Addictive Behaviors 39 (2014) 1311–1317

Contents lists available at ScienceDirect

Addictive Behaviors

Age Cohort Differences in the Nonmedical Use of Prescription Zolpidem: Findings from a Nationally Representative Sample Ty S. Schepis ⁎ Department of Psychology, Texas State University, 601 University Drive, San Marcos, Texas 78666

H I G H L I G H T S • • • • •

Age cohort differences in nonmedical zolpidem use correlates were examined. Nationally representative data from nearly 175,000 participants were used. Substance use correlates differed most by age, with highest odds in adolescents. Mental health correlates operated more uniformly across cohorts. Findings may indicate age-based differences in nonmedical zolpidem use motives.

a r t i c l e

i n f o

Available online 24 April 2014 Keywords: nonmedical use zolpidem misuse age cohort differences

a b s t r a c t Background: Recent warnings from the FDA have highlighted the potential risks associated with zolpidem use. These risks may be especially acute in nonmedical users of zolpidem, but little work has examined the characteristics of such nonmedical users. This study aims to investigate the correlates of nonmedical use of zolpidem (NUPZ) across the lifespan and potential age cohort-based differences in NUPZ correlates. Methods: Data from the 2009–2011 versions of the National Survey on Drug Use and Health were used (n = 174,667). Analyses used weighted design-based logistic regressions to examine a set of substance use and mental health correlates within five separate age cohorts and differences in correlate magnitude between these cohorts. Results: Most examined substance use and mental health variables were significant correlates of NUPZ, though odds ratio (OR) magnitude tended to drop with increasing age. Age-based differences were most apparent for substance use correlates of both lifetime and past year NUPZ, with significantly higher ORs in adolescent nonmedical users. Mental health variables operated more consistently across age, with OR magnitudes that were generally in the same range, regardless of age cohort. Conclusions: Age-based differences in NUPZ correlates suggest motives may change for NUPZ through the lifespan, though this cannot be established with the cross-sectional data used in this work. Clinicians screening for NUPZ should emphasize such screening in high-risk individuals with substance use and/or mental health problems. © 2014 Elsevier Ltd. All rights reserved.

1. Introduction Zolpidem, marketed most commonly as Ambien® or Ambien CR®, is a controlled non-benzodiazepine indicated for insomnia treatment. Oral zolpidem has a long history of use, with significant evidence of its effectiveness in short-term treatment (Dang, Garg, & Rataboli, 2011). The benefits of zolpidem extend to improving work performance over a 6-month period in those with chronic insomnia (Erman, Guiraud, Joish, & Lerner, 2008), and zolpidem may help those with insomnia related to psychopathology (Fava et al., 2011).

⁎ Tel.: +1 512 245 6805; fax: +1 512 245 3153. E-mail address: [email protected].

http://dx.doi.org/10.1016/j.addbeh.2014.04.018 0306-4603/© 2014 Elsevier Ltd. All rights reserved.

Despite these benefits, concern about the effects of zolpidem prompted the US Food and Drug Administration to recommend lower dosing to prevent next-day impairment, particularly in women (Kuehn, 2013). Recent evidence also associates zolpidem with increases in suicidality (Brower et al., 2011), emergency department utilization (Mitka, 2013), motor vehicle accidents (Yang, Lai, Lee, Wang, & Chen, 2011) and various other injuries, especially in elderly individuals (Chung, Lin, Wang, Lin, & Kang, 2013). The increased risk of zolpidemrelated injury may result from similar effects of both zolpidem and traditional benzodiazepines on memory (Pompeia, Lucchesi, Bueno, Manzano, & Tufik, 2004), balance (Frey, Ortega, Wiseman, Farley, & Wright, 2011) and attention (Troy et al., 2000). Zolpidem is on Schedule IV of the US Controlled Substances Act, denoting therapeutic benefits with some degree of abuse potential. The

1312

T.S. Schepis / Addictive Behaviors 39 (2014) 1311–1317

World Health Organization echoed this assessment, recommending zolpidem for international control (WHO, 2001). Research evidence largely supports the controlled status of zolpidem, with evidence of modest reinforcing effects in drug-naïve individuals (Licata, Mashhoon, Maclean, & Lukas, 2011), and concern about abuse in those with substance use disorder histories (Hajak, Muller, Wittchen, Pittrow, & Kirch, 2003). Furthermore, zolpidem is used nonmedically (Ford & McCutcheon, 2012), either for the mild reinforcing effects, or as selftreatment for undertreated or undiagnosed insomnia. That said, only one study has examined the correlates of nonmedical use of prescription zolpidem (NUPZ). Ford and McCutcheon (2012) used data from the 2009 National Survey on Drug Use and Health (NSDUH) to examine the correlates of lifetime NUPZ in adolescents. Their findings were consistent with other work (e.g., Boyd, McCabe, Cranford, & Young, 2007) examining correlates of adolescent nonmedical use of other medications: lifetime other substance use, major depression and permissive parental and peer attitudes towards substance use increased NUPZ odds. After controlling for sociodemographic characteristics, they found a more limited set of correlates, including other nonmedical use and lifetime marijuana use (Ford & McCutcheon, 2012). More thorough examination of NUPZ correlates is warranted, particularly because of the lack of research on adult correlates and ongoing concerns about adverse outcomes associated with zolpidem use. Nonmedical users would also experience these effects, possibly at higher rates due to improper medication use. Furthermore, nonmedical use rates decrease with age (SAMHSA, 2012), indicating the potential for different processes related to nonmedical use throughout the lifespan. Nonetheless, no work has directly examined age-based differences in nonmedical use generally or NUPZ specifically. Furthermore, more thorough examination of adverse characteristics associated with NUPZ can inform prevention efforts to limit it across the lifespan. Examination of a diverse set of potential substance use and mental health correlates of NUPZ across different age cohorts will allow this work to evaluate the fit of Problem Behavior Theory on NUPZ at different timepoints in the lifespan. Problem Behavior Theory posits that deviant behavior clusters, particularly in adolescents, and that such behavior is influenced by both risk (e.g., chaotic home environment) and protective (e.g., church attendance) factors (Jessor, 1987; Jessor & Jessor, 1977). While this work will not evaluate protective factors, it will evaluate potential risk factors to allow for partial examination of NUPZ through the lens of Problem Behavior Theory. The primary aims of this work are to examine the sociodemographic, substance use and mental health correlates of lifetime and past year NUPZ across five cohorts, with separate examination of correlate magnitudes by age cohort. We hypothesized that substance use and mental health problems would increase NUPZ odds, with the highest odds linked to nonmedical and illicit substance use. We also wanted to evaluate the utility of Problem Behavior Theory in understanding NUPZ across the lifespan. We posited that NUPZ and other substance use would clearly cluster in adolescents, with less marked clustering as individuals aged, given that some deviant behaviors in adolescents (e.g., alcohol use) are not deviant in adults and that deviant behavior likely decreases through the young adult period (Jessor, Donovan, & Costa, 1991). 2. Methods 2.1. NSDUH Design The NSDUH is an annual survey conducted by the US Substance Abuse and Mental Health Services Administration (SAMHSA) to investigate substance use and associated behaviors in a sample that is representative of non-institutionalized population of the US. The NSDUH is designed to oversample adolescents, young adults, African–Americans and Hispanics. It used an independent, multistage area probability sample for all states and the District of Columbia, with yearly population estimates from the U.S. Census Bureau informing population-based weights (SAMHSA,

2009). More information on survey sampling is available elsewhere (Research Triangle Institute, 2012). The NSUDH combined both computer-assisted interviewing and audio computer-assisted selfinterviewing (ACASI) methods. During the ACASI portion of the survey, the participant wore headphones to hear questions and the field interviewer remained out of view of the computer screen; these procedures were employed to maximize honest responding. All substance use (including NUPZ-related) and psychopathology measures were asked in ACASI format. The 2009–2011 NSDUH versions included automatic skip-outs and questions serving as consistency checks based on previous answers to increase full responding and data consistency. 2.2. Participants For the 2009–2011 NSDUH versions, 174,667 respondents were included in the public use files. Of those, 33.0% were young adults (aged 18–25; n = 57,678) and 32.7% were adolescents (aged 12–17; n = 57,142). Females composed 51.6% of the sample (n = 90,161), with Caucasian (n = 108,016; 61.8%), Hispanic/Latino (n = 28,541; 16.3%) and African–American individuals (n = 22,545; 12.9%) comprising the three largest ethnic groups. 2.3. Measures 2.3.1. Primary Outcomes and Control Variables The primary outcome measures are lifetime and past year NUPZ. Current age is used in two ways: first, it separates participants for analyses of correlates by age cohort; and second, it is used in an interaction term with the examined correlates to evaluate potential cohort differences in associations between NUPZ and correlates. Control variables were sex, race/ethnicity, family income, educational attainment and metro status in area of residence. Current age was a five-level variable, with age group choices restricted by the available variables in the NSDUH public use file. Groups came from the CATAG3 variable and were ages 12–17, 18–25, 26–34, 35–49, and 50 and older. Lifetime NUPZ is defined as zolpidem use when “the drug was not prescribed for you, or you took the drug only for the experience or feeling it caused.” Past year NUPZ is queried in participants endorsing lifetime NUPZ by asking whether the respondent engaged in NUPZ within the past 12 months. 2.3.2. Correlates Correlates were selected from previous examinations of NUPZ, with selection of articles ensuring coverage of the lifespan. Correlates were chosen to be consistent across NUPZ timeframes (allowing comparison between timeframes), with greater public health impact (i.e., daily smoking over any 30-day smoking), while ensuring sufficient base rates engaged in the behavior. Substance use correlates were: daily smoking, 30-day binge alcohol use, 30-day heavy drinking, 30-day marijuana use, lifetime cocaine use, lifetime opioid nonmedical use, lifetime tranquilizer nonmedical use, lifetime stimulant nonmedical use and past year substance use treatment. For all substance use variables, assessment began with queries about any lifetime use, followed up by questions about recency of use (e.g., past 30-day). Smoking on all of the past 30 days was defined as daily smoking. 30-day binge alcohol use was defined as one occasion (“at the same time or within a couple of hours of each other”) of consuming 5 or more alcoholic drinks. 30-day heavy drinking was defined as five episodes of binge alcohol use in the past 30 days. Finally, nonmedical use was assessed via a similar query as was used to assess NUPZ (above). Mental health correlates were: past year anxiety diagnosis, past year major depression, past year mental health treatment, past year serious psychological distress (SPD; adult only), past year suicidal ideation (adult only). Anxiety diagnosis was assessed via a single item asking if

T.S. Schepis / Addictive Behaviors 39 (2014) 1311–1317

“a doctor or other medical professional…that you had [anxiety] in the past 12 months.” Major depression was assessed via questions from the National Comorbidity Study-Replication and National Comorbidity Study-Adolescent, based on the DSM-IV (American Psychiatric Association, 2001). This assessment has good reliability and validity (Zanarini & Frankenburg, 2001). The SPD assessment comes from the K6 assessment of nonspecific psychological distress (Kessler et al., 2003) for the worst month in the past year in adults. An SPD dichotomous variable was created, with ≥ 13 (of 24) as positive for SPD. Finally, suicidal ideation is queried by asking adults if in the past 12 months “did you seriously think about trying to kill yourself?” 2.4. Analyses Analyses were performed in SUDAAN 10.1 (Research Triangle Park, NC). Data were sorted to account for the 50% overlap between successive years in variance and standard error estimates, and adjusted person-level weights (weight/3) were applied to create unbiased population-based estimates. For all hypotheses, design-based logistic regression was utilized. First, analyses were run for the correlate under examination while controlling for (other) sociodemographic variables separately in each age cohort. One set of analyses was performed for lifetime and one performed for past year NUPZ. Then, each correlate was included in an interaction term with the age cohort variable for both lifetime and past year NUPZ to test whether the relationship between the correlate and NUPZ varied by current age. Because of the large sample, we set a conservative a priori significance value of .001 to maximize clinical significance. 3. Results 3.1. Sociodemographics and Lifetime NUPZ Lifetime NUPZ rates varied from 1.06% in adolescents to 4.35% in the 26–34 year-old cohort. The most consistent NUPZ correlate across cohorts was Caucasian race/ethnicity. With the exception of the 50 and older cohort, African–Americans had lifetime NUPZ odds that were over 3 times lower than those in Caucasians. While not as consistent, other participants of non-Caucasian backgrounds were also protected from

1313

lifetime NUPZ. These results, including odds ratios (ORs) and 95% confidence intervals (95% CIs), are in Table 1. 3.2. Sociodemographics and Past Year NUPZ Past year NUPZ prevalence ranged from 0.43% in the 50 and older cohort to 1.28% in the 26–34 year-old cohort. African–Americans had past year NUPZ odds that were over 4 times lower than those of Caucasians in both adolescents and young adults; this seemed to drive the significant interaction between age cohort and racial/ethnic status for past year NUPZ. Across cohorts, no other correlates were significantly associated with past year NUPZ. These results, including ORs and 95% CIs, are in Table 2. 3.3. Substance Use and Mental Health Correlates of Lifetime NUPZ In adolescents, young adults and those in the 26–34 year-old cohort, every examined correlate was significantly associated with lifetime NUPZ. Mental health variables generally had lower ORs than all substance use variables, though only adolescents had higher alcohol and tobacco-related ORs than mental health-related ORs. Other forms of nonmedical use consistently had the highest ORs for lifetime NUPZ. This pattern roughly held for the 35–49 year-old cohort, though past 30-day binge alcohol use was the only significant licit substance use correlate of lifetime NUPZ. In the 50 and older cohort, fewer substance use correlates were significantly associated, but other nonmedical use remained a robust correlate. All mental health variables, except for past year anxiety disorder diagnosis, were significant correlates in this group. All substance use variables, except past year substance use treatment, evidenced a significant interaction with cohort for lifetime NUPZ. The primary driver of each interaction seemed to be the much larger OR in adolescents. Indeed, post hoc pairwise comparisons indicated the ORs for all significant substance use variables were higher in adolescents than any other cohort. The only mental health variable with a significant age-based interaction was past year anxiety disorder diagnosis. Again, post hoc pairwise comparisons indicated that adolescents had highest ORs, and young adults had a higher OR than those aged 50 and older.

Table 1 Sociodemographic Correlates for Lifetime Nonmedical Use of Prescription Zolpidem.

Subsample % and 95% CI Sex Female Male Race/Ethnicity Caucasian African–American Hispanic Other Metro Status Large Metro Small Metro Nonmetro Family Income b $20,000 $20,000–49,999 $50,000–74,999 N $75,000 Education No High School High School Some College College Graduate

Adolescents (12–17 years) N = 57,756

Young Adults (18–25 years) N = 58,646

26–34 years N = 17,841

35–49 years N = 24,244

50 and older N = 19,373

1.06% (0.92–1.22%)

3.48% (3.27–3.72%)

4.35% (3.72–5.07%)

2.67% (2.39–2.97%)

1.54% (1.32–1.81%)

1.00 (reference) 0.83 (0.60–1.15)

1.00 (ref.) 1.08 (0.91–1.27)

1.00 (ref.) 0.95 (0.73–1.25)

1.00 (ref.) 1.04 (0.80–1.36)

1.00 (ref.) 0.62 (0.42–0.93)

1.00 (reference) 0.36 (0.21–0.62) 0.67 (0.41–1.08) 0.64 (0.39–1.04)

1.00 (ref.) 0.21 (0.14–0.32) 0.39 (0.30–0.51) 0.40 (0.27–0.59)

1.00 (ref.) 0.27 (0.15–0.49) 0.49 (0.29–0.83) 0.20 (0.10–0.37)

1.00 (ref.) 0.38 (0.22–0.66) 0.31 (0.18–0.55) 0.65 (0.43–0.99)

1.00 (ref.) 0.81 (0.42–1.56) 0.32 (0.14–0.75) ***No Use***

1.00 (reference) 1.37 (1.03–1.83) 1.28 (0.81–2.01)

1.00 (ref.) 0.94 (0.80–1.10) 0.68 (0.53–0.87)

1.00 (ref.) 0.85 (0.65–1.10) 0.68 (0.42–1.11)

1.00 (ref.) 1.07 (0.80–1.43) 1.04 (0.69–1.55)

1.00 (ref.) 0.97 (0.62–1.53) 0.95 (0.54–1.68)

1.00 (reference) 0.69 (0.44–1.08) 0.83 (0.46–1.48) 0.68 (0.44–1.04)

1.00 (ref.) 0.94 (0.76–1.16) 1.08 (0.82–1.43) 0.98 (0.76–1.26)

1.00 (ref.) 1.03 (0.67–1.59) 1.16 (0.70–1.91) 1.01 (0.62–1.63)

1.00 (ref.) 0.70 (0.46–1.08) 0.61 (0.34–1.10) 0.70 (0.46–1.07)

1.00 (ref.) 0.51 (0.29–0.89) 0.77 (0.38–1.57) 0.81 (0.39–1.65)

not applicable not applicable not applicable not applicable

1.00 (ref.) 1.24 (1.00–1.52) 1.34 (1.06–1.69) 1.41 (1.06–1.87)

1.00 (ref.) 1.17 (0.61–2.26) 1.43 (0.81–2.53) 1.46 (0.78–2.73)

1.00 (ref.) 1.35 (0.71–2.57) 1.51 (0.85–2.70) 1.84 (0.96–3.51)

1.00 (ref.) 1.96 (1.03–3.73) 3.86 (1.79–8.35) 4.57 (1.95–10.69)

Interaction

p = .020

p b .0001

p = .121

p = .378

p = .222

Note: Bolded values indicate pairwise comparisons with the reference group or interactions that are significant at the p b .001 level. 95% CI = 95% Confidence Interval, NUPZ = Nonmedical Use of Prescription Zolpidem.

1314

T.S. Schepis / Addictive Behaviors 39 (2014) 1311–1317

Table 2 Sociodemographic Correlates for Past Year Nonmedical Use of Prescription Zolpidem.

Subsample % and 95% CI Sex Female Male Race/Ethnicity Caucasian African–American Hispanic Other Metro Status Large Metro Small Metro Nonmetro Family Income b $20,000 $20,000–49,999 $50,000–74,999 N $75,000 Education No High School High School Some College College Graduate

Adolescents (12–17 years) N = 57,142

Young Adults (18–25 years) N = 57,678

26–34 years N = 17,156

35–49 years N = 23,614

50 and older N = 19,077

0.55% (0.46–0.66%)

1.16% (1.02–1.32%)

1.28% (1.01–1.62%)

0.91% (0.74–1.12%)

0.43% (0.29–0.64%)

1.00 (reference) 1.01 (0.68–1.50)

1.00 (ref.) 0.95 (0.71–1.26)

1.00 (ref.) 0.99 (0.69–1.42)

1.00 (ref.) 0.85 (0.53–1.36)

1.00 (ref.) 0.84 (0.36–1.92)

1.00 (reference) 0.20 (0.09–0.43) 0.56 (0.29–1.10) 0.46 (0.20–1.04)

1.00 (ref.) 0.23 (0.12–0.46) 0.48 (0.28–0.80) 0.39 (0.22–0.69)

1.00 (ref.) 0.33 (0.15–0.73) 0.47 (0.21–1.05) 0.25 (0.07–0.84)

1.00 (ref.) 0.44 (0.20–0.95) 0.24 (0.10–0.58) 0.63 (0.27–1.47)

1.00 (ref.) 0.95 (0.24–3.80) 0.65 (0.14–2.91) ***No use***

1.00 (reference) 1.19 (0.76–1.85) 1.05 (0.55–2.01)

1.00 (ref.) 0.81 (0.62–1.06) 0.60 (0.41–0.87)

1.00 (ref.) 0.69 (0.42–1.13) 0.35 (0.16–0.75)

1.00 (ref.) 0.91 (0.55–1.50) 1.07 (0.54–2.13)

1.00 (ref.) 1.17 (0.54–2.50) 2.16 (0.77–6.07)

1.00 (reference) 0.76 (0.44–1.34) 0.84 (0.37–1.90) 0.65 (0.36–1.15)

1.00 (ref.) 0.68 (0.48–0.95) 0.74 (0.44–1.25) 0.98 (0.66–1.45)

1.00 (ref.) 0.97 (0.48–1.94) 0.81 (0.41–1.58) 1.00 (0.55–1.82)

1.00 (ref.) 1.09 (0.54–2.17) 1.08 (0.52–2.23) 0.99 (0.51–1.92)

1.00 (ref.) 0.39 (0.17–0.88) 0.41 (0.11–1.55) 0.56 (0.16–1.97)

not applicable not applicable not applicable not applicable

1.00 (ref.) 1.20 (0.78–1.87) 1.33 (0.83–2.12) 1.12 (0.70–1.78)

1.00 (ref.) 0.74 (0.32–1.72) 0.71 (0.36–1.40) 0.96 (0.47–1.98)

1.00 (ref.) 0.66 (0.26–1.66) 0.65 (0.25–1.67) 0.84 (0.34–2.08)

1.00 (ref.) 1.45 (0.36–5.85) 2.32 (0.62–8.70) 5.38 (1.34–21.58)

Interaction

p = .975

p b .0001

p = .028

p = .448

p = .360

Note: Bolded values indicate pairwise comparisons with the reference group or interactions that are significant at the p b .001 level, 95% CI = 95% Confidence Interval, NUPZ = Nonmedical Use of Prescription Zolpidem.

All lifetime NUPZ correlate results, including ORs, 95% CIs, interaction p-values and post hoc comparisons are in Table 3. 3.4. Substance Use and Mental Health Correlates of Past Year NUPZ The examined correlates were less consistently associated with past year NUPZ than with lifetime NUPZ. As with lifetime NUPZ, though, all examined substance use and mental health correlates in adolescents and young adults were significantly associated with past year NUPZ. In adolescents, the highest odds were from other nonmedical use, with mental health variables having the lowest odds; in young adults, odds from nonmedical use were also highest, with the ORs for other substance use and mental health variables largely co-equal. The 26–34 year-old and 35–49 year-old cohorts had a less consistent pattern of significant

correlates, with illicit substance use, particularly nonmedical use, consistently associated with past year NUPZ. In the 50 and older cohort, only the three other nonmedical use variables were associated with past year NUPZ. With the exception of substance use treatment, all substance use variables evidenced a significant age-based interaction. Again, significantly higher ORs in the adolescent cohort drove the results, with two relative exceptions. The exceptions were past 30-day marijuana use, where the adolescent OR was only higher than those of young adults and 26–34 year-olds, and lifetime stimulant nonmedical use, where the adolescent OR was higher than every cohort’s OR except for the age 50 and older group. All past year NUPZ correlate results, including ORs, 95% CIs, interaction p-values and post hoc comparisons are in Table 4.

Table 3 Substance Use and Mental Health Correlates for Lifetime Nonmedical Use of Prescription Zolpidem.

Daily Smoking 30d Binge Alc. Heavy Drinking 30d Marijuana LT Cocaine LT Opioid LT Tranquilizer LT Stimulant PY SUD Tx PY Anxiety Dx PY MDD PY MH Tx PY SPD PY Suicidality

Adolescents (12–17 years)

Young Adults (18–25 years)

26–34 years

35–49 years

50 and older

Interaction

10.92 (7.11–16.77) 8.24 (6.24–10.87) 12.30 (8.10–18.68) 10.98 (8.40–14.37) 22.43 (16.17–31.12) 23.95 (18.22–31.48) 29.93 (22.08–40.58) 39.55 (28.00–55.86) 11.46 (7.58–17.33) 6.54 (4.56–9.37) 2.78 (1.80–4.27) 2.87 (2.23–3.70) not available not available

2.42 (2.01–2.92) 2.46 (2.10–2.88) 2.63 (2.10–3.30) 4.10 (3.53–4.75) 6.91 (5.66–8.45) 11.59 (9.55–14.06) 14.20 (11.98–16.84) 9.00 (7.36–11.01) 4.85 (3.50–6.72) 3.18 (2.48–4.07) 2.60 (2.10–3.23) 2.96 (2.46–3.56) 2.80 (2.36–3.32) 2.55 (1.96–3.33)

1.74 (1.30–2.34) 1.65 (1.25–2.17) 1.83 (1.28–2.62) 4.23 (2.88–6.20) 5.33 (3.91–7.28) 8.15 (6.14–10.82) 12.45 (9.45–16.41) 5.74 (4.46–7.37) 5.77 (2.97–11.23) 2.33 (1.48–3.67) 2.61 (1.80–3.79) 2.51 (1.71–3.66) 3.46 (2.53–4.74) 3.35 (1.98–5.69)

1.51 (1.05–2.17) 2.11 (1.56–2.86) 2.06 (1.32–3.23) 3.68 (2.52–5.38) 3.93 (3.02–5.11) 8.59 (6.17–11.96) 10.04 (7.63–13.21) 5.67 (4.14–7.78) 5.39 (3.02–9.63) 2.62 (1.83–3.75) 2.66 (1.90–3.71) 3.14 (2.49–3.95) 3.17 (2.26–4.43) 3.08 (1.97–4.80)

1.19 (0.67–2.11) 2.32 (1.45–3.71) 1.42 (0.56–3.57) 3.11 (1.48–6.56) 3.52 (2.31–5.37) 8.28 (5.34–12.83) 10.68 (6.82–16.73) 4.56 (3.13–6.65) 5.11 (1.83–14.22) 2.39 (1.41–4.06) 3.46 (2.10–5.72) 2.13 (1.36–3.32) 3.16 (1.95–5.13) 4.18 (2.06–8.47)

p b .0001a p b .0001a p b .0001a p b .0001a p b .0001a, d p b .0001a p b .0001a p b .0001a p = .024 p = .0002a, e p = .777 p = .709 p = .671 p = .545

Notes: Bolded values are significant at a p b .001 level, 30d = 30-day, Binge Alc. = Binge Alcohol Use (5 or more alcoholic drinks on one occasion), LT = Lifetime, PY = Past Year, Tx = Treatment, Dx = Diagnosis, MDD = Major Depressive Disorder, MH = Mental Health, SPD = Serious Psychological Distress. a OR for adolescents N all. b OR for adolescents N all except 50 and older. c OR for adolescents N young adults and 26–34 years. d OR for young adults N 35–49 years and 50 and older. e OR for young adults N 50 and older.

T.S. Schepis / Addictive Behaviors 39 (2014) 1311–1317

4. Discussion These results concur with other investigations of nonmedical use correlates (e.g., Schepis & Krishnan-Sarin, 2008), with mental health problems and substance use generally serving as significant correlates of NUPZ. These correlates were less likely to be associated with past year than lifetime NUPZ or with advancing age, but nonmedical use of other medications consistently remained a significant correlate. Perhaps more noteworthy, this investigation found a pattern of age-based differences in correlate OR magnitudes for both past year and lifetime NUPZ that were generally consistent with Problem Behavior Theory. While mental health correlates were generally associated with a similar level of increased NUPZ odds across the lifespan, substance use was most prominently associated with NUPZ in adolescents. Such clustering of deviance, including NUPZ, in adolescence and decreases in deviance with increased age are consistent with Problem Behavior Theory. With advancing age, the ORs for the associations between NUPZ and the substance use correlates decreased, though this was not universal, and the decreases were small from young adults to other adult cohorts. The decrease in substance use correlate ORs from adolescents to young adults, though, was often quite large. The differences in NUPZ correlates between adolescents and the adult cohorts may indicate divergence in the processes that promote NUPZ. While causality cannot be inferred from cross-sectional data, the age-based differences seen here are suggestive. One possibility is that adolescents have different motives for NUPZ than older cohorts. Work in adolescents (McCabe, Boyd, Cranford, & Teter, 2009) indicates that those with recreational motives for opioid nonmedical use (e.g., to get high) had a greater likelihood of other substance use than those with only self-treatment motives (e.g., to relieve pain). Young adults with recreational motives for nonmedical prescription use also had higher odds of binge alcohol use, other drug use and problematic illicit drug use than those with self-treatment motives only (McCabe, Boyd, & Teter, 2009). Furthermore, motives for nonmedical use appear to change over time (McCabe, West, & Boyd, 2013), which could explain how the pattern of correlate magnitudes for NUPZ changes through development. Adolescents may be more likely than other developmental groups to initiate and maintain NUPZ for recreational motives; at older ages, the importance of recreational motives may decrease, while the importance of self-treatment motives remains relatively stable. Again, though, notable caution should be exercised in interpreting these results in a causal fashion. Also, it is not clear that the mental health correlates examined here necessarily associate with self-treatment motives, though it is likely

1315

that some proportion of those with mental health problems who engaged in NUPZ do so to treat the insomnia that often accompanies such conditions. Thus, in older adults, NUPZ may be consistent with the self-medication theory of substance use (e.g., Khantzian, 1985). In addition, NUPZ may be a more general marker of deviance than in older cohorts. Adolescents engaged in NUPZ may have a stronger general propensity towards risk-taking and substance use, with NUPZ simply an expression of a higher-order deviance factor. This interpretation is consistent with Problem Behavior Theory (e.g., Jessor, 1987), insofar as NUPZ tends to cluster with other deviant behaviors most clearly in adolescents. Other evidence indicates that deviance decreases through young adulthood, which would suggest that deviant behaviors would less clearly cluster in older cohorts. We found such a pattern here. Adolescents appear to have neurobiological differences, persisting into young adulthood, that predispose them to poorer inhibitory control and greater risk engagement. Furthermore, adolescents may have poorer stress-coping skills and more recent experience with major life trauma, like abuse, that could alter NUPZ propensity. This deviance explanation is not mutually exclusive with the possibility that motives differ as a function of age. Indeed, it is logical that more deviant adolescents are also likely to engage in NUPZ for recreational motives such as experimenting or getting high.

4.1. Limitations First, the NSDUH is cross-sectional, meaning that only correlation can be established here. Second, roughly one-quarter of those asked to participate across the 2009–2011 NSDUH versions did not complete the survey, potentially introducing self-selection bias. Third, while honest reporting was likely maximized by NSDUH methods, self-report data may contain misreporting by respondents. Fourth, the NSDUH item on nonmedical use is somewhat complex (Boyd & McCabe, 2008), potentially resulting in some misclassification for NUPZ and other forms of nonmedical use examined here. Fifth, the assessment of past year anxiety disorder diagnosis is a single-item question with diminished reliability and validity versus a standardized interview. While inclusion of this item seemed warranted, given the significance of insomnia in anxiety disorders, the related ORs should be interpreted conservatively. Finally, the binge drinking variable captures episodes of alcohol use including consumption of five or more alcoholic drinks in one occasion, but National Institute on Alcohol Abuse and Alcoholisn (NIAAA) guidelines recommend a cut-off of four drinks for women (NIAAA, 2004). Use of five as a threshold here may have led to misclassification of some female alcohol users by the

Table 4 Substance Use and Mental Health Correlates for Past Year Nonmedical Use of Prescription Zolpidem.

Daily Smoking 30d Binge Alc. Heavy Drinking 30d Marijuana LT Cocaine LT Opioid LT Tranquilizer LT Stimulant PY SUD Tx PY Anxiety Dx PY MDD PY MH Tx PY SPD PY Suicidality

Adolescents (12–17 years)

Young Adults (18–25 years)

26–34 years

35–49 years

50 and older

Interaction

10.91 (6.89–17.26) 13.44 (9.28–19.45) 17.34 (10.34–29.10) 15.76 (10.11–24.55) 22.92 (13.78–38.12) 27.37 (18.48–40.52) 35.51 (23.70–53.20) 47.35 (29.77–75.31) 12.46 (6.71–23.14) 8.11 (4.79–13.73) 3.76 (2.24–6.29) 3.37 (2.25–5.05) not available not available

2.19 (1.64–2.92) 2.62 (1.93–3.56) 2.69 (1.87–3.87) 4.58 (3.43–6.11) 4.74 (3.49–6.42) 9.74 (7.28–13.03) 11.54 (8.78–15.16) 7.28 (5.38–9.86) 5.16 (2.99–8.89) 4.54 (3.31–6.24) 3.65 (2.61–5.11) 3.47 (2.60–4.62) 3.83 (2.98–4.93) 3.16 (2.21–4.51)

1.61 (0.97–2.65) 1.97 (1.24–3.13) 2.82 (1.64–4.83) 3.07 (1.84–5.12) 4.56 (2.93–7.12) 6.84 (4.20–11.14) 8.28 (5.23–13.10) 4.89 (2.78–8.62) 7.69 (3.14–18.80) 2.05 (1.02–4.12) 3.06 (1.84–5.07) 2.06 (1.18–3.59) 4.01 (2.66–6.03) 3.29 (1.44–7.53)

1.62 (0.96–2.73) 2.07 (1.16–3.70) 1.54 (0.80–2.56) 4.85 (2.52–9.34) 3.45 (2.12–5.61) 9.15 (6.11–13.70) 9.42 (5.76–15.43) 5.70 (3.40–9.55) 3.87 (1.36–11.02) 3.50 (1.73–7.10) 2.75 (1.71–4.44) 4.39 (2.70–7.15) 3.50 (2.03–6.03) 3.07 (1.54–6.12)

1.03 (0.29–3.60) 1.48 (0.64–3.41) 0.57 (0.07–4.66) 2.08 (0.44–9.79) 2.84 (1.46–5.54) 6.01 (3.02–11.96) 13.13 (7.09–24.29) 6.04 (2.87–12.70) 7.44 (2.10–26.35) 2.17 (0.71–6.70) 1.82 (0.66–5.07) 1.50 (0.55–4.12) 1.98 (0.74–5.31) 4.06 (1.25–13.19)

p b .0001a p b .0001a p b .0001a p b .0001c p b .0001a p b .0001a p = .0001b p b .0001a p = .159 p = .011 p = .757 p = .175 p = .665 p = .995

Notes: Bolded values are significant at a p b .001 level, 30d = 30-day, Binge Alc. = Binge Alcohol Use (5 or more alcoholic drinks on one occasion), LT = Lifetime, PY = Past Year, Tx = Treatment, Dx = Diagnosis, MDD = Major Depressive Disorder, MH = Mental Health, SPD = Serious Psychological Distress. a OR for adolescents N all. b OR for adolescents N all except 50 and older. c OR for adolescents N young adults and 26–34 years.

1316

T.S. Schepis / Addictive Behaviors 39 (2014) 1311–1317

NIAAA guidelines and some alterations to the ORs for analyses involving binge drinking. 4.2. Clinical Implications and Summary While NUPZ prevalence rates are lower than for other forms of nonmedical use, these results nonetheless reiterate the importance of education about the risks and benefits of zolpidem use and the particular importance of education about the risks of NUPZ. For those prescribed zolpidem, education should extend to the risks of diverting the medication and to the importance of proper medication storage and disposal. This is especially important for adults with children living in the home, as access to a parent’s controlled medication serves as an important source for nonmedical use (Schepis & Krishnan-Sarin, 2009). The literature on nonmedical prescription opioid use suggests other ways to reduce NUPZ, including use of prescription database monitoring programs and insurance databases to prevent inappropriate medication refills and “doctor shopping” (CDC, 2012); improvements in clinician education about optimal prescribing practices and non-medication treatment options (Logan, Liu, Paulozzi, Zhang, & Jones, 2013); guidelines for treatment, particularly for higher-risk patients, and/or the development of treatment algorithms (Voon & Kerr, 2013); and, adequate treatment of underlying mental and physical health conditions (Katz, El-Gabalawy, Keyes, Martins, & Sareen, 2013; Schepis & Hakes, 2011). Clinicians can also incorporate screening for NUPZ into a larger, brief screening for other forms of nonmedical and illicit drug use, given the high degree of correlate overlap. Universal, primary prevention programs appear to limit other forms of nonmedical use in a cost-effective fashion (Crowley, Jones, Coffman, & Greenberg, 2014; Spoth et al., 2013), and in line with this work’s support for the validity of evaluating NUPZ through Problem Behavior Theory, the emphasis of universal prevention on a variety of risk and protective factors suggests effectiveness at limiting NUPZ as well. In all, this work found that the correlates of both lifetime and past year NUPZ differ between cohorts, particularly between adolescents and older cohorts. In adolescents, other substance use was the most robust correlate of NUPZ. Other substance use was salient in other age groups, but the OR magnitudes were smaller than in adolescents. Conversely, the ORs for mental health issues were more consistent across ages. These cohortbased differences may indicate changes in motives through development, though this cannot be established based on the cross-sectional NSDUH design. Future work should employ longitudinal methods to investigate how risk factors, including ones not examined here like neurobiologcal factors, stress-coping and motives, change through development. Given the overlap between the correlates uncovered here and past work, previous longitudinal work on other substance use and work using Problem Behavior Theory can direct predictor selection to ensure theoretical guidance and maximize the likelihood of significant results. Role of the Funding Sources This manuscript was not directly funded by any source. The NSUDH is funded by SAMHSA; SAMHSA had no further role (above collecting the NSDUH data) in study design, the selection of relevant variables, analysis or interpretation of data, the writing of the report, or the decision to submit the paper for publication.

Contributors I was the primary writer of the manuscript and conducted all statistical analyses. I also decided upon study design, crafted the interpretation of the results and completed all editing of the manuscript. No other individuals made contributions to this manuscript.

Conflicts of Interest I have no conflicts of interest.

Acknowledgements None.

References American Psychiatric Association (2001). Diagnostic and statistical manual of mental disorders (4th ed., text revision). Washington, DC: Author. Boyd, C. J., & McCabe, S. E. (2008). Coming to terms with the nonmedical use of prescription medications. Substance Abuse Treatment, Prevention, and Policy, 3, 22. Boyd, C. J., McCabe, S. E., Cranford, J. A., & Young, A. (2007). Prescription drug abuse and diversion among adolescents in a southeast Michigan school district. Archives of Pediatrics and Adolescent Medicine, 161, 276–281. Brower, K. J., McCammon, R. J., Wojnar, M., Ilgen, M.A., Wojnar, J., & Valenstein, M. (2011). Prescription sleeping pills, insomnia, and suicidality in the National Comorbidity Survey Replication. Journal of Clinical Psychiatry, 72, 515–521. Centers for Disease Control and Prevention (2012). CDC grand rounds: Prescription drug overdoses — A U.S. epidemic. MMWR. Morbidity and Mortality Weekly Report, 61, 10–13. Chung, S. D., Lin, C. C., Wang, L. H., Lin, H. C., & Kang, J. H. (2013). Zolpidem use and the risk of injury: A population-based follow-up study. PLoS ONE, 8, e67459. Crowley, D.M., Jones, D. E., Coffman, D. L., & Greenberg, M. T. (2014). Can we build an efficient response to the prescription drug abuse epidemic? Assessing the cost effectiveness of universal prevention in the PROSPER trial. Preventive Medicine, 62C, 71–77. Dang, A., Garg, A., & Rataboli, P. V. (2011). Role of zolpidem in the management of insomnia. CNS Neuroscience & Therapeutics, 17, 387–397. Erman, M., Guiraud, A., Joish, V. N., & Lerner, D. (2008). Zolpidem extended-release 12. 5 mg associated with improvements in work performance in a 6-month randomized, placebo-controlled trial. Sleep, 31, 1371–1378. Fava, M., Asnis, G. M., Shrivastava, R. K., Lydiard, B., Bastani, B., Sheehan, D.V., et al. (2011). Improved insomnia symptoms and sleep-related next-day functioning in patients with comorbid major depressive disorder and insomnia following concomitant zolpidem extended-release 12.5 mg and escitalopram treatment: A randomized controlled trial. Journal of Clinical Psychiatry, 72, 914–928. Ford, J. A., & McCutcheon, J. (2012). The misuse of Ambien among adolescents: Prevalence and correlates in a national sample. Addictive Behaviors, 37, 1389–1394. Frey, D. J., Ortega, J.D., Wiseman, C., Farley, C. T., & Wright, K. P., Jr. (2011). Influence of zolpidem and sleep inertia on balance and cognition during nighttime awakening: A randomized placebo-controlled trial. Journal of the American Geriatrics Society, 59, 73–81. Hajak, G., Muller, W. E., Wittchen, H. U., Pittrow, D., & Kirch, W. (2003). Abuse and dependence potential for the non-benzodiazepine hypnotics zolpidem and zopiclone: A review of case reports and epidemiological data. Addiction, 98, 1371–1378. Jessor, R. (1987). Problem-behavior theory, psychosocial development, and adolescent problem drinking. British Journal of Addiction, 82, 331–342. Jessor, R., Donovan, J. E., & Costa, F. M. (1991). Beyond adolescence: Problem behavior and young adult development. Cambridge England; New York: Cambridge University Press. Jessor, R., & Jessor, S. L. (1977). Problem behavior and psychosocial development: A longitudinal study of youth. New York: Academic Press. Katz, C., El-Gabalawy, R., Keyes, K. M., Martins, S. S., & Sareen, J. (2013). Risk factors for incident nonmedical prescription opioid use and abuse and dependence: Results from a longitudinal nationally representative sample. Drug and Alcohol Dependence, 132, 107–113. Kessler, R. C., Barker, P. R., Colpe, L. J., Epstein, J. F., Gfroerer, J. C., Hiripi, E., et al. (2003). Screening for serious mental illness in the general population. Archives of General Psychiatry, 60, 184–189. Khantzian, E. J. (1985). The self-medication hypothesis of addictive disorders: Focus on heroin and cocaine dependence. The American Journal of Psychiatry, 142, 1259–1264. Kuehn, B.M. (2013). FDA warning: Driving may be impaired the morning following sleeping pill use. JAMA, 309, 645–646. Licata, S.C., Mashhoon, Y., Maclean, R. R., & Lukas, S. E. (2011). Modest abuse-related subjective effects of zolpidem in drug-naive volunteers. Behavioural Pharmacology, 22, 160–166. Logan, J., Liu, Y., Paulozzi, L., Zhang, K., & Jones, C. (2013). Opioid prescribing in emergency departments: The prevalence of potentially inappropriate prescribing and misuse. Medical Care, 51, 646–653. McCabe, S. E., Boyd, C. J., Cranford, J. A., & Teter, C. J. (2009). Motives for nonmedical use of prescription opioids among high school seniors in the United States: Self-treatment and beyond. Archives of Pediatrics and Adolescent Medicine, 163, 739–744. McCabe, S. E., Boyd, C. J., & Teter, C. J. (2009). Subtypes of nonmedical prescription drug misuse. Drug and Alcohol Dependence, 102, 63–70. McCabe, S. E., West, B. T., & Boyd, C. J. (2013). Medical use, medical misuse, and nonmedical use of prescription opioids: Results from a longitudinal study. Pain, 154, 708–713. Mitka, M. (2013). Zolpidem-related surge in emergency department visits. JAMA, 309, 2203. National Institute on Alcohol Abuse and Alcoholism (2004). NIAAA Newsletter (NIH Publication No. 04–5346). Bethesda, MD: Office of Research Translation and Communications, NIAAA, NIH, DHHS. Pompeia, S., Lucchesi, L. M., Bueno, O. F., Manzano, G. M., & Tufik, S. (2004). Zolpidem and memory: A study using the process-dissociation procedure. Psychopharmacology, 174, 327–333. Research Triangle Institute (2012). 2011 National Survey on Drug Use and Health: Methodology report. Rockville, MD: United States Department of Health and Human Services. Substance Abuse and Mental Health Services Administration. Center for Behavioral Health Statistics and Quality. Schepis, T. S., & Hakes, J. K. (2011). Nonmedical prescription use increases the risk for the onset and recurrence of psychopathology: Results from the National Epidemiological Survey on Alcohol and Related Conditions. Addiction, 106, 2146–2155. Schepis, T. S., & Krishnan-Sarin, S. (2008). Characterizing adolescent prescription misusers: A population-based study. Journal of the American Academy of Child and Adolescent Psychiatry, 47, 745–754.

T.S. Schepis / Addictive Behaviors 39 (2014) 1311–1317 Schepis, T. S., & Krishnan-Sarin, S. (2009). Sources of prescriptions for misuse by adolescents: Differences in sex, ethnicity, and severity of misuse in a population-based study. Journal of the American Academy of Child and Adolescent Psychiatry, 48, 828–836. Spoth, R., Trudeau, L., Shin, C., Ralston, E., Redmond, C., Greenberg, M., et al. (2013). Longitudinal effects of universal preventive intervention on prescription drug misuse: Three randomized controlled trials with late adolescents and young adults. American Journal of Public Health, 103, 665–672. Substance Abuse and Mental Health Services Administration (2009). Results from the 2008 National Survey on Drug Use and Health: National findings. Office of Applied Studies, NSDUH Series H-36, HHS Publication No. SMA 09-4434 (Rockville, MD). Substance Abuse and Mental Health Services Administration (2012). Results from the 2011 National Survey on Drug Use and Health: Summary of national findings. NSDUH Series H-44, HHS Publication No. (SMA) 12-4713. Rockville, MD: Substance Abuse and Mental Health Services Administration. Troy, S. M., Lucki, I., Unruh, M.A., Cevallos, W. H., Leister, C. A., Martin, P. T., et al. (2000). Comparison of the effects of zaleplon, zolpidem, and triazolam on memory, learning, and psychomotor performance. Journal of Clinical Psychopharmacology, 20, 328–337.

1317

Voon, P., & Kerr, T. (2013). "Nonmedical" prescription opioid use in North America: A call for priority action. Substance Abuse Treatment, Prevention, and Policy, 8, 39. WHO Expert Committee on Drug Dependence (2001). WHO Expert Committee on Drug Dependence. Thirty-second report. World Health Organization technical report series, 903, i-v. (pp. 1–26). Yang, Y. H., Lai, J. N., Lee, C. H., Wang, J.D., & Chen, P. C. (2011). Increased risk of hospitalization related to motor vehicle accidents among people taking zolpidem: A case-crossover study. Journal of Epidemiology/Japan Epidemiological Association, 21, 37–43. Zanarini, M. C., & Frankenburg, F. R. (2001). Attainment and maintenance of reliability of axis I and II disorders over the course of a longitudinal study. Comprehensive Psychiatry, 42, 369–374.