from the country, excepting as needed in voluntary methamphetamine treatment admissions in federally approved research, surveillance and toxicologi- California changed sharply in association with precursor cal analysis [22]. In so doing, Mexico superseded its controls, while coerced admissions had little such asso- prior import and pharmaceutical restrictions. It also ciation [10]. Unlike the United States, Mexico’s metham- became the first North American country to ban these phetamine treatment admissions are essentially chemicals. voluntary. For example, in 2006, only 4% of Mexico’s Mexico’s precursor control efforts have received methamphetamine admissions were associated with law little study. There has been research on US precursor enforcement [35], limiting the likelihood of a practical controls, however, and it indicates they were associated confounding effect. However, a substantial percentage of with decreases in (i) methamphetamine-related arrests Texas’ methamphetamine treatment admissions are in California [23]; (ii) methamphetamine purity in referred by the legal system—37% in 2006. Recognizing the United States [9]; (iii) methamphetamine inject- this, the analysis for Texas uses voluntary admissions ing, swallowing and snorting in California [24]; (iv) exclusively.

SUPPRESSION INTERACTIONS WITH (SISVEA), a national drug monitoring system [47]. Treat- TREATMENT ADMISSIONS ment client characteristics, including primary drug con- sumed, are recorded using an admission form. During the Cunningham & Liu have suggested that methamphet- study period (1996–2008), there were 67 408 metham- amine treatment admissions drop in association with pre- phetamine admissions, 84 827 cocaine admissions, cursor controls—a drug suppression policy—partly 66 005 heroin admissions and 103 339 alcohol admis- because the controls lessen methamphetamine-related sions in Mexico. In Mexico’s northwest region, the states problems, and thus motivation to seek treatment [10]. with relatively high rates of methamphetamine use (Baja Consistent with this, research indicates that treatment California Norte, Baja California Sur, Sonora, Chihua- seeking can rise and fall with problems related to drug hua, Durango and Sinaloa) had 61 544 methamphet- use [36]. And US precursor control-associated decreases amine admissions collectively. In Mexico’s central region, in methamphetamine-related hospital admissions and the states with relatively high rates of methamphetamine arrests [23,25] suggest reductions in methamphetamine- use (Jalisco, Michoacán, Mexico City/Distrito Federal related illness and criminal justice problems, respectively. and Hidalgo) had 2775 methamphetamine admissions Drops in methamphetamine treatment admissions collectively. following precursor controls (suppression) stand in Treatment admissions data in Texas come from the apparent contrast to studies which report that heroin Department of State Health Services’ state-wide treat- treatment admissions increased, at least initially, with ment admissions system, which also uses a form to record suppression. For example, in Lynn, Connecticut, heroin- characteristics of clients. Admissions in the system are related treatment admissions increased more than 85% publicly funded. During the study period, the system’s in the 10 months following the law enforcement closure voluntary treatment admissions numbered as follows: of an open-air heroin market [37]. In Zurich, Switzer- methamphetamine, 38 074; cocaine, 126 211; heroin, land, ARIMA-intervention time–series analysis indicated 65 197; and alcohol, 128 266 (note: both the Texas and that heroin-related treatment admissions increased Mexican systems originated with the Client Oriented Data sharply in association with the closure of an open-air Acquisition Process, an earlier version of the US Treat- drug scene [38]. In the Australian Capital Territory ment Episode Data Set). during the end of 1999 through the beginning of 2002, In ARIMA-based intervention analysis, interventions a decrease in heroin purity (indicating decreased heroin are identified a priori, have specific start-points and divide supply, due possibly to suppression or factors with the time–series into segments. The pre- and post- similar-acting effects [39]) was associated with a signifi- intervention segments of the series are examined to esti- cant increase in heroin-related treatment admissions mate the size of the intervention effects, and assess (correlation =-0.549) [40]. whether they occurred by more than random chance. This apparent contrast may be due partly to the ARIMA analysis adjusts for the presence of serially cor- nature of and treatment for methamphetamine and related errors among data points and appropriately heroin. That is, during times of suppression, heroin users adjusts for drift, trends, seasonality and outliers in the may have been drawn to treatment to obtain substitute series [48,49]. drugs (e.g. methadone) (cf. [41]), whereas no similar ARIMA-intervention models were obtained via the motivation for methamphetamine users exists, as their joint estimation method of model parameters and outlier treatment includes no substitute medication. Also, meth- effects [50,51], using SCA Statistical System software amphetamine users often turn to readily available drugs [52]. The admission series were non-stationary as indi- such as alcohol to help address withdrawal symptoms (cf. cated by the following: they did not exhibit a fixed mean [42]), which may further lessen their perceived need for level, and they had high, positive autocorrelations that treatment. In contrast, heroin users typically need other decreased slowly.To help achieve stationarity, each series opioids—drugs generally not readily available outside of was first-order differenced. The autocorrelation function treatment—to address withdrawal symptoms (cf. [43]). (ACF), partial autocorrelation function (PACF) and Finally, heroin withdrawal is associated with agitation extended autocorrelation function (EACF) of the differ- and insomnia [44,45], while methamphetamine with- enced series were used to identify the ARMA models. drawal is associated with depression and long periods of Sample autocorrelations and Box–Ljung Q-tests for the sleep [46], two factors that could limit activity, including first 24 lags indicated that all of the models had residuals possibly treatment seeking. consistent with white noise. Mexico’s precursor restrictions can be considered a METHODS two-step intervention, beginning in November 2005— Mexico’s treatment admissions data come from the the point when pseudoephedrine imports were reduced to Epidemiological Surveillance System of Addictions a level of legitimate need—and then followed in February

2006 with the domestic restrictions (e.g. customer docu- for Mexico overall, barring little apparent change at the mentation) described previously.Coding for this interven- time of the precursor ban (Fig. 1). Texas also had year- tion was 0 to November 2005, 1 during December 2005 end seasonal declines, albeit smaller than Mexico’s. to February 2006 and 2 thereafter. The quasi-control series (cocaine, heroin and alcohol TheYe Gon/Unimed Pharm Chem de Mexico interven- admissions) in Mexico (Fig. 2) and Texas (Fig. 3) changed tion will be referred to as the rogue precursor chemical little at the times of Mexico’s precursor restriction and company closure, with March 2007 as the start date. rogue company closure. Mexico’s quasi-control series all September 2007 and July 2008 are the start dates for appear to have declined around the time of the 2008 Mexico’s prescription requirement and methamphet- precursor ban; however, these were probably, in whole or amine precursor chemical ban, respectively. In addition, part, year-end (seasonal) declines. In contrast, Texas’ impacts were considered for US and Canadian interven- heroin and alcohol series exhibited little change at the tions—the US federal regulation of (i) pseudoephedrine time of the ban, although the state’s cocaine series may products implemented October 1997 and (ii) over-the- have declined. counter pseudoephedrine products (purchases limited to 9 g or less) implemented October 2001 [9]; and the ARIMA analysis Canadian federal regulation of (i) precursor imports/ exports implemented January 2003, (ii) precursor domes- The ﬁnal ARIMA model for Mexico’s methamphetamine tic distribution implemented July 2003 and (iii) essential treatment admissions series is: chemicals implemented January 2004 [53,54]. Texas’ ω ()− =+ω () − + ⎛ 2 ⎞ ()− + state precursor regulation (implemented August 2005) of 11BBYCtt111 I( − ) ⎝ ⎠ 1 B I21( t− ) 1−δ2B over-the-counter pseudoephedrine and ephedrine prod- 12 ω33()1− B I (ttt−1112) +−()11θθBB()− a ucts (placed behind the counter with purchases limited to

6 g per transaction and customer driver’s license and sig- where B is the backshift operator such that BYt = Yt-1; C is nature required and recorded) was examined in the Texas a constant term representing the overall trend of the series analyses [55,56]. These interventions were coded 0 series; at’s are independent random errors following to the month of initial implementation and 1 thereafter. a normal distribution; wi represents an intervention’s

initial effect; di represents the intervention effect’s impact

RESULTS rate; q1 is a non-seasonal moving average parameter; q12 is a seasonal moving average parameter; I is a two-step Descriptive analysis 1t function for Mexico’s 2005 precursor restrictions coded

In Mexico, relatively few methamphetamine treatment as described earlier; I2t and I3t are step functions for the admissions occurred during 1996–99 (Fig. 1). Metham- 2007 rogue company closure and 2008 precursor ban, phetamine admissions, however, generally rose from late respectively. I1t was associated with a downward shift

1999/early 2000 until the end of 2005/beginning of in methamphetamine admissions; I2t was associated

2006, the time of the precursor restrictions, and then with a gradual downward shift to a lower level; and I3t declined for about 1 year. Around March 2007, the time was associated with an abrupt downward shift (Table 1). of the rogue precursor chemical company closure, admis- No significant impacts were found at the times of the sions declined again, and remained lower to the end of US and Canadian regulations or the Mexican September the study period. Whether admissions shifted at the times 2007 prescription requirement, in this or any other of the methamphetamine prescription intervention in model. As such, these interventions were not included September 2007 or the precursor ban in August 2008 is in the final models [49]. In ARIMA analyses for each of difficult to assess visually,as both were followed shortly by Mexico’s three quasi-control series (cocaine, heroin and substantial year-end (‘seasonal’) declines (note the sub- alcohol) statistically significant declines were found at stantial year-end declines beginning in 2001). the time of the ban, but not at the times of the precursor The states in Mexico’s northwest region accounted for restriction or the rogue company closure (equations most of the methamphetamine treatment admissions in and parameter estimates for the quasi-control series are Mexico (Fig. 1). The series in this region generally paral- not presented). leled that for Mexico overall. The series for states in Mexi- The final ARIMA model for Texas is: co’s central region exhibited substantial fluctuation/ ω ()− =+ω () − + ⎛ 2 ⎞ ()− + variance, probably because monthly admission counts 11BBYCtt111 I( − ) ⎝ ⎠ 1 B I21( t− ) 1−δ2B were relatively small. Barring that, the series was fairly 12 ω ()1− B I ( − ) +−()11θθBB()− a similar to the series for Mexico overall. 33ttt1112

Texas’ series of voluntary methamphetamine treat- I1t was associated with a downward shift in methamp- ment admissions was also generally similar to the series hetamine admissions. I2t was associated with a gradual

Mexico’s Methamphetamine Interventions

Mexico: Methamphetamine Treatment Admissions 1250 Canada essential Mexico chemical regulation precursor Canada chemical precursor ban 1000 import/export US pseudoephedrine regulation retail sales regulation 750

Mexico precursor 500 Admissions chemical restriction Mexico: rogue US pseudoephedrine precursor Mexico 250 product regulation chemical precursor Canada precursor company prescription domestic closure requirement distribution regulation 0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Months

States in Northwest Mexico: Methamphetamine Treatment Admissions 1100 Mexico Canada precursor 1000 US pseudoephedrine precursor chemical 900 retail sales import/export ban 800 regulation regulation

700 Mexico 600 precursor US pseudoephedrine chemical 500 product regulation restriction

Admissions 400 Canada essential chemical regulation Mexico: rogue 300 Mexico precursor precursor 200 chemical Canada precursor prescription company 100 domestic requirement distribution regulation closure 0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Months

States in Central Mexico: Methamphetamine Treatment Admissions 100 Canada essential Mexico Canada chemical regulation precursor 90 US pseudoephedrine precursor prescription 80 retail sales import/export requirement Canada precursor regulation regulation 70 domestic distribution Mexico 60 regulation precursor Mexico US pseudoephedrine chemical ban 50 precursor product regulation 40 chemical Admissions restriction 30 Mexico: rogue 20 precursor 10 chemical company closure 0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Months

Texas: Voluntary Methamphetamine Treatment Admissions 700 Canada essential Mexico chemical regulation 600 Canada precursor precursor chemical ban 500 import/export regulation US pseudoephedrine Mexico 400 retail sales precursor Texas regulation chemical precursor 300 restriction

Admissions retail sales US pseudoephedrine regulation 200 product regulation Mexico: rogue Mexico Canada precursor precursor precursor domestic chemical 100 prescription company distribution requirement regulation closure 0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Months

Figure 1 Methamphetamine treatment admissions in Mexico, states in Mexico’s northwest region*, states in Mexico’s central region** and Texas, USA (1996–2008). *Baja California Norte, Baja California Sur, Sonora, Chihuahua, Durango and Sinaloa. **Jalisco, Michoacán, Mexico City/Distrito Federal and Hidalgo

downward shift (Table 1). No significant impact was restriction (I1t) occurred fairly close in time, raising the found for I3t.Texas’ regulation was not included in the final possibility that I1t somehow masked theTexas regulation’s model as it was not found to be significant. Note thatTexas’ impact. In analyses not detailed here, however, the Texas

August 2005 regulation and Mexico’s November 2005 regulation was insigniﬁcant even when I1t was excluded

Mexico’s Methamphetamine Interventions

Mexico: Cocaine Treatment Admissions Mexico 1500 precursor Canada essential chemical Canada 1250 chemical regulation ban precursor import/export 1000 US pseudoephedrine regulation retail sales regulation 750 Mexico

Admissions US pseudoephedrine precursor 500 Mexico: rogue product regulation chemical restriction precursor Mexico chemical precursor 250 Canada precursor company prescription domestic closure requirement distribution regulation 0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Months

Mexico: Heroin Treatment Admissions 1000 Mexico Mexico precursor precursor chemical US pseudoephedrine chemical restriction 800 retail sales ban regulation

600

US pseudoephedrine product regulation 400 Canada essential Admissions chemical regulation Mexico: rogue Canada precursor Mexico 200 precursor chemical precursor Canada precursor import/export prescription domestic company regulation closure requirement distribution regulation 0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Months

Mexico: Alcohol Treatment Admissions 2500 Mexico precursor Canada chemical 2000 precursor restriction import/export regulation Mexico 1500 Canada US pseudoephedrine essential precursor retail sales chemical chemical regulation ban 1000 regulation Admissions Mexico: rogue US pseudoephedrine precursor Mexico 500 product regulation chemical precursor Canada precursor company prescription domestic closure requirement distribution regulation 0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Months

Figure 2 Cocaine, heroin and alcohol treatment admissions in Mexico (1996–2008) from the model, indicating that masking was not an issue. Percentage change was estimated here by comparing In analyses for the Texas quasi-control series, no impacts the average number of admissions during the 3 months for any interventions were found (equations and para- preceding an intervention to the change in admissions meter estimates for these series are not presented). indicated by the model parameter estimates (Table 1) Estimates in Table 1 for the 2005 Mexico restriction (note: to help avoid artifact associated with the year-end and the abrupt impacts can be interpreted directly. For drops, a 2-month average was used for the March 2007 example, the estimate for the impact associated with company closure). In Mexico overall, the estimated Mexico’s precursor restriction indicates a downward shift reductions for the 2005 precursor restriction, 2007 of 141.39 admissions in the series. The approximate rogue company closure and 2008 ban are 12%, 56% and size of the eventual effects of the gradual impacts 15%, respectively. In Texas, the estimated reductions for can be assessed using the formula w/(1 - d) [49]. For the Mexico precursor restriction and the company example, the approximate eventual impact in Mexico closure are 11% and 48%, respectively. associated with the company closure is: -152.23/ (1 - 0.75) =-608 (an eventual downward shift of 608 DISCUSSION admissions in the series). Estimates of percentage changes in a series related In Mexico, methamphetamine treatment admissions to interventions can be useful for policy planning [49]. decreased in association with both the country’s

Mexico’s Methamphetamine Interventions

Texas: Voluntary Cocaine Treatment Admissions 1400 Canada essential Mexico chemical regulation precursor 1200 chemical US pseudoephedrine ban retail sales 1000 regulation

Mexico: rogue 800 precursor Texas chemical precursor retail 600 company

Admissions sales regulation Canada closure 400 precursor Mexico US pseudoephedrine Canada precursor Mexico import/export precursor product regulation domestic precursor 200 regulation prescription distribution chemical requirement regulation restriction 0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Months

Texas: Voluntary Heroin Treatment Admissions 800 Mexico precursor Canada essential chemical 700 US pseudoephedrine chemical regulation retail sales restriction 600 regulation Mexico precursor 500 chemical ban 400 Texas precursor retail Admissions 300 Canada sales regulation Mexico: rogue precursor 200 precursor Mexico import/export Canada precursor US pseudoephedrine chemical precursor regulation domestic 100 product regulation company prescription distribution closure requirement regulation 0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Months

Texas: Voluntary Alcohol Treatment Admissions 1600 Mexico precursor 1400 US pseudoephedrine Canada essential chemical retail sales chemical regulation restriction 1200 regulation Mexico 1000 precursor chemical 800 Texas ban precursor retail Admissions 600 sales regulation Canada Mexico: rogue precursor precursor 400 Canada precursor Mexico import/export chemical US pseudoephedrine domestic precursor regulation company 200 product regulation distribution prescription closure regulation requirement 0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 Months

Figure 3 Cocaine, heroin and alcohol treatment admissions in Texas (1996–2008)

Table 1 Mexico precursor chemical control impacts on methamphetamine treatment admissions in Mexico and Texas: model parameter estimates.

Mexico Texas

Parameter Estimate t SE Estimate t SE

Mexico precursor chemical restriction: November 2005 w1 -141.39 -5.34 26.49 -67.48 -3.54 19.04

Rogue precursor chemical company closure: March 2007 w2 -152.23 -5.93 25.66 -29.34 -2.62 11.22

d2 0.75 13.56 0.06 0.90 15.85 0.06

Mexico precursor chemical ban: August 2008 w3 -120.99 -2.05 59.03 -15.96 -0.50 31.22 ARIMA C 7.12 2.00 3.56 4.79 2.49 1.92

q1 0.64 8.97 0.07 0.41 5.26 0.08

q12 -0.79 -14.87 0.05 -0.25 -2.85 0.09

SE: standard error; w: intervention impact; d: impact rate; q1: moving average parameter; q12: seasonal moving average parameter; C: trend; ARIMA: autoregressive integrated moving average.

© 2010 The Authors, Addiction © 2010 Society for the Study of Addiction Addiction, 105, 1973–1983 1980 James K. Cunningham et al. precursor restrictions and the closing of a major rogue This is one of the first quasi-experimental analyses of precursor chemical company. The decreases appeared to an illicit-drug policy in Mexico. While the research design be specific to methamphetamine, as no concurrent used is considered relatively powerful [57,58], this study changes in cocaine, heroin and alcohol treatment admis- has limitations. For example, it is possible that unidenti- sions were indicated. These findings constitute the first fied factors (e.g. undocumented changes in methamphet- evidence outside the United States that a country’s pre- amine enforcement or other efforts by the Mexican cursor chemical controls can be associated with positive government) occurred at the times of Mexico’s interven- domestic impacts (cf. [26]). tions, possibly resulting in the impacts found. Treatment Compared to the decrease following Mexico’s precur- admittees may not be representative of the general drug sor restrictions (-12%), the decrease following the rogue user population. SISVEA and the Texas treatment admis- company closure was substantially larger (-56%). sions system, being administrative in nature, have more However, this larger decrease may not have been associ- error than research data systems. ated solely with the closure. Rather, it is possible that the Consistent with this study’s finding of no significant company, while operating, limited the restrictions’ changes in cocaine, heroin and alcohol admissions at the impact; and the decrease following the closure may have times of Mexico’s precursor restrictions and rogue been in large part the realization of the restrictions’ full company closure, multiple other studies have found that impact on admissions, one delayed temporarily by the methamphetamine precursor impacts were unaccompa- company’s illicit import activities. nied by changes in cocaine or heroin purity, arrests or The analysis indicated that methamphetamine admis- treatment admissions [9,10,23,59]. Also note that sions in Mexico declined at the time of the 2008 precur- behavioral economic research by Sumnall et al. [60] in sor ban. This finding is indeterminate, however, as the the United Kingdom found no evidence that polysub- analysis also indicated ban-related declines in the coun- stance users would increase cocaine purchases when try’s quasi-control series. Moreover, the ban’s post-series amphetamine prices increased. These findings call into was short (4 months) and concluded with a pronounced question the so-called ‘water-balloon theory’, the notion year-end seasonal drop (for methamphetamine and the that suppression of one drug leads simply to increased other drugs). This may have confounded estimates for the problems with another (cf. [10]). ban, leading possibly to an artifactual indication of sig- Corruption in Mexico, a widely discussed issue (e.g. nificant declines for all the drugs [49]. To resolve this [1,61,62]), may have presented challenges to the execu- issue, analyses using series extending beyond 2008 will tion of Mexico’s precursor controls. In particular, the be needed. rogue chemical company may have been able to import The precursor restriction and rogue company closure massive amounts of precursor chemicals due partly to impacts in Texas essentially paralleled those in Mexico, cooperation from corrupt officials (cf. [21]). If such col- indicating that Mexico’s interventions were associated lusion occurred, the closure of the company would seem with positive international impacts as well, a finding with to underscore the Mexico federal government’s efforts to little precedent in illicit-drug policy research [9,26]. No address corruption and realize policy goals (cf. [63]). ban impact was found in Texas. However, due to the ban’s This study, together with previous research, indicates short post-series this finding will need confirmation using that Mexican and US precursor controls were associated series extending beyond 2008. (Note: while the United with pronounced positive domestic impacts, while Cana- States’ 1997 regulation had a significant impact on da’s were not. A possible explanation is that metham- methamphetamine treatment admissions in California phetamine production in Mexico and the United States [10], no such impact in Texas was found. Texas’ volun- was substantially greater than that in Canada [1] and, tary methamphetamine admissions generally followed other things being equal, demonstrating impacts may be the pattern of California’s, but in 1997 the volume of more feasible when production is on a larger scale. Texas’ admissions was small, possibly limiting measure- Another possible factor is that Canada’s controls may ment sensitivity.) have coincided with a shift from local producers to sophis- Significant declines in methamphetamine treatment ticated criminal organizations readily positioned to admissions were not found at the times of Mexico’s pre- smuggle methamphetamine and precursor chemicals scription requirement and Texas’ state precursor regula- from Asia [26]. In contrast, when US precursor controls tion. Both interventions largely targeted precursors sold were implemented, US producers required time to locate in over-the-counter medicinal products. Multiple studies alternative sources of chemicals [9,23–25,64]. This, too, indicate that federal interventions targeting over-the- may be the case now facing Mexico’s producers [65]. counter products have had little or no impact on meth- There are reports of methamphetamine producers amphetamine hospitalizations, arrests, route of attempting to circumvent Mexico’s interventions by administration or purity [9,23–25]. turning to Central and South America for precursor

© 2010 The Authors, Addiction © 2010 Society for the Study of Addiction Addiction, 105, 1973–1983 Mexico’s methamphetamine interventions 1981 chemicals ([66], cf. [67]). This notwithstanding, Mexico’s 11. Secretaria de la Funcíon Pública (SFP). Informe de Rendición interventions, coupled with those of the United States de Cuentas de la Administración 2000–2006. Clave Presupues- and Canada, amount to a new, relatively cohesive level of taria SOO. Nombre: Comisión Federal para la Protección contra Riegos Sanitarios [Report on Accountability of Management methamphetamine precursor control across North 2000–2006. Budget key SOO. Name: Federal Commission for America’s largest nations, raising the possibility that the Protection against Health Risks]. México, DF: SFP; 2006. impacts found here could continue for an extended 12. International Narcotics Control Board. Report of the Interna- period. tional Narcotics Control Board for 2005. New York: United Nations; 2006, p. 29. 13. Tandy K. P. Statement of Karen P. Tandy, Administrator, Drug Declarations of interest Enforcement Administration, U.S. Department of Justice. Regarding International Meth Trafficking. Before the Senate None. Foreign Relations Committee, Subcommittee on International Economic Policy, Export and Trade Promotion, and Subcommit- Acknowledgements tee on Western Hemisphere, Peace Corps and Narcotics Affairs, 21 June 2006. Washington, DC: Drug Enforcement Admin- We wish to thank the Directorate General of Epidemiol- istration; 2006. ogy, Ministry of Health, Mexico (especially Mario Cortes, 14. Comisión Federal para la Protección contra Riesgos MSC) and the Texas Department of State Health Services Sanitarios (COFEPRIS). Control de la Pseudoefedrina en for the use of their data. James Cunningham’s work on México [Control of Pseudoephedrine in Mexico]. México, DF: COFEPRIS; 2007. this study was supported by the Mexico–US Fulbright– 15. Comisión Federal para la Protección contra Riesgos Sani- García Robles Program and the Canada–US Fulbright tarios (COFEPRIS). Convenio de Concertación de Acciones para Program, as part of the Fulbright Canada–Mexico Joint un Efectivo Control de Medicamentos que Contienen Pseudoe- Award in North American Studies. Jane Maxwell’s work fedrina [Agreement of Concerted Actions for Effective Control of was supported by grant R21 DA 025029 from the Drugs Containing Pseudoephedrine]. México, DF: COFEPRIS; 2006. National Institute on Drug Abuse. The study authors are 16. Cordova Villalobos J. A. Entrevista al Secretario de Salud, José solely responsible for all findings and conclusions. Ángel Córdova Villalobos, al termino de XXII Congreso Interna- cional de Farmacias [Interview with Secretary of Health Jose References Angel Cordova Villalobos following the XXII International Con- gress of Pharmacists]. México, DF: Secretaría de Salud; 1. United Nations Office on Drugs and Crime. 2008 World Drug 2007. Report. New York: United Nations; 2008. 17. Fármacos. Pseudoefedrina: se prohíbe su comercialización 2. Brouwer K. C., Case P., Ramos R., Magis-Rodríguez C., en México [Pseudoephedrine: commercialization is prohib- Bucardo J., Patterson T. L. et al. Trends in production, traf- ited in Mexico]. Fármacos 2008; 11: 21–2. ficking, and consumption of methamphetamine and 18. Procuraduría General de la República (PGR). Boletín 121/ cocaine in Mexico. Substance Use Misuse 2006; 41: 707–27. 07. Los 205 millones de dólares asegurados por la Policía 3. Instituto Nacional de Salud Pública. Encuesta Nacional De Federal, producto de las transacciones ilegales de pseudoefedrina Adicciones 2008 [National Survey on Addictions 2008]. Cuer- con empresas de China [Bulletin 121/07. 205 million dollars navaca, Morelos: Instituto Nacional de Salud Pública; secured by the Federal Police, a product of illegal pseudoephedrine 2008. transactions with companies in China]. México, DF: PGR; 4. Medina-Mora M. E., Guiot E. R. Demand of drugs: Mexico in 2007. the international perspective. Salud Mental 2003; 26: 1–11. 19. McCampbell C. A. Statement by Christy A. McCampbell, 5. Maxwell J. C., Cravioto P., Galván F., Cortés M., Wallisch L. Deputy Assistant Secretary, Bureau for International Narcotics S., Spence R. T. Drug use and risk of HIV/AIDS on the and Law Enforcement Affairs, U. S. Department of State. Testi- Mexico-USA border: a comparison of treatment admissions mony before the Senate Finance Committee September 18, in both countries. Drug Alcohol Depend 2006; 82: S85–S93. 2007. Bilateral and Multilateral Efforts to Combat the Interna- 6. Tapia-Conyer R., Medina-Mora M. E., Sepúlveda J., De la tional Production and Trafficking of Methamphetamine. Wash- Fuente R., Kumate J. La encuesta nacional de adicciones de ington, DC: US Government Printing Office; 2007: 86–95. México [National survey on addictions of Mexico]. Salud 20. Procuraduría General de la República (PGR). Boletín 338/ Publica Mex 1990; 32: 507–22. 07. La PGR tiene un caso jurídico sólido y contundente en contra 7. Lupsha P. A. Drug trafficking: Mexico and Columbia in de Zhenli Ye Gon por delitos contra la salud en varias modal- comparative perspective. J Int Affairs 1981; 35: 95–115. idades; delincuencia organizada; posesión de armas de fuego y 8. Haislip G. R. Methamphetamine Precursor Chemical Control in cartuchos reservados para las fuerzas armadas; y operaciones con the 1990s. Washington, DC: Drug Enforcement Administra- recursos de procedencia ilícita. [Bulletin 338/07. The PGR has tion; 1996. a solid and overwhelming legal case against Zhenli Ye Gon for 9. Cunningham J. K., Liu L.-M., Callaghan R. Impact of US and different types of crimes against health; organized crime, posses- Canadian precursor regulation on methamphetamine sion of firearms and cartridges reserved for the Armed Forces, purity in the United States. Addiction 2009; 104: 441–53. and illegal procedural operations. México, DF: PGR; 2007. 10. Cunningham J. K., Liu L.-M. Impact of methamphetamine 21. Comisión Federal para la Protección contra Riesgos Sani- precursor chemical legislation, a suppression policy, on the tarios (COFEPRIS). La COFEPRIS presentó ante la PGR 41 demand for drug treatment. Soc Sci Med 2008; 66: 1463– denuncias penales por 161 permisos sanitarios de importación 73. apócrifos [COFEPRIS filed before the PGR 41 criminal com-

plaints due to 161 fraudulent importation health permits]. 38. Falcato L., Stohler R., Dürsteler-Mac Farland K. M., Eichen- México, DF: COFEPRIS; 2007. berger A., Eich D., Rössler W. Closure of an open drug 22. Consejo de Salubridad General. Acuerdo por el que se scene—a case register-based analysis of the impact on the establecen medidas de protección en material de salud demand for methadone maintenance treatment. Addiction humana para prevenir el uso y consumo de pseudoefedrina 2001; 96: 623–8. y efedrina. [General Health Council. Agreement establish- 39. Weatherburn D., Jones C., Freeman K., Makkai T. The ing protective measures for human health to prevent the Australian Heroin Drought and Its Implications for Drug Policy. use and consumption of pseudoephedrine and ephedrine]. [Crime and Justice Bulletin Number 59]. Sydney: NSW Diario Oficial 13 June 2008; p. 31–3. Bureau of Crime Statistics and Research; 2001. 23. Cunningham J. K., Liu L.-M. Impacts of federal precursor 40. Smithson M., McFadden M., Mwesigye S.-E., Casey T. chemical regulations on methamphetamine arrests. Addic- The impact of illicit supply reduction on health and social tion 2005; 100: 479–88. outcomes: the heroin shortage in the Australian Capital 24. Cunningham J. K., Liu L.-M., Muramoto M. Methamphet- Territory. Addiction 2004; 99: 340–8. amine suppression and route of administration: precursor 41. American Psychiatric Association. Treatment of patients regulation impacts on snorting, smoking, swallowing and with substance use disorders (2nd edn). Am J Psychiatry injecting. Addiction 2008; 103: 1174–86. 2007; 164: S5–123. 25. Cunningham J. K., Liu L.-M. Impacts of federal ephedrine 42. Rothman R. B., Partilla J. S., Dersch C. M., Carroll F. I., Rice and pseudoephedrine regulations on methamphetamine- K. C., Baumann M. H. Methamphetamine dependence: related hospital admissions. Addiction 2003; 98: 1229–37. medication development efforts based on the dual deficit 26. Callaghan R. C., Cunningham J. K., Victor J. C., Liu L.-M. model of stimulant addiction. Ann NY Acad Sci 2000; 914: Impact of Canadian federal methamphetamine precursor 71–81. and essential chemical regulations on methamphetamine- 43. Threlkeld M., Parran T. V., Adelman C. A., Grey S. F., Yu J. related acute-care hospital admissions. Drug Alcohol Depend 2006). Tramadol versus buprenorphine for the manage- 2009; 105: 185–93. ment of acute heroin withdrawal: a retrospective matched 27. Medina-Mora M. E., Cravioto P., Ortíz A., Kuri P.,Villatoro J. cohort controlled study. Am J Addict 2006; 15: 186–91. Mexico: systems for epidemiological diagnosis of drug 44. National Institute on Drug Abuse. NIDA InfoFacts: Heroin. abuse. Bull Narc 2003; LV: 105–19. Bethesda, MD: National Institutes of Health; 2006. 28. Procuraduría General de la República (PGR). National Drug 45. Wickler A. Opioid Dependence: Mechanisms and Treatment. Control Program Actions and Results, First Semester 2005. New York: Plenum Press; 1980. México, DF: PGR; 2006. 46. McGregor C., Srisurapanont M., Jittiwutikarn J., Laobhri- 29. National Drug Intelligence Center. National Drug Threat patr S., Wongtan T., White J. M. The nature, time–course Assessment 2008. Johnstown, PA: US Department of Justice; and severity of methamphetamine withdrawal. Addiction 2007. 2005; 100: 1320–9. 30. US Department of Commerce. Statistical Abstract of the 47. Kuri P., López-Gatell H., Bojorquez I., Cortes M. Update on United States, 2010. Washington, DC: US Census Bureau; the epidemiologic surveillance system of addictions 2010. (SISVEA) in Mexico: 2007. Epidemiologic Trends in Drug 31. Lee R. W.,Wilson J. T. Trace Elements and Organic Compounds Abuse: Proceedings of the Community Epidemiologic Work Associated with Riverbed Sediments in the Rio Grande/Rio Group June 2008. Rockville, MD: National Institute on Drug Bravo Basin, Mexico and Texas. Austin, TX: U.S. Geological Abuse; 2009, p.321–8. Survey; 1997. 48. Box G. E. P., Tiao G. C. Intervention analysis with applica- 32. Woodward D., Durall R. United States–Mexico Border Area, tions to economic and environmental problems. JAmStat As Delineated by a Shared-Water Resources Perspective. Albu- Assoc 1975; 70: 70–9. querque, NM: US Geological Survey; 1996. 49. McCleary R., Hay R. A. Applied Time Series Analysis for the 33. Drug Enforcement Agency. DEA Briefs and Background, Social Sciences. Beverly Hills: Sage Publications; 1980. Drugs and Drug Abuse, State Factsheets, Texas. Washington, 50. Liu L.-M., Chen C. Recent developments of time series analy- DC: US Department of Justice; 2008. sis in intervention and environmental impact studies. 34. Global Assessment Programme. Guidance for the Measure- J Environ Sci Health 1991; A26: 1217–52. ment of Drug Treatment Demand. New York: United Nations; 51. Chen C., Liu L.-M. Joint estimation of model parameters and 2006. outlier effects in time series. J Am Stat Assoc 1993; 88: 284– 35. Dirección de Investigación Operativa Epidemiológica. 97. Sistema de Vigilancia Epidemiológica de las Adicciones 52. Liu L.-M., Hudak G. B., Box G. E. P., Muller M. E., Tiao G. C. (SISVEA), Informe 2006 [Epidemiologic Surveillance System Forecasting and Time Series Analysis Using the SCA Statistical of Addictions (SISVEA) 2006 Report]. México, DF: Subsecre- System. Chicago: Scientific Computing Associates Corpora- taría de Prevención y Promoción de la Salud, Dirección tion; 1992. General de Epidemiología; 2006. 53. Government of Canada. Controlled drugs and substances 36. Orford J., Kerr C., Copello A., Hodgson R., Alwyn T., Black R. act: precursor control regulations. Canada Gaz 2002; 136: et al. Why people enter treatment for alcohol problems: find- 2139–200. ings from UK Alcohol Treatment Trial pre-treatment inter- 54. Criminal Intelligence Directorate. Federal and International views. J Subst Use 2006; 11: 161–76. Operations—Coming into Force of the Precursor Control 37. Kleiman M. A. R. Crackdowns: the effects of intensive Regulations. Beginning January 9, 2003. Ottawa, ON: Royal enforcement on retail heroin dealing. In: Chaiken M. R., Canadian Mounted Police; 2003. editor. Street-Level Drug Enforcement: Examining the Issues, 55. H. B. No. 164. House of Representatives. Seventy- Washington, DC: National Institute of Justice; 1988, p. Ninth Legislature, 2005. An Act Relating to the Civil and 3–34. Criminal Consequences of Engaging in Conduct Related to the

Manufacture of Methamphetamine and to the Distribution and representations among the legal community of the Federal Retail Sales of Certain Chemical Substances. Austin: State of District. Crim Law Forum 2005; 16: 257–77. Texas; 2005. 63. Cejudo G. M. Explaining change in the Mexican public 56. H. B. No. 164. House of Representatives. Summary of sector: the limits of New Public Management. Int Rev Admin Enactments—Seventy-Ninth Legislature, 2005. Austin: State Sci 2008; 74: 111–27. of Texas; 2005. 64. Constantine T. A. Statement by Thomas A. Constantine, 57. Cook T. D., Campbell D. T. Quasi-Experimentation: Design and Administrator,Drug Enforcement Administration, United States Analysis Issues in Field Settings. Chicago: Rand-McNally; Department of Justice. Before the House Subcommittee on the 1979. Western Hemisphere, House International Relations Committee. 58. Shadish W. R., Cook T. D., Campbell D. T. Experimental and Regarding: Drug Control in the Western Hemisphere. June 6, Quasi-Experimental Designs for Generalized Causal Inference. 1996. Washington, DC: Drug Enforcement Administration; Boston: Houghton Mifﬂin; 2002. 1996. 59. McKetin R. The holy grail of supply control: have we found 65. United States Government Accountability Ofﬁce (GAO). the evidence? Addiction 2009; 104: 454–6. Drug Control: U. S. Assistance has Helped Mexican Counterna- 60. Sumnall H. R., Tyler E., Wagstaff G. F., Cole J. C. A behav- rcotics Efforts, but Tons of Illicit Drugs Continue to Flow into the ioural economic analysis of alcohol, amphetamine, cocaine United States. GAO-07-1018, August 2007. Washington, and ecstasy purchases by polysubstance misusers. Drug DC: GAO; 2007. Alcohol Depend 2004; 76: 93–9. 66. International Narcotics Control Board. Report of the Inter- 61. Bailey J., Paras P. Perceptions and attitudes about corrup- national Narcotics Control Board for 2008. New York: United tion and democracy in Mexico. Mexican Studies/Estudios Nations; 2009, p. 56–63. Mexicanos 2006; 22: 57–81. 67. McKetin R. Methamphetamine precursor regulation: are we 62. Garcia Gonzales J. C., Gonzalez Robles R. O., Mir Araujo A., controlling or diverting the drug problem? Addiction 2008; Garcia Delsoto A. Public sector corruption in Mexico: social 103: 521–3.

© 2010 The Authors, Addiction © 2010 Society for the Study of Addiction Addiction, 105, 1973–1983 This document is a scanned copy of a printed document. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material.