The Brief Addiction Science Information Source (BASIS)

STASH, Vol. 8(1) – Who are dual alcohol and marijuana users?

When people drink alcohol, they are more likely also to use marijuana than any other illicit drug (Degenhardt, Hall et al. 2001). However, little is known about this group of dual alcohol/marijuana users, or about how different levels of alcohol and marijuana use affect other risky behaviors. This week’s STASH reviews a study that investigates behavioral and psychosocial characteristics of those who use both alcohol and marijuana (Harrington, Baird et al. 2012). The investigators sought to determine if there are distinct clusters of alcohol/marijuana users and examine the range of potential negative consequences associated with this pattern of substance use.

Method

• Investigators conducted a secondary data analysis of data previously collected for a randomized control trial focusing on brief intervention at an emergency service within a  trauma center. 
• Participants were 427 individuals who (1) agreed to participate in a study about risk behavior reduction among alcohol and marijuana users and (2) reported using both alcohol and marijuana during the past 30 days.
• Participants reported how frequently they used alcohol and how frequently they used marijuana during the past 30 days.
• The researchers used cluster analysis to identify sub-types of dual users based on frequency of alcohol and marijuana use.In addition, participants completed the following self-report measures of negative consequences:
  • Alcohol Use Disorders Inventory Test (AUDIT): A measure of alcohol use frequency, binge drinking episodes, and alcohol-related negative consequences for the past year.
  • Marijuana Problem Scale (MPS): A measure of negative psychological, social, occupational, and legal consequences of marijuana use.
  • Noteworthy Index of Problems (NIP): A measure of frequency of psychosocial events related to alcohol and/or marijuana use during the past three months (e.g., saying harsh or cruel things to someone, getting in a physical fight, saying embarrassing things). The researchers calculated overall scores as well as separate alcohol-related and marijuana-related scores.
  • Injury Behavior Scale (IBC): A measure of injuries during the past year, with separate scores for total injuries, alcohol-related injuries, and marijuana-related injuries. The researchers calculated overall scores as well as separate alcohol-related and marijuana-related scores.
  • High Risk Behavior Scale (HRB): A measure of 13 risky behaviors, some related to alcohol and/or marijuana use (e.g., driving after marijuana use) and others not necessarily related to alcohol and/or marijuana use (e.g., speeding, being involved in a physical fight).
  • For each of these measures, except the HRB, items were combined to form a single score. For the HRB, items were kept separate.
  • The researchers used Analysis of Variance (ANOVA) and Tukey post-hoc tests to compare how participants in different clusters differed in terms of psychosocial and behavioral characteristics. 

Results

• The analyses revealed four distinct clusters of drug users based on the frequency of alcohol and marijuana use. The clusters were replicated independently using two sub-samples. 

Table 1. Identified clusters

Frequency of Alcohol Use
Frequency of Marijuana Use		Daily	Weekly
	Daily	Cluster 3 (N = 56)	Cluster 1 (N = 93)
	Weekly	Cluster 4 (N = 55)	Cluster 2 (N = 223)

• The researchers observed statistically significant differences among the four clusters for the AUDIT (F (3, 420)=39.92, p < .001), the MPS (F (3, 417)= 6.01, p < .001), the overall NIP (F (3, 410)=10.63, p < .001), the overall IBC (F (3, 423)=6.77, p < .001), and several of the HRB items.
• Figure 1a and Figure 1b show AUDIT, MPS, overall IBC, and overall NIP scores as a function of the four clusters of drug users.  

Figure 1a. AUDIT scores as a function of clusters of participants.

Figure 1b. MPS, total IBC and total NIP scores as a function of clusters of participants

• Post-hoc Tukey tests revealed the following statistically significant specific group differences:
  • For the AUDIT, Cluster 3 scores and Cluster 4 scores were both higher than Cluster 1 and Cluster 2 scores 
  • For the MPS Cluster 3 scores were higher than Cluster 2 and Cluster 4 scores
  • For the NIP, Cluster 3 scores were higher than Cluster 1 and Cluster 2 scores
  • For the IBS, Cluster 3 scores were higher than Cluster 1 and Cluster 2 score

Limitations

• This study uses a self-reported data.
• The study measured only frequency of alcohol and marijuana use, but not quantity.

Conclusion

These analyses identified four distinct groups of alcohol and marijuana users according to the frequency of using both substances. The results showed that individuals who used alcohol daily experienced more alcohol-related problems than individuals who used it weekly, independent of marijuana use. The pattern of marijuana use was less clear. For instance, people who used marijuana daily reported more instances of driving under the influence of marijuana than did weekly marijuana users, regardless of how frequently they used alcohol. However, for the number of marijuana related problems, daily marijuana users were not significantly different from weekly marijuana users, if the marijuana user also used alcohol weekly. The daily users of both substances reported more marijuana-related problems than weekly marijuana users. The authors speculate that individuals who use both substances daily might need targeted intervention and treatment efforts because of the highest number of associated problems.

-Julia Braverman

What do you think?  Please use the comment link below to provide feedback on this article.

References

Degenhardt, L., W. Hall, et al. (2001). "The relationship between cannabis use and other substance use in the general population." Drug and Alcohol Dependence 64(3): 319-327.

Harrington, M., J. Baird, et al. (2012). "Identifying subtypes of dual alcohol and marijuana users: A methodological approach using cluster analysis." Addictive Behaviors 37(1): 119-123.

STASH, Vol. 7(11): Do smart kids become adult users of illegal drugs?

Children who score relatively high on intelligence tests become relatively healthy adults, in some important respects. For instance, White and Batty (2011) noted that they are less likely to smoke, more likely to engage in physical activity, and more likely to eat healthy foods than their peers. However, research is equivocal regarding a link between childhood IQ scores and adult use of illegal drugs (as reviewed by White & Batty, 2011). In this week’s STASH, we review a recent longitudinal, population-based study investigating the connection between childhood IQ and adult use of illegal drugs (White & Batty, 2011).                   

Methods

• The authors used data from the 1970 British Cohort Study, an ongoing longitudinal study of children born in Great Britain during April, 1970.
• The children in the study were enrolled at birth and were invited to participate at ages 5, 10, 16, 26, and 29-30.
• At age 5, participants completed four tests purported to measure general cognitive ability (e.g., the Human Figure Drawing test, shown at right). The researchers transformed scores to the standard IQ distribution (mean = 100, SD = 15). Also at these assessments, the researchers made note of parents’ social class.
• At age 16, participants indicated their lifetime use of cannabis and cocaine (yes/no).
• At age 30, participants indicated their past-year use of several illegal drugs (yes/no).
• The researchers completed logistic regression to estimate the relation between childhood IQ and participant gender with illegal drug use at ages 16 and 30. They modeled childhood IQ as tertiles, with the lowest IQ group as the reference group. Positive odds ratios therefore indicate that a higher risk of illegal drug use is associated with moderate or high childhood IQ.*  In the results presented here, researchers statistically controlled for parental social class at age 5. 

Results

• The researchers present drug use outcomes separately for men and women at both 16 and 30 years.**
• The strongest effects emerged from comparisons between the highest childhood IQ tertile and to the lowest IQ tertile. Figure 1 illustrates only these comparisons.*** 
• Male and female participants with relatively high childhood IQ were more likely to report having used cannabis but not cocaine at age 16.
• Male participants with relatively high childhood IQ were more likely to report past-year use of cannabis and cocaine at age 30.
• Female participants with relatively high childhood IQ were more likely to report past-year polydrug use and use of cannabis and cocaine at age 30.
• Associations between childhood IQ and adult illegal drug use were stronger for women than for men.

Figure 1
Odds ratios and 95% confidence intervals for illegal drug use outcomes at 16 and 30 years, separately for men and women. Adapted from White and Batty, 2011. (Click to enlarge.)

Limitations

• Participants did not report important features of their drug use, such as frequency, quantity, and consequences. Accordingly, these results do not speak to the potential association between childhood IQ and substance use disorders (e.g., dependence, abuse).
• As with most longitudinal studies, this research was marked by some attrition. For instance, only 46% of the participants at the 30-year follow-up completed all earlier surveys. Indeed, participants who completed the 30-year follow-up had slightly higher childhood IQ scores than participants who did not. It is unclear how this selective attrition might have influenced observed relationships between childhood IQ and adult drug use. 

Conclusion
In this large, population-based cohort study, the authors observed a consistent pattern: higher childhood IQ predicted greater likelihood of illegal drug use during adulthood, independent of parental social class. The authors speculate about a potential pathway involving personality. According to this reasoning, high IQ individuals use drugs because they are characterized by traits linked with drug use, such as stimulation seeking and openness to experience. The presence of stronger associations for women than men requires further exploration, as does the potential association between childhood IQ and adult substance dependence and abuse.

-Heather Gray

What do you think?  Please use the comment link below to provide feedback on this article.

Reference

White, J., & Batty, G. D. (2011). Intelligence across childhood in relation to illegal drug use in adulthood: 1970 British Cohort Study. Journal of Epidemiology and Community Health, Online first: November 14, 2011.



* Preliminary analyses suggested that modeling IQ as a linear term was inappropriate.
** The interaction term between childhood IQ and participant gender for drug use outcomes was significant at 30 years but not at 16 years.
*** Comparisons between the moderate IQ tertile and the lowest IQ tertile were generally in the same direction. Readers interested in more details can see White and Batty (2011).

 

STASH, Vol. 7(10) – How does marijuana smoking impact driving? A randomized placebo- controlled study

About 10 million Americans report that they have driven under the influence of an illicit drug at least once during the past year (US Department of Health and Human Services, Substance and Mental Health Services Association, Office of Applied Studies, 2007). Marijuana is the most commonly used illicit drug (ibid.). This week’s STASH reviews a randomized control study that examines the effect of smoking marijuana versus a placebo on a driving simulator (Anderson, Rizzo, Block, Pearlson & O’Leary, 2010). 

Method

• Investigators recruited 50 men and 35 women 18  - 31 y/o who used marijuana at least once but fewer than 10 times per month during the past year. The researchers excluded participants whose urine-screening test detected any drug other than tetrahydrocannablnol (THC) the psychoactive ingredient in marijuana.
• During the baseline session, participants became familiarized with a math task that would serve as a distracting task during the driving session.
• During the smoking session, participants smoked a cigarette that contained 2.9% (active) or 0% (placebo) of THC. The instructions encouraged participants to consume the entire cigarette, but participants could stop anytime if they felt uncomfortable. 
• Using a driving simulator, participants then performed a driving assessment (see Figure 1). Each participant drove for approximately 15 miles. An uneventful section of the drive lasted one minute. Then, the driving was interrupted by the following events:
  • Multitasking. Drivers completed a math test designed to distract them from the driving. The investigators measured the number of math errors made compared to the baseline, as well as the speed and the steering wheel position.
  • Response to emergency vehicle.  The task assessed attention to the appearance of a police car. The dependent measures were speed, steering position and reaction time.  
  • Go/NoGo. The task measured safe driving through the yellow light. Safe driving was operationalized as no hesitation in making a decision.
  • Dog incursion avoidance. The task measured safe avoidance of a dog, measured as the ability to stop the car or steer clear of the dog.
  • Intersection incursion avoidance. The task measured the speed at first contact and avoidance tactic.

Figure 1. View of a driving simulator (copied from Anderson, Rizzo, Block, Pearlson & O’Leary, 2010)



• The researchers assessed participants’ heart rate, self-reported level of “highness” (0: no effect – 10:highest) and sleepiness (Stanford Sleepiness Scale; Hoddesetal. 1973) at baseline, after smoking and after driving. 

Results

• The results include only those participants who completed the entire cigarette. The analytic sample included 49 men (25 in the active group) and 24 women (9 in the active group)
• As expected, marijuana smoking significantly increased heart rate (F = 66.4, p < .001) immediately after smoking, and subjective feeling of “highness” compared to placebo (F = 65.1, p < .001) immediately after smoking.
• Women rated themselves as being “higher” than men for both active marijuana and placebo (F = 4.6, p < .05).
• Men were less sleepy than women after driving (F = 4.6, p < .04). Women who smoked active marijuana reported to be more sleepy than men after driving (F = 6.0, p < .02), but not immediately after smoking.
• Both marijuana and placebo groups performed similarly on all driving tasks; there were no sex differences.

Limitations

• The driving simulator results are not necessary directly applicable to real life driving situations.
• This study only investigated the first 15 miles of driving under certain conditions. It is still possible that marijuana affects prolonged driving, or driving under conditions that were not investigated in the study (e.g., slippery road or obscure vision).
• This study uses a small sample, especially in the active marijuana female group. This might be a reason for a failure to find sex X drug effect interaction for all measures, but sleepiness.

Conclusion

A meta analytic study concluded that there was a subtle effect of marijuana on driving performance (Berghaus, Sheer, & Shmidt, 1995). Anderson et al. (2010) provided results that did not support these earlier findings. Using a modern driving simulator might have influenced these results; in addition, the different finding might have emerged because of Anderson et al. investigated different driving tasks and driving evaluation methods.

Results obtained in a lab are not always generalizable to real life. Specifically, previous studies showed that marijuana influenced driving performance impairments are more likely to be manifest within a driver stimulator test compared to on road settings (US Department of Transportation, 1993). The present study did not find any marijuana effect on simulated driving. However, we should interpret this null finding with caution because there are many methodological (e.g., large measurement error) and analytical (e.g., small sample size; small effect size) reasons for failing to finding differences between groups. In addition, it is possible that smaller undetected effects exist or that marijuana impacts actual driving. 

-Julia Braverman

What do you think?  Please use the comment link below to provide feedback on this article.

References

Anderson, B.M., Rizzo, M., Block, R. I., Pearlson, G. D., & O’Leary, D.S. (2010) Sex differences in the effects of marijuana on simulated driving performance. Journal of Psychoactive Drugs, 40, 19 – 30.

Berghaus G, Sheer N, & Schmidt P. (1995) Effects of cannabis on psychomotor skills and driving performance–A meta-analysis of experimental studies. In CN Kloeden and AJ McLean (eds.), Proceedings of the 13th International Conference on Alcohol, Drugs and Traffic Safety. Adelaide, Australia: The University of Adelaide, NHMRC Road Accident Research Unit, 403–409.

Hoddes. E.: Zarcone. V.; Smythe. H.: Phillips. R. & Dement. W.C. (1973)

Quantificalion of sleepiness: A new approach. Psvchophysiology 10, 11-36.

US Department of Transportation, National Highway Traffic Safety Administration. (1993) It appears performance is more affected by THC in laboratory (settings) than (in) actual driving tests. Marijuana and Actual Driving Performance: Final Report.

US Department of Health and Human Services, Substance and Mental Health Services Association, Office of Applied Studies. (2007) 2006 National Survey on Drug Use and Health: National Results.

STASH, Vol. 7(9): Taking Matters into Their Own Hands: Cannabis Use Following Inpatient Treatment for PTSD among a Sample of Military Veterans

Military Veterans who have experienced trauma are at increased risk for PTSD symptoms and cannabis use among other disorders (Bremner et al., 1996). In line with Khantzian’s self-medication hypothesis (Khantzian, 1985), people might turn to cannabis as an attempt to cope with PTSD symptoms, including anxious arousal and sleep disturbance. Research suggests that individuals who have experienced trauma are a particularly vulnerable population for relapsing to cannabis use after a period of discontinuation (Bremner et al., 1996).
The study we review in this week’s STASH (Bonn-Miller et al., 2011) was the first to examine cannabis use among American Veterans following inpatient treatment for PTSD. This study prospectively examined the relationship between treatment changes in PTSD symptom severity and cannabis use four months after discharge from residential treatment for PTSD. The researchers hypothesized that lower levels of change in PTSD symptom severity (i.e., PTSD symptoms not improving) would predict more frequent cannabis use at 4-month follow-up post-treatment than their symptom improving counterparts. In addition, the researchers examined patterns of substance use specifically by looking at two additional substance use classes: alcohol and opiates. Use of these substances tends to co-occur among those with PTSD at high rates (Chilcoat & Menard, 2003).

Methods

• The sample was composed of 432 male military Veterans (56.9% white, Mage= 51.06 years)

• All Veterans had a primary diagnosis of PTSD and were inpatients at the Department of Veterans Affairs (VA) Medical Center in Palo Alto.

• The researchers assessed the sample at 3 points:

              1)    At treatment intake
              2)    At treatment discharge
              3)    At 4-month follow-up which was administered by mail

   Measures

• The researchers used the PTSD Checklist—Military Version (PCL-M; Weathers et al., 1993) to measure PTSD symptom severity at treatment intake and treatment discharge.

• The PCL-M asks questions about DSM-IV PTSD symptoms; respondents indicate the degree to which they have been bothered by each of the 17 symptoms within the past month.

• The researchers calculated PTSD symptom severity change as the PCL-M score difference between intake and discharge. Lower levels of change indicate less symptom improvement.

• The researchers used the Northeast Program Evaluation Center (NEPEC) to assess substance use at intake and follow-up. Patients reported the number of days that they used cannabis, alcohol, cocaine, opiates, and amphetamines during the 2 months before intake assessment as well as 1 month before follow-up assessment (i.e., 3 months after discharge)

Results

• Roughly half (54.3%; n=19) of participants who used cannabis pre-treatment relapsed to cannabis use after treatment; 10.1% (n= 40) of participants who did not report cannabis use pre-treatment reported cannabis use after treatment.

• Participants with lower levels of PCL-M score improvement reported significantly greater frequency of cannabis use during the 30 days before the 4-month follow-up, even after controlling for frequency of cannabis use pre-treatment (β = .11; p < .05)

• Using post-hoc analyses, the researchers explored whether changes in certain PTSD symptom clusters were predictive of cannabis use at follow-up after controlling for frequency of cannabis use pre-treatment (see Table 1):

        -    Levels of improvement in PTSD “re-experiencing” severity were not associated             with cannabis use at follow-up (β = -0.09; p= .07),
        -    Lower levels of improvement in PTSD “avoidance/numbing” severity significantly             predicted more cannabis use at follow-up (β =-0.11; p< .05)
        -    Lower levels of improvement in PTSD “hyperarousal” symptoms significantly             predicted more cannabis use at follow-up (β = -0.09; p= .05).

• Levels of improvement in PTSD symptoms were not associated with the frequency of any other non-cannabis substance use, suggesting a substance specificity effect for cannabis.

Table 1: Relationship between PCL-M Symptom Improvement during Treatment and Follow-Up Cannabis Use

PCL-M Change (Intake to Follow-Up)	β	P
Re-experiencing	- .09	.07
Avoidance/numbing	- .11	<.05
Hyperarousal	- .09	.05

Limitations

• The researchers did not collect information regarding co-occurring Axis-I or Axis-II diagnoses. As a result, they could not study the relationships between substance use and co-morbid conditions.

• Because the VA Palo Alto Medical Center was responsible for collection of data and determination of benefits, participants might have been reluctant to disclose some patterns of drug use. A third party assessment (both in terms of person and institution) might have helped to prevent this potential bias.

• This study used a limited sample, only male military Veterans. This might mean that these findings will not generalize to women or victims of other kinds of trauma.

Discussion
Veterans who experienced less improvement in their PTSD symptoms used cannabis more frequently after treatment. This was particularly true for the PTSD symptoms avoidance/numbing and hyperarousal.  These results support the idea that Veterans with PTSD might turn to cannabis to cope with hyperarousal symptoms like irritability and sleep disturbances, and numbing symptoms like emotional detachment, in line with with Khantzian’s self-medication theory (1997). The substance specificity effect suggests that cannabis might be particularly helpful in alleviating these PTSD symptoms, at least temporarily.
Future research should clarify the perceived and actual benefits of cannabis use among the Veteran population, as well as the detrimental effects of cannabis use.  Future research also could seek to identify the perceived and actual benefits of other kinds of coping mechanisms among those who have PTSD.

References
Bonn-Miller, M.O, Vujanovic, A.A., Drescher, K.D. (2011) Cannabis use among military veterans after residential treatment for posttraumatic stress disorder. Psychology of Addictive Behaviors, 25(3), 485-491.

Bremner, J. D., Southwick, S. M., Darnell, A., & Charney, D. S. (1996).Chronic PTSD in Vietnam combat Veterans: Course of illness and substance abuse. The American Journal of Psychiatry, 153, 369–375.

Chilcoat, H. D., &Mennard, C. (2003). Epidemiological investigations: Comorbidity of posttraumatic stress disorder and substance use disorder. In P. Ouimette& P. J. Brown (Eds.), Trauma and substance abuse: Causes, consequences, and treatment of comorbid disorders. Washington, DC: American Psychological Association.

Khantzian, E. J. (1985). The self-medication hypothesis of addictive disorders: Focus on heroin and cocaine dependence. American Journal of Psychiatry, 142, 1259–1264.

Weathers, F., Litz, B., Herman, D., Huska, J., & Keane, T. (October, 1993). The PTSD checklist (PCL): Reliability, validity, and diagnostic utility. Paper presented at the Annual Convention of the International Society for Traumatic Stress Studies, San Antonio, TX.

STASH, Vol. 7(8) - Patterns of cocaine use among male and female undergraduates

Recent studies reveal that cocaine use is more prevalent among young adults compared to other user population groups (Substance Abuse and Mental Health Administration (SAMHSA), 2008). Estimates of past-year cocaine use among young adults have steadily increased over the last ten years (Kasperski et al., 2011). However, male and female college students might be different in terms of cocaine use in some important ways.  While some studies cite higher prevalence rates of cocaine use in young males, others cite riskier behavior in female cocaine users, suggesting that young females might be more susceptible than males to developing serious problems with cocaine (Kasperski et al., 2011). This week’s STASH features a study that analyzed gender differences in prevalence of cocaine exposure and actual cocaine use in a longitudinal sample of 1,253 college students.

Methods

• The researchers administered annual questionnaires from the National Survey on Drug Use and Health to track a sample of college students throughout their four years at a large, mid-Atlantic university.
• The analyses examined frequency of cocaine use and DSM-IV criteria for cocaine dependence in the sample of students. Analyses covered two additional kinds of data: age at first offer of cocaine and age at first use of cocaine.
  • Age at first offer of cocaine was operationalized as the participant’s age during the first assessment in which an offer was reported. For individuals who were offered cocaine prior to Year 1, age at first offer was captured by self-report.
  • Age of first use of cocaine was computed similarly.
• 1,253 college students took part in the study (70.8% White; 48.5% Male). Follow-up rates ranged from 91.1% in Year 2 to 87.6% in Year 4. Researchers excluded 192 students because of missing data.
• Data from the first four years of college were analyzed to detect changes over time and possible gender differences.

Results
Gender differences appeared in opportunities to use and patterns of actual usage.

• As displayed in Figure 1, with the exception of the first year of college, males reported significantly higher opportunities to use than women (41% vs. 31% overall). Males reported an increase in exposure opportunity over time (Year 1 to Year 2:  p <.03; Year 2 to Year 3: p=.005), while females did not. This suggests a gender difference in opportunities to use.

 

• However, there was no significant difference between males and females in the sub-sample (n=584) of those who used cocaine, given the opportunity. Both genders showed an overall rate of 37%. While men were more likely to be presented with opportunities to use cocaine, men and women were equal in terms of use when given the opportunity to use.

• Women tended to use cocaine more frequently than men, both in terms of lifetime use as of Year 4 (27.78 days vs. 12.75 days during Year 4) and during the peak year of use (18.39 days vs. 8.83 days). For females, annual changes in frequency of cocaine use were statistically significant from Year 1 to Year 2 (p=.001) and Year 2 to Year3 (p=.009).

 

• Women were more likely than men to meet DSM-IV criteria for dependence (9.3% vs. 2.5%).

Limitations

• The authors note several limitations, including that their study cannot rule out the possibility of gender differences in reporting dependence symptoms. Males may report fewer symptoms that would meet DSM-IV criteria for dependence.
• They also note that the study’s findings have limited generalizability to other types of colleges and geographic regions.

Discussion

This study’s longitudinal design and high follow-up rate is a strong start for identifying patterns of risky behavior. The results suggest that there are possible gender differences in how college students experience opportunities for use of illicit drugs, in line with previous studies. The results help resolve apparent paradoxes in gender-based patterns of cocaine use. Men reported more opportunities to use cocaine than women. Research that uncovers the settings in which young men typically find opportunities to use cocaine might inform primary prevention efforts. However, among the smaller sample of students who actually used cocaine given the opportunity, there was no significant difference between the two genders. Women who chose to use cocaine used it more frequently and were more likely to meet DSM-IV criteria for dependence. Animal studies have suggested sex differences in endocrine-related cocaine response (Kasperski et al., 2011). Early interventions targeted for women could emphasize these differences by teaching girls about their potential heightened sensitivity to cocaine use. More research remains to be done in specific gender differences in developing substance abuse dependence.

- Kat Belkin

What do you think?  Please use the comment link below to provide feedback on this article.

References

Kasperski, S. J., Vincent, K.B., Caldeira, K.M., Garnier-Dykstra, L.M., O'Grady, K.E., Arria, A.M. (2011). "College students' use of cocaine: Results from a longitudinal study." Addictive Behaviors 36, 408-4011.

Substance Abuse and Mental Health Services Administration (2008). Results from the 2007 National Survey on Drug Use and Health: National findings (DHHS Publication No. SMA 08-4343). Rockville, MD: Office of Applied Studies.

STASH, Vol. 7(7) - Flying “High”: Drug Use Prevalence among Aviation Employees

Since the early 1980s, more employers have required their employees to complete random drug tests, for drugs such as marijuana, cocaine, and more (Li et al., 2011). In particular, the federal government mandates drug testing among employees who hold “safety-sensitive” occupations, including airline pilots. This week, STASH reviews recent research that examined the prevalence rates of positive drug screens among “safety-sensitive” aviation employees who either completed random drugs tests or recently were involved in an aviation accident (Li et al., 2011). The authors examined whether the use of illicit drugs is associated with an increased risk of airline accident involvement.

Method

• The researchers analyzed aviation employer (i.e., all major airlines, commuter air carriers/air taxi, and non-FAA air traffic controllers) records that detailed the results of drug testing for marijuana, cocaine, amphetamines, opiates, and phencyclidine among employees who held “safety-sensitive” positions (e.g., flight crewmembers, flight attendants, flight instructors, aircraft dispatchers). The records covered the years from 1995 to 2005.
• Analyses utilized two sources of data: standard random drug testing and post-accident drug testing.
  • Random testing accounts for about 96% of all drug related records, and involve urine testing a random selection of at least 25% of “safety-sensitive” employees.
  • Post-accident urine testing occurs within 32 hours of an accident for employees whose performance might have contributed to an aviation accident (i.e., an event resulting in fatality or serious injury, or substantial damage to an aircraft).

Results

• From 1995-2005, the prevalence of positive screens among random drug tests was 0.64% (i.e., 7,211 of 1,129,922). The prevalence of positive screens among post-accident tests was 1.82% (i.e., 91 of 4,977).
• The odds ratio of accident involvement for employees who tested positive for drugs was 2.90 (95% CI = 2.35-3.57). In other words, employees who tested positive for drugs were almost three times as likely to be involved in an accident as employees who tested negative for drugs.
• Accounting for the prevalence of drug violations observed from random tests, this odds ratio suggests that 1.2% of all aviation accidents were attributable to drug violations during this period.^[i]
• Figure 1 (adapted from Li et al., 2011) shows the distribution of drug violations among aviation employees who hold safety-sensitive positions. Overall, marijuana accounted for 67.3% of violations, cocaine accounted for 23.9%, amphetamines accounted for 6.1%, opiates accounted for 2.1%, and phencyclidine accounted for 0.6%.

 

Limitations

• The authors noted several limitations, including that their analyses were restricted to a segment of all aviation events; they could not include general (e.g., non-commercial) aviation accidents, which account for 90% of all aviation accidents.
• They also note that they could not assess potential confounding individual-level variables, due to nature of data (i.e., aggregate reports to FAA).

Discussion

Drug testing for employees who have “safety-sensitive” occupations is popular among the public and policy-makers. Research related to this practice is limited, however. These results suggest that drug violations generally are rare among flight crews; about one in every 2,000 random drug tests produces a positive result. Although the use of illicit drugs is rare, it is consequential; it nearly triples the risk of accident involvement. More research is needed to determine whether this observation extends to general aviation accidents.

-Debi LaPlante

What do you think?  Please use the comment link below to provide feedback on this article.

Reference

Li, G., Baker, S. P., Zhao, Q., Brady, J. E., Lang, B. H., Rebok, G. W., & DiMaggio, C. (2011). Drug violations and aviation accidents: Findings from the US mandatory Drug Testing Programs. Addiction, 106, 1287-1292.

--

^[i]Attributable Risk = [p x (OR-1)]/[1+ p x (OR-1)], where p=the prevalence of drug violations detected in random testing.

STASH, Vol. 7(6) – Primary Care Treatment for People with Opiate Addiction: Stronger Support for a Team Effort

Fewer than 25% of individuals who are opioid dependent receive addiction treatment (SAMHSA, 2007). Within general medical practices, buprenorphine is a promising tool for the treatment of opioid dependence; this strategy is commonly referred to as office-based opioid treatment (OBOT). However, many physicians cite insufficient nursing support as a barrier to their use of OBOT (Walley et al., 2008). Boston Medical Center has instituted a collaborative care OBOT program that emphasizes collaboration among nurses, physicians, and pharmacists. In this week’s STASH, we review a new study that describes the effectiveness of this program and explores potential patient-level moderators of treatment success (Alford et al., 2011).

Methods

• Four hundred eight patients with opioid addiction took part in the program (M age = 39 years; SD = 11; 66.5% White, 16.5% Black, and 15.4% Hispanic; 66.0% Male). Researchers excluded 26 patients who left treatment because of preexisting legal or medical conditions or a need to re-locate from the analyses. 
• The collaborative care OBOT program included a full-time nurse program director, nurse care managers (NCMs), a program coordinator, and generalist physicians with part-time clinical practices. NCMs’ clinical responsibilities included assessment, patient education, development treatment plans, overseeing medication management, monitoring treatment adherence, and communicating with other providers. Physicians examined patients, prescribed buprenorphine, and provided follow-up as needed. 
• The treatment model included 3 stages:
  • Assessment:  The nurse program director screened patients and triaged them to intake appointments. At intake, NCMs and physicians jointly administered a thorough medical evaluation and developed a treatment plan. 
  • Induction and Stabilization: Physicians prescribed buprenorphine and NCMs monitored induction, which occurred on-site on Day 1, and dose stabilization, which occurred at home on Days 2-7. NCMs assessed for withdrawal and adverse effects.
  • Maintenance: Staff provided ongoing monitoring at follow-up appointments at least weekly for the first 4 to 6 weeks followed by visits every 2 to 4 weeks if adherent, gradually tapering out to once a month and then once every 3 months. NCMs conducted scheduled and random callback urine drug tests at least once every 3 months, and encouraged patients to attend self-help groups and mandatory weekly addiction counseling.
• The researchers defined treatment adherence as 4 consecutive urine drug tests with negative results for illicit drugs; study participants also needed to evidence positive results for buprenorphine, and attend at least 3 of 4 counseling sessions per month.
• To fulfill the criteria for successful treatment, patients had to meet three conditions:retention at 12 months with a consistent pattern of treatment adherence
  • buprenorphine dosage taper after 6 months of treatment adherence 
  • absence of illicit opiate use for at least 6 months 

Results

• At 1 year, 196 of 382 patients (51.3%) underwent successful treatment, 162 patients (42.4%) showed an unsuccessful outcome, and 24 (6.3%) transferred to methadone maintenance treatment for more structured care or for full opioid agonist therapy.
• Urinalysis indicated a 4.7% opioid positive test rate for patients who were still in the OBOT program at 12 months.
• As Figure 1 illustrates, moderators of treatment success include age, race, employment, and illicit buprenorphine use. Patients who were older, employed, and self-maintained with illicit buprenorphine had significantly higher odds of treatment success (all p’s < 0.05). Those of African American or Hispanic/Latino race had significantly lower odds of treatment success (p’s < 0.05).

Figure 1
Odds Ratios for Factors that Influence Treatment Success at 12 Months for Patients Entering OBOT in Primary Care

 

 
   
Limitations

• This study did not have a control group. This makes it impossible to directly compare the effectiveness of the program to other traditional methods of care.
• Follow-up information was not available after patient departure for the program. A broader scope of outcomes is needed to assess the success of the OBOT program.

Discussion

This study shows some support for OBOT within a collaborative care model. The results compare favorably to those of similar programs that treat opioid addiction. Gerra et al. (2006) found a 43.8% retention rate at 12 months for heroin users accepting buprenorphine treatment in an outpatient setting, and Pinto et al. (2010) report a 50% retention rate at 6 months for a buprenorphine treatment program with less clinical support. Methadone treatment consistently has evidenced a higher retention rate than buprenorphine treatment, but has shown no difference in suppression of illicit opioid use for those who remain in treatment (Mattick et al., 2008; Fischer et al., 1999). The high level of clinical support and role sharing in the collaborative care model might account for some treatment success, but a 51.3% retention rate is difficult to applaud – treatment can do better. 

The finding that self-maintaining with illicit buprenorphine at admission predicted later success is interesting.  Schuman-Olivier et al. (2010) found that individuals who are dependent on opioids often illicitly use buprenorphine to manage pain and withdrawal rather than to get high. Similarly, Alford et al. speculate that those who used buprenorphine illicitly at pre-admission are highly motivated to find a safe and affordable source of buprenorphine for treatment of withdrawal symptoms. Further study is needed to better understand the ways that race, age, and employment status predict success in OBOT.

-Kat Belkin

What do you think?  Please use the comment link below to provide feedback on this article.

References

Alford, D. P., LaBelle, C.T., Kretsch, N., Bergeron, A., Winter, M., Botticelli, M., Samet, J.H.  (2011). "Collaborative Care of Opioid-Addicted Patients in Primary Care Using Buprenorphine: Five-Year Experience." Archives of Internal Medicine 171(5): 425-431.

Fischer, G., Gombas, W., Eder, H., Jagsch, R., Peternell, A., Stühlinger, G., Pezawas, L., Aschauer, H.N., Kasper, S. (1999). "Buprenorphine versus methadone maintenance for the treatment of opioid dependence." Addiction 94(9): 1337-1347.
   
Gerra, G., Leonardi, C., D’Amore, A., Strepparola, G., Fagetti, R., Assi, C., Zaimovic, A., Lucchini, A. (2006). "Buprenorphine treatment outcome in dually diagnosed heroin dependent patients: A retrospective study." Progress in Neuropsychopharmacology and Biological Psychiatry 30(2): 265-272.

Mattick, R. P., Kimber, J., Breen, C., Davoli, M. (2008). “Buprenorphine maintenance versus placebo or methadone maintenance for opioid dependence.” Cochrane Database of Systematic Reviews (2).
   
Pinto, H., Maskrey, V., Swift, L., Rumball, D., Wagle, A., Holland, R. (2010). "The SUMMIT trial: A field comparison of buprenorphine versus methadone maintenance treatment." Journal of Substance Abuse Treatment 39: 340-352.
   
SAMHSA (2007). Results from the 2006 National Survey on Drug Use and Health: National Findings. Substance Abuse and Mental Health Services Administration. Rockville, MD, Office of Applied Studies.

Schuman-Olivier, Z., Albanese, M., Nelson, S. E., Roland, L., Puopolo, F., Klinker, L., & Shaffer, H. J. (2010). Self-treatment: Illicit buprenorphine use by opioid dependent treatment seekers. Journal of Substance Abuse Treatment, 39, 41-50.
   
Walley A.Y., A. J. K., Cheng D.M. (2008). "Office-based management of opioid dependence
with buprenorphine: clinical practices and barriers." Journal of General Internal Medicine 23(9): 1393-1398.

STASH, Vol. 7(5) – Pathways from childhood maltreatment to adolescent cannabis abuse

Both childhood maltreatment and personality functioning are linked to adolescent cannabis abuse and dependence (CAD; Oshri, Rogosch, Burnette, & Cicchetti, 2011). However, little is known about the developmental sequence of childhood maltreatment, personality changes, and adolescent CAD. This week’s STASH reviews a longitudinal study that explored (a) the association between childhood maltreatment and its severity as well as individual personality differences, and (b) the extent to which these personality individual differences help explain the connection between childhood maltreatment and adolescent CAD (Oshri, et al., 2011).

Methods

• The research sample consisted of 415 children, 259 of whom experienced maltreatment as indicated by Child Protective Services reports, and 156 of whom reportedly had no maltreatment experiences. These two groups were comparable on socio-demographic characteristics.
  • Researchers characterized maltreatment according to severity and type (i.e., neglect, physical abuse, emotional abuse, sexual abuse).
• Researchers tracked and interviewed participants three times from age 7 through age 15.
  • Wave 1 occurred when the children were 7-9 years old. The primary measures were ego control (i.e., EC; an individual’s tendency to express emotional impulses) and ego resiliency (i.e., ER; the capacity to adapt one’s level of EC to meet situational needs).
  • Wave 2 occurred when the children were 10-12 years old. The primary measures were externalizing and internalizing behaviors. Two camp counselors independently rated children using the Teacher Report Form (Achenbach, 1991).
  • Wave 3 occurred when the children were 13-15 years old. Researchers interviewed participants about their current (i.e., past 3 months) CAD using the Diagnostic Interview Schedule for Children (DISC; Shaffer et al., 1993).
• Researchers created and tested a model predicting Wave 3 CAD based on maltreatment severity and type, Wave 1 EC and ER, and Wave 2 internalization and externalization problems.

Results

• As expected, childhood maltreatment was associated with problem behaviors at Wave 2 and CAD symptoms at Wave 3.
• Tests of mediation revealed that EC and ER at Wave 1 significantly mediated the relationship between maltreatment severity and problem behaviors at Wave 2. Externalizing problem behaviors at Wave 2 mediated the relationship between maltreatment severity and CAD symptom count at Wave 3. Figure 1 displays the direct paths of these relationships.

Figure 1. Structural model of relations among child maltreatment, ego control and ego resiliency, externalizing behaviors, and cannabis abuse and dependence symptoms. 

Unstandardized and standardized (in parentheses) path coefficients, p < .05. **p < .01.
For simplification, we have removed gender and internalizing behavior from this model and we have not depicted the factor compositions of the latent variables (i.e., maltreatment severity and externalizing behavior). For full details, please refer to Oshri et al., 2001, p. 7)

Limitations

• Measures focused heavily on peer and observer reports, which might not capture internal experiences accurately and might reflect observer biases.
• The sample included high-risk, economically disadvantaged youth and the results might not generalize to other groups of children.
• The control group might have had unreported childhood maltreatment. 

Discussion

Findings suggest that early childhood maltreatment produces less adaptive childhood personality functioning, marked by a diminished ability to control emotional impulses and relatively inflexible emotional coping strategies. These personality changes, in turn, put children at a higher risk for externalizing symptoms and eventual substance abuse and dependence. These results shed light on the developmental pathways linking childhood maltreatment to subsequent substance abuse and dependence. As a result, they might be useful in the development of drug prevention and treatment programs.

- Heather Gray & Daniel Tao

What do you think? Please use the comment link below to provide feedback on this article.

References

Oshri, A., Rogosch, F. A., Burnette, M. L., & Cicchetti, D. (2011). Developmental pathways to adolescent cannabis abuse and dependence: Child mistreatment, emerging personality, and internalizing versus externalizing psychopathology. Psychology of Addictive Behaviors, online first.

 

STASH, Vol. 7(4) - Sex, Drugs and...Increased Stress: Perceived stress and risky sexual behaviors among young women who have used ecstasy

Ecstasy use and misuse among young adults is an important public health concern.  A recent report from the Substance Abuse and Mental Health Services Administration’s (SAMHSA) National Survey on Drug Use and Health (NSDUH, 2009) found an increase in both the incidence rate of new ecstasy users (i.e., 1.1 million new users during 2009 compared to 0.6 million during 2004, a recent low year) and the number of ecstasy-related emergency department visits (i.e., 22,816 visits, nearly double those recorded just two years earlier) (DAWN, 2007-2009), and it predicts a resurgence in ecstasy use for the years to come (CESAR FAX, 2011).  However alarming, these data do not reveal the underlying reasons for ecstasy use. To develop effective interventions, researchers need to understand what factors influence young adults to use and misuse ecstasy. Recently, Wu and colleagues (2011) explored the hypothesis that ecstasy use among young, low-income women is associated with both increased stress levels and with increased risky sexual behaviors as compared to other types of illicit drug use or no drug use. This week’s STASH reviews their findings.

Method

• Participants were 407 low-income women (M age = 24.6 years; SD = 3.6; 30 % Black, 34 % White, and 36 % Hispanic) who had made a gynecological appointment at a community-based family-planning clinic in Galveston, Texas.  Women were eligible if they were between 18 and 31 years old and not currently pregnant or post-partum.  Of the 443 women who agreed to participate, 46 were excluded because they did not respond to survey questions.
• Participants completed an in-person survey on the day of their clinic appointment, which included the following measures:
  • Self-reported substance use: Participants reported their lifetime use of the following three categories of substances: (1) club drugs (e.g., ecstasy, PMA); (2) other illicit drugs (e.g., marijuana, cocaine); and (3) licit substances (e.g., tobacco and alcohol).
  • Self-perceived stress: Participants completed Cohen’s 10-Item Perceived Stress Scale (PSS-10) (Cohen, Mermelstein, Kamarck, & Hoberman, 1985). Researchers calculated a summary score ranging from 0 (low stress) to 40 (high stress).
  • Risky sexual behaviors: Participants indicated their number of sexual partners, number of sexually transmitted infections (i.e., STI), consistency of contraceptive use, and engagement in other risky sexual behaviors (e.g., having sex with a stranger).
• The researchers formed three user groups based on participants’ lifetime drug use: (1) ecstasy and other illicit drug-users (including marijuana); (2) exclusive marijuana users; and (3) other illicit drug users excluding ecstasy.

Results

• As Figure 1 illustrates, women in all three drug user groups reported having significantly more stress than non-users (all t’s > 3.02, all p’s < .003).¹  However, there were no significant stress differences among the drug user groups (all t’s < .87, all p’s > .38). 
• Similarly, as Figure 2 illustrates, all drug user groups reported having significantly more lifetime sexual partners than non-users (all t’s > 4.32, all p’s < .001.). However, ecstasy users also had a significantly higher number of sexual partners than either marijuana users (t (142) = 3.57, p < .001) or other illicit drug users (t (88) = 2.49, p < .01).
• Additionally, compared to non-users, ecstasy users were more likely to have (1) contracted a lifetime STI (χ² (1) = 17.0, p < .001), (2) used contraceptives inconsistently (χ² (1) = 4.80, p < .03), and (3) experienced another risky sexual behavior (χ² (1) = 3.35, p < .067). Users of other illicit drugs also surpassed non-users on each of these measures (all χ² (1) > 6.11, all p’s < .014), but marijuana users only surpassed non-users on the STI contraction measure (X² (1) = 10.86, p < .001; for the other two measures, all χ² (1) < 1.13, all p’s > .29).

Figure 1:  Average Perceived Stress Score and Variance (+ 1 Standard Deviation) by Drug User Group

 

Figure 2:  Average Lifetime Sexual Partners and Variance (+ 1 Standard Deviation) by Drug User Group

 

Limitations

• This study’s cross-sectional design limits our understanding of causality among the relationships between ecstasy use, stress, and risky sexual activity. Additionally, the authors do not report correlational and mediational analyses, which would indicate the strength and direction of potential relationships among these variables.
• Because this study’s sample consisted entirely of young, low-income females, the findings might not generalize to other populations such as males and higher income groups.
• Absent physiological stress measures as a more objective alternative to self-reports, we cannot be certain about the nature of the self-reported stress.

 Discussion

Women who had used ecstasy during their lifetime reported higher levels of life stress, more lifetime sexual partners, and more risky sexual practices than non-users. Increased self-reported stress was not unique to ecstasy users; indeed, the initiation of substance use in general has been linked to self-medication following stressful or potentially traumatic life events (e.g., Najavits, Weiss, & Shaw, 1997). However, ecstasy users reported having more lifetime sexual partners than women who reported using other drugs. The researchers suggest that the tactile sensitivity and altered emotions (e.g., heightened interpersonal sense of closeness) associated with ecstasy use contributes to promiscuity. Future research, informed by these findings, would appear to have potential for developing effective, targeted interventions.

Note:
1. The researchers do not present these t-test results. We have used raw data provided in the report to generate these straightforward comparisons. For additional analyses and findings, please refer to Wu, Grady, Rosales, and Berenson (2011). 

-Brittany Bannon

What do you think? Please use the comment link below to provide feedback on this article.

References

CESAR FAX. (April 4, 2011). Resurgence in Ecstasy Use on the Horizon? A
Weekly FAX from the Center for Substance Abuse Research (CESAR), 20 (12). <www.cesar.umd.edu>.

Cohen, S., Mermelstein, R., Kamarck, T., & Hoberman, H. M. (1985). Measuring the
functional components of social support.  In I. G. Sarason & B. R. Sarason (Eds.), Social Support: Theory, research and applications (pp. 73-94). The Hague, The Netherlands: Martines Nijhoff.

Najavits, L. M., Weiss, R. D., & Shaw, S. R. (1997). The link between substance abuse and
posttraumatic stress disorder in women. A research review. The American Journal on Addictions, 6, 273-283.

U.S. Department of Health and Human Services Office of Applied Studies. (2009).
2009 SAMHSA National Survey on Drug Use and Health: Detailed Tables. Retrieved 20 April, 2011, from http://oas.samhsa.gov/WebOnly.htm#NSDUHtabs

U.S. Department of Health and Human Services Office of Applied Studies. (2009). Drug
Abuse Warning Network (DAWN) Detailed Tables: National Estimates, Drug-Related Emergency Department Visits for 2004-2009. Retrieved 20 April, 2011, from https://dawninfo.samhsa.gov/data/

Wu, Z. H., Grady, J. J., Rosales, S., & Berenson, A. B. (2011). Ecstasy Use and Its
Correlates Among Young, Low-Income Women. Substance Use and Misuse, 46, 404-410.

STASH, Vol. 7(3) – Can using pot lead to an increased incidence of psychotic experiences

Use of cannabis – the most popular illicit drug in the world - is associated with psychotic disorders (Wittchen et al., 2007). However, it is not clear if using cannabis actually prompts psychosis or vice versa. This week’s STASH reviews a study that examines if cannabis use among adolescents increases the risk for having psychotic experiences 10 years later (Kuepper et al., 2011).

Method

• Adolescents from Germany, aged 14-24 at baseline, participated in the study. The baseline sample was randomly drawn from the population registry office of Munich metro area. Kuepper et al. analyzed data from 1923 participants who had no missing information for three periods of time.
• Researchers collected demographic and psychiatric information at baseline. Researchers assessed cannabis use at the baseline, 3.5 years later (T2), and 8.4 years later (T3) with a dichotomous question: “Have you ever used cannabis five times or more?”
• A clinical psychologist collected data about psychotic experiences at time T2 and T3 during face-to-face interviews using a version of an international diagnostic interview M-CIDI (Wittchen, Lachner, Wunderlich, & Pfister, 1998). If the participant declared occurrence of any of 20 classic psychotic symptoms (e.g., hallucinations, persecution), the presence of psychosis was recorded.

Results

• At baseline, 247 participants (13%) reported cannabis use.
• Table 1 demonstrates that the cannabis exposure at baseline was associated with psychotic experiences at T2 and T3. In addition, the incidence of cannabis use between baseline and T2 significantly increased the incidence of psychotic experience between T2 and T3 (odd ratio =1.9; 95% confidence interval = 1.1 to 3.1; p < .03).  This relationship is independent of age, sex, socioeconomic status, use of other drugs, childhood trauma, and urban/rural environment, and only among people who had no cannabis use at baseline and no lifetime prevalence of any psychotic symptoms at T2. 
• There was a significant correlation between the continuation of cannabis use and risk of persisting psychotic experiences (?2= 16.22, p < .01).
• On the other hand, people who reported experiencing psychotic symptoms at T2 were not more likely to begin using cannabis between T2 and T3 (odds ratio = 0.8; confidence interval = 0.6 to 1.2; p = 0.3). Psychotic experience at T2 or T3 was not associated with cannabis use at earlier stages.

Table 1. Psychotic experiences at times T2 and T3 for individuals exposed and not exposed to the cannabis at baseline and T2. N = 1923

Cannabis use*	Psychotic symptoms at T2		Psychotic symptoms at T3
	Yes	No	Yes	No
Baseline
Yes	81 (4%)	166 (9%)	42 (2%)	205 (11%)
No	355 (18%)	1321 (69%)	189 (10%)	1487 (77%)
T2
Yes	126 (7%)	267 (14%)	69(4%)	324 (17%)
No	310 (16%)	1220 (64%)	162 (8%)	1368 (71%)

Limitations

• This is a correlational study that cannot determine causality, but only the temporal relationship between use of cannabis and psychosis experiences.
• The results are based on self-reported cannabis use and psychotic experiences
• The researchers did not collect information about the family history of psychosis.

Conclusion

The results of this study clarify the temporal relationships between use of cannabis and psychosis experience. These results imply that cannabis use can predict the incidence and persistence of psychotic experiences later, but not vice versa. Psychotic experiences are not uncommon in general population and - when abnormally persistent - they can progress to clinical state. It is unclear, however, if the psychotic experiences associated with cannabis use (rather than, for example, family history) are also associated with psychotic disorders. The mechanism underlying the effect of prolonged cannabis on persistence of psychotic experiences is not clear yet.

-Julia Braverman

What do you think? Please use the comment link below to provide feedback on this article.

References

Kuepper, R., van Os, J., Lieb, R., Wittchen, H. U., Hofler, M., & Henquet, C. (2011). Continued cannabis use and risk of incidence and persistence of psychotic symptoms: 10 year follow-up cohort study. BMJ, 342, 738.

Wittchen, H. U., Lachner, G., Wunderlich, U., & Pfister, H. (1998). Test-retest reliability of the computerized DSM-IV version of the Munich-Composite International Diagnostic Interview (M-CIDI). Social Psychiatry Epidemiology, 33(11), 568-578.