Across all participants, Alcohol was the substance with Reasons spontaneously mentioned by the greatest percentage of the participants , followed by Marijuana and Other Drugs . Within each Type of substance, “recreation” was given as a reason for use by the highest percentage of participants for Alcohol , Marijuana , and for Other Drugs , and “avoidance of consequences” was given as a reason for non-use by the highest percentage of participants for Alcohol , Marijuana , and Other Drugs .The QIS was an add-on study of a subset of the well-characterized MTA sample of young adults with and without a diagnosis of childhood-onset ADHD. It was designed to explore additional relationships between persistent SU and ADHD that emerged in the prospective LTF phase of the MTA and to generate new hypotheses about SU in adults with a history of ADHD. The main findings related to the four purposes of the study will be summarized here. First, the origins of the QIS sample were described. Subgroups were based on the two factors under investigation — Diagnosis in childhood and Persistence of substance use into adulthood. The unbalanced groups were explained. Even though ADHD represents a small minority of the population , by deliberate oversampling the MTA sample recruited only half as many LNCG cases as ADHD . Thus, based on the design of the MTA, greenhouse benches the available cases for the Diagnostic groups were unbalanced, with more ADHD cases than LNCG cases.
They remained unbalanced in the QIS sample after the selection of 124 ADHD cases and 58 LNCG cases. Also, as expected, a small minority of the cases had Persistent SU . Since these cases were expected to be highly informative about the relationship between ADHD and SU, all were included in the QIS. Despite this relative enrichment by including all of the Persistent SU cases and random selection of only a minority of the available Non-Persistent SU cases, the selected groups remained unbalanced, with 58 in Persistent SU group and 125 in the Non-Persistent SU group. Second, the exploratory analyses of demographic variables known to be associated with SU in the population were described. Since the QIS sample was not intended to be representative of the population, processes that operated to form the subgroups were evaluated to determine if these important demographic factors were confounded with the Diagnosis and Persistence factors. significant differences between the ADHD group and the LNCG were identified for two baseline variables , but no significant differences were documented between the QIS subgroups with Persistent SU and Non-Persistent SU. Thus, the observed difference between the Persistent and Non-Persistent SU groups on the SUQ and M-EFI outcome measures are not likely to be due to these demographic factors. Also, treatment history within the ADHD group was evaluated to determine if use of stimulant medication differed between the subgroups with Persistent SU and Non-Persistent SU. Potential medication effects were evaluated by extensive analyses of the distribution of cases from the 4 groups defined by the assigned treatment conditions of the RCT phase and the distribution of cases from 3 naturalistic subgroups formed by self-selected patterns of extended treatment during the LTF.
Although additional analyses are in progress in the full MTA sample regarding associations over time, the current findings are consistent with previous reports that did not provide statistical support for associations between substance use and either assigned or self-selected treatment with stimulant medication. Third, the analyses of two measures of frequency of substance use revealed remarkably similar findings related to the Diagnosis and Persistence factors. The quantitative SUQ measure and the qualitative M-EFI measure were distinctly different outcome measures of frequency of SU, and the methods of analyses were distinctly different also. Despite these methodological differences, the patterns of significant and non-significant effects of Diagnosis and Persistence were very similar: for both measures of frequency, very few significant differences between the ADHD group and the LNCG were observed, but many of the differences between the Persistent SU and Non-Persistent SU groups were significant. However, the qualitative and quantitative measures of frequency did show differences in the evaluation of Type of substances. For example, the self-reported use of Marijuana on the SUQ was extremely high in the Persistent SU subgroups compared to Alcohol . This should be interpreted in the context of a strong secular trend of increasing marijuana use over the time when the MTA was conducted , which more than doubled from 1989 to 2008 .
Also, in the unconstrained narratives of the M-EFI, the ADHD cases with Persistent SU apparently did not deny or under-estimate substance use compared to the LNCG cases: the percentages of participants with frequent use were significantly higher in the ADHD group than the LNCG for all 3 types of substances , and within the Non-Persistent group, the percentages were higher for the ADHD group than the LNCG for Marijuana and Alcohol . This generates the hypothesis that under-reporting was not manifested in self-reporting of SU. Fourth, the exploratory analysis of Reasons for “use” of substances suggests some insights regarding possible motivations for the use of Marijuana. After strategic selection for Persistent SU, ADHD group and the LNCG did not differ significantly in frequency of use on either the SUQ or on the M-EFI . However, the diagnostic groups did differ significantly on why substances were used: higher percentage of the ADHD group than the LNCG gave “stability” as a Reason for use. These findings build on recent studies indicating individuals with ADHD may have some unique perceptions about marijuana . Also, they are consistent with the findings of the companion paper by Mitchell et al. , who addressed the topic of emotionality and marijuana in the excerpts from the M-EFI and reported that a higher percentage of ADHD than LNCG cases with Persistent SU perceived beneficial effects of marijuana use . Taken as a whole in the QIS, in the unconstrained narratives of the M-EFI, many of the ADHD cases expressed the belief that marijuana provides emotional “stability” and improves symptoms of ADHD . This may reflect perceived self-medication of symptoms of ADHD and/or dysregulated mood.The QIS has several limitations. First, many analyses were presented here. For example, the analyses of 10 Reasons for the 3 types of substances resulted in 30 analyses, with comparisons of main effects based on the two factors of interest shown in Table 7 , as well as for their interaction . The significance level was not adjusted for multiple tests, which was justified by considering these analyses exploratory . Therefore, the significant and non-significant main effects and interactions effects in these exploratory analyses must be interpreted cautiously. The patterns of effects may provide the most relevant findings. It is worth noting that very similar patterns of significant and non-significant effects emerged across multiple analyses, with many significant differences for comparisons of the Persistent SU and Non-Persistent SU group, but few for comparisons of the ADHD group and the LNCG. Second, even though there was relative enrichment of the QIS sample for Persistent SU, the Non-Persistent SU group was larger by a factor of two. Therefore, within the group with Persistent SU, growere equipment the comparison of ADHD and LNCG cases would be based on smaller subgroups than contrasts of the Non-Persistent subgroups.
Also “by design”, the ADHD group and the LNCG were unbalanced, and within the LNCG, the comparison of subgroups with Persistent and Non-Persistent SU cases would be based on smaller subgroups than contrasts of the ADHD subgroups. In the exploratory analyses of QIS, the unbalanced groupsizes for the levels of Diagnosis and Persistence , as well as the relatively small number of cases in some of the subgroups, complicates and limits the statistical power of the comparisons described here. For example, even though the magnitude of the difference between the Persistent and Non-Persistent subgroups similar for the LNCG and the ADHD group , the difference did not reach statistical significance for the LNCG . Due to these issues related to sample sizes, caution is required for interpretation of lack of statistical significance for comparisons of subgroups within the groups with the smaller number of cases. Third, the participants of the QIS were drawn from only 4 of the 7 original MTA sites. Some of the demographic characteristics differed for the participating and non-participating sites, with reduced risk for SU in the participating sites. If the other 3 sites had been included, effects on SU may have been different in geographical locations where risk for SU was higher than in the locations of the QIS sites. Fourth, the subgroups evaluated here represented the total QIS sample described by Weisner et al . However, different criteria could be applied to select subsets of the total QIS sample, as in evaluation of “turning points” by Jensen et al , who selected equal subgroups of n = 20 SU Persisters, n = 20 SU Desisters, and n = 20 Abstainers for comparison, or as in the evaluation of “emotional functioning” by Mitchell et al , who selected n = 67 SU Persisters and n = 25 SU Desisters for comparison. In Appendix A, the characteristics of these different subsets of the sample across the four papers of the Special Issues on the QIS are described and compared. The findings reported here may be different if the same analyses were performed for other subsets of the sample. Fifth, Persistent SU was not evaluated in the cases that did not complete the MTA assessments in adulthood. Therefore, differential retention of the cases with Persistent SU compared to those with Non-Persistent SU could not be determined for the QIS study. If the classification of SU was known for the cases that were not retained, a different pattern of significant and non-significant effects may have been documented in the analyses of the Diagnosis and Persistence factors. Sixth, the evaluation of demographic factors known to be associated with SU did not show the expected differences between the groups with Persistent SU and Non-Persistent SU. Alternatively, this may be due to multiple effects related to the origin and selection of the QIS sample that may have interacted to counteract other demographic effects. The exploratory analyses reported here suggest the hypothesis that the innovative design of the QIS and the strategic recruitment of the participants of the study may have masked some traditional associations with demographic variables.More high school students smoked little cigars and cigarillos than cigarettes in 33 US states in 2015. Concern is growing about co-use of tobacco and marijuana among youth, particularly among African-American youth. In a 2015 survey, for example, one in four Florida high school students reported ever using cigars or cigar wraps to smoke marijuana. 2 One colloquial term for this is a “blunt.” Adolescent cigar smokers were almost ten times more likely than adults to report that their usual brand offers a flavored variety. Since the US ban on flavored cigarettes , the number of unique LCC flavors more than doubled. Anticipating further regulation, the industry increasingly markets flavored LCCs with sensory and other descriptors that are not recognizable tastes. For example, after New York City prohibited the sale of flavored cigars, blueberry and strawberry cigarillos were marketed as blue and pink, but contained the same flavor ingredients as prohibited products. Among the proliferation of such “concept” flavors , anecdotal evidence suggests that references to marijuana are evident. Cigar marketing includes the colloquial term, “blunt”, in brand names and product labels . Other marketing techniques imply that some brands of cigarillos make it easier for users to replace the contents with marijuana. For example, the image of a zipper on the packaging for Splitarillos and claims about “EZ roll” suggest that products are easily manipulated for making blunts. We use the term “marijuana co-marketing” to refer to such tobacco industry marketing that may promote dual use of tobacco and marijuana and concurrent use .In addition to flavoring, low prices for LCCs also likely increase their appeal to youth. In California, 74% of licensed tobacco retailers sold cigarillos for less than $1 in 2013. 11 Before Boston regulated cigar pack size and price in 2012, the median price for a popular brand of grape-flavored cigars was $1.19. 12 In 2012, 78% of US tobacco retailers sold single cigarillos, which suggests that the problem of cheap, combustible tobacco is widespread.