Relation of Mental Health to Alcohol and Substance Use Among Texas College Students Texas Medicine April 2017

Texas Medicine Logo(1)

The Journal — April 2017

Tex Med. 2017;113(4):e1.

By Alan B. Shafer, PhD; Jessica A. Koenig, MD; and Emilie A. Becker, MD 

Drs Shafer and Becker, Texas Department of State Health Services; and Dr Koenig, The University of Texas at Austin Dell Medical School. Send correspondence to Alan Shafer, 909 W 45th St, Bldg 634 Mail Code 2114, Austin, TX 78751; email:

We examined the effect of mental health problems and difficulties on alcohol, tobacco, and other drug (ATOD) use among college students by using the 2013 Texas College Survey of Substance Use (n=11,216), which includes the K6 screening scale for severe mental illness (SMI). Students' K6 scores were used to classify them into 3 groups: those likely to have SMI (9% with scores ≥ 13), those with some mental health problems (36%), and those without mental health issues (55% with scores ≤ 4). Questions regarding ATOD use were analyzed using these 3 groups.  Alcohol use was not significantly associated with K6 scores, although problematic alcohol behaviors as measured by the CAGE test were. Higher cigarette use was significantly associated with higher K6 scores. Finally, both higher marijuana and higher drug use (across 9 other individual drugs) were significantly associated with higher K6 scores. Although higher K6 scores were associated with higher rates of drug use, most students with high K6 scores did not use drugs. However, given the higher level of risk, drug and alcohol interventions should be made available for those students who receive mental health counseling. 


Prevalence studies have found that at least 1 in 5 adults in the general public are affected by mental illness, and 75% of them have symptom onset before age 24 years.1,2 Population prevalence surveys need a reliable and efficient screening tool to assess serious mental illness (SMI). The K6 is a brief test that screens for nonspecific psychological distress that can indicate the likelihood of an SMI, which is defined as at least one 12-month Diagnostic and Statistical Manual disorder other than a substance use disorder, and serious impairment.3 Significant K6 scores are associated with mental illness but are not able to provide a specific diagnosis. Considerable evidence shows that the K6 has been a reliable screening test to assess communities for the prevalence of SMI.4,5 It is used by the Centers for Disease Control and Prevention (CDC) Behavioral Risk Factors Surveillance Survey, the Substance Abuse and Mental Health Services Administration's National Household Survey on Drug Use and Health, and the US National Health Interview Survey, representing samples of approximately 500,000 people each year.3

The K6 may also differentiate serious psychological distress from mild-to-moderate psychological distress and nondistress. Some studies have divided K6 responses into strata and assigned probabilities of SMI based on the score intervals. Specific cut points in the scoring of the K6 have also been used to determine differences in severity. The K6 is a 6-question test with a score from 0 to 24; a K6 score of 13 or more has been shown to be the cut point that would suggest a high likelihood of nonspecific SMI.3-5 A K6 score between 5 and 12 has been used as a marker for moderate mental distress. Although less severe than the K6 scores of 13 or more, persons with moderate mental distress still have more mental health care utilization, impairment, and substance use than persons with no psychological distress.6

As most persons diagnosed with SMI have their first symptoms before age 24 years, college students are a higher risk age group that could be screened and identified for mental health treatment. One study of university students found SMI at rates higher than those in the general population. Students reporting mental health problems had lower academic achievement, even among students with moderate psychological distress.7 Moderate severity mental illness can contribute to disability and lower academic achievement, and puts students at risk for having untreated symptoms worsen and transition to SMI.5

Assessing the mental health needs of university students is important to identify groups that may be at elevated risk to develop SMI. Multiple factors may impact the rate of mental illness among college age students. Biologically, college students are at an age when several mental illnesses first manifest.1 Additionally, college can be the first exposure to a more independent environment with more responsibilities, requiring increased amounts of self-discipline and control. Along with this often comes exposure to alcohol, tobacco, and other drugs (ATOD). A study of childbearing-age women found that more than 50% used alcohol and 4.1% had heavy use. Women with SMI were more likely to have alcohol use, and particularly heavy alcohol use.8

Drug use is also of concern for college students. In adults aged 18-25 years, the likelihood of nonprescription stimulant use increases when SMI is present.9 Cannabis use has also been associated with higher levels of SMI in adults.10 Adults with SMI are more likely to report cannabis use and are more likely to be dependent on or abusing cannabis than are adults who do not have SMI. Furthermore, despite a higher proportion reporting a desire to quit, adults with SMI had less successful quit attempts than did adults without SMI.11

Following a similar trend, a US population study found that more people with SMI were nicotine dependent and had lower quit ratios than those without. Not surprisingly, those with SMI had a greater likelihood of lifetime tobacco use, recent tobacco use (defined as tobacco use within the last 30 days), and daily tobacco use.12

The Texas College Survey of Substance Use13 is a biennial online self-reported survey of alcohol and drug use, mental health status, risk behaviors, and perceived attitudes and beliefs among college students in Texas conducted for public substance abuse treatment and prevention programs. Starting in 2013, the survey is conducted biennially to estimate prevalence and risk for alcohol and substance use among college students. The Texas College Survey also contains the K6 screening scale,4 which is used to estimate mental health problems and difficulties among Texas college students. This study explores the relationship of mental health difficulties with alcohol and substance use. Students' K6 scores were used to examine responses to ATOD questions on the Texas College Survey. Based on previous studies, higher levels of mental health difficulties and problems should be associated with more alcohol, tobacco and drug use. 



The study population consists of undergraduate students between the ages of 18 and 26 years who enrolled in more than 4 hours of classes. In 2014, there were 11,216 valid survey responses; the raw unweighted respondents were 62% female and 37% male. The race and ethnicity of the sample was 64% non-Hispanic white, 23% any race Hispanic, 5.3% non-Hispanic African American, and 5.8% non-Hispanic Asian. The students' average age was 20.8 years (SD = 1.9) with 22% freshmen, 23% sophomores, 26% juniors, and 29% seniors. The types of college consisted of 65% large 4-year schools, 15% small 4-year schools, and 15% 2-year schools. 

Texas College Survey of Substance Use 

The Texas College Survey of Substance Use collects data online from a secure website. Colleges are sampled at the state level. Participating colleges provided students' email addresses. A total of 45 colleges participated, and nearly 320,000 invitations were sent out over the course of about 5 weeks. About 11,300 valid and complete survey responses were received. The survey is divided into thematic sections asking questions about related areas. These sections include student life (eg, living situation and major); alcohol use; use of drugs other than alcohol; prescription drug use; other personal behaviors (drunk driving and sexual behaviors); mental health; campus policies; and demographic information. The primary alcohol and drug questions analyzed concerned actual student use rather than attitudes or availability. 

K6 Screening Scale

The mental health section of the Texas College Survey contains the K6 screening scale,4 a standardized and validated 6-item scale of nonspecific psychological distress developed to screen for the presence of SMI. The K6 was specifically designed to estimate adults with SMI to meet population prevalence requirements for the US federal block grant for states to fund community mental health services for adults with SMI. The test has a high concordance with clinical diagnoses of SMI in general population samples when a diagnostic threshold score is reached on the scale.3 The questions consist of rating the frequency of 6 items over the last 30 days ― nervous, hopeless, restless or fidgety, depressed, feeling that everything is an effort, and feeling worthless ― on a Likert type scale from 0 (none) to 4 (all the time). Many studies3-5 have validated that scores of 13 and greater indicate a relatively high probability of SMI. Standard validation studies have classified respondents with scores of 13–24 as having probable SMI and those with scores of 0–12 as probably not having SMI. For this study, we also defined a second group of students who had levels of general psychological distress above average but below the SMI indicator score of 13 who were more likely to have a mild or moderate mental health (MH) problem. 


K6 Scores and Groups

The K6 survey had a mean of 5.38 (SD = 4.9) with a range from 0 to 24; the median was 4 with an inter-quartile range of 2 to 8. Reliability of the K6 as measured by coefficient alpha was 0.87. 

Scores of 13 and higher on the K6 indicate that the respondent is likely to have an SMI. In this sample, students with scores of 13 or more comprised 9% of the sample. Some studies have divided K6 responses into strata and assigned probabilities of SMI based on the scores. For this study, we also defined a second group of students, the moderate MH problem group, who had levels of psychological distress above average (scoring over the median of 4) but below the SMI indicator score of 13 and who are more likely to have a mild or moderate MH problem. This group of students with moderate MH problems (K6 scores between 5 and 12) constituted 36% of the sample. The remaining students with K6 scores of 0 to 4 were classified as no MH problems and represented 55% of the sample. Descriptive statistics are provided in Table 1.

When the three groups of students were compared on questions similar to the CDC Health Related Quality of Life questions asking how often in the past 30 days they were unable to work or carry out their normal activity due to the mental health difficulties, the SMI group mean (M = 4.3 days, SD = 5.74) was higher than that of the moderate MH problems group (M = 1.1 days, SD = 3.01), which was in turn higher than the no MH problems group (M = 0.37 days, SD = 2.13). Similar results were found for a question asking about the additional number of days that students were unable to complete half of their tasks or activity (SMI = 8.03 days, moderate MH problems = 2.79 days, and no MH problems = 0.81 days).

Alcohol Use

The Texas College Survey contains a variant of the CAGE screening test14 for alcohol abuse. Three of the questions are the same as those on the original CAGE test; however, the first CAGE item, "Have you ever felt you needed to cut down on your drinking?" has been replaced with an item that asks, "Have you ever thought you had a drinking problem?"  Using the 3 original questions and the alternative first question, students' CAGE test scores were computed. A crosstab (Table 2) of CAGE test scores and the three K6 groups revealed significant differences ( [8] = 263, P < .01) with the SMI group having a higher percentage of students scoring at each level above 1 than the moderate MH problems group, which in turn had a higher percentage of students at each level above 1 than the no MH problems group. CAGE test scores ≥ 2 have been used to screen for excessive drinking and alcoholism. A crosstab (far right column of Table 2) of CAGE test scores ≥ 2 and the three K6 groups revealed significant differences ( [2] = 149, P < .01). 

For binge drinking in the past 30 days, defined as 5 or more drinks for males and 4 or more drinks for females, the results were inconclusive (Table 3). When analyzing 6 levels of binge drinking frequency ranging from none to 10 or more times in the last 30 days, we found no significant difference between the SMI groups ( [10] = 16.9, P = 0.07) for males but a significant difference for females ( [10] = 30.4, P < .01).  For females, SMI students and moderate MH students appeared to binge drink slightly more often than did the no MH problem group at higher frequencies of binge drinking; whereas for males, SMI students appear to binge drink slightly less at all levels. Further analyses when dichotomized into any binge drinking in the last 30 days (Table 3, far right column), showed that males ( [2] = 11.5, P < .01) had significantly less binge drinking for the SMI group than for other groups, but females had no significant difference in binge drinking among the groups ( [2] = 5.1, P = .07). 

Although the SMI group (M = 5.98 days, SD =5.91) reported the highest number of days that alcohol was consumed over the last 30 days compared with the moderate MH problems group (M = 5.76 days, SD =5.38) or the no MH problems group (M = 5.56 days, SD =5.30), the differences were not statistically significant (F[2, 6832] = 2.12, P = 0.12). Similarly, the SMI group (M = 3.80 drinks, SD =2.82) reported the highest average number of drinks on each drinking occasion compared with the moderate MH problems group (M = 3.64 drinks, SD =2.68) and the no MH problems group (M = 3.62 drinks, SD =2.89), but these differences were also not statistically significant (F[2, 7099] = 7.97, P = 0.36). 

Tobacco Use

No significant differences were found among the three groups in the rate of use of smokeless tobacco (83% never using) or cigars (70% never using). However, a statistically significant difference ( [8] = 73.9, P < .01) was seen for cigarette use (Table 4), with the SMI group reporting less "never use" and higher "last month use" than did either the moderate MH problems group or the no MH problems group, indicating higher rates of cigarette use with higher levels of mental health difficulties and problems. 

Drug Use

The SMI group (M = 22.21 occasions, SD = 53.43) reported a statistically significant (F[2, 4600] = 5.23, P < .01) higher number of occasions since the start of the academic year in which they used drugs compared with both the moderate MH problems group (M = 14.74 occasions, SD = 40.93) and the no MH problems group (M = 15.73 occasions, SD = 45.45).

Turning to the analysis of individual drugs, marijuana (Table 5) was the most frequently used drug; 60% of all students across all three groups reported never having used marijuana (compared with only 15% who described themselves as alcohol abstainers). For marijuana use, the SMI group reported statistically significant ( [6] = 37.5, P < .01) more frequent marijuana use compared with both the moderate MH problems group and the no MH problems group. 

After marijuana, the next most frequently used drug was stimulants, with 87% of students reporting "never used." These drugs included medications such as Adderall and Ritalin for treatment of attention deficit hyperactivity disorder. A large group of drugs had approximately 10% of students reporting ever used; the drugs included synthetic marijuana, with 89% never used; narcotics other than heroin (primarily prescription drugs), with 89.9% never used; ecstasy, with 90.6% never used; and other psychedelics, with 90.9% never used. Less frequently used drugs were cocaine (92.6% never used) and sedatives (93.7% never used). The least frequently used drugs were inhalants (96.3% never used) and heroin (98.9% never used). In testing these 9 additional individual drugs, differences among the three K6 groups (SMI, moderate MH problems, and no MH problems) were statistically significant for every single drug. The detailed cross tabulations are available from the authors. Typically, the SMI group's last month use and last year use were the highest and their never used, the lowest; the moderate MH problems group was in between; and the no MH problems group's last month use and last year use were the lowest, and their never used was the highest. In all cases, the drug use levels in both MH groups were higher than in the no MH problems group. 

Because use of the 9 additional individual drugs was relatively low, we combined them together for analysis into a single categorical variable of any drug other than marijuana used (Table 6). The SMI group reported statistically significant ( [6] = 86.1, P < .01) more frequent use of any drug other than marijuana compared with both the moderate MH problems group and the no MH problems group.


Several results are notable: first, that while frequency of alcohol use or amounts of drinking were not significantly associated with K6 scores, problematic alcohol-related behaviors as measured by the CAGE test were; second, that higher cigarette use was significantly associated with higher K6 scores; third, that higher marijuana use was associated with higher K6 scores; and, finally, that higher use of other drugs was also significantly associated with higher K6 scores. 

The consistency of the results across all the other drugs both individually and as a group would seem to suggest a relatively strong relationship such that the SMI group of students are much more likely than typical students to use drugs.  However, even with this strong relationship, it should be emphasized that most students (75%) have never used any of these drugs and only 7% report using these other drugs in the last month.  

Overall, unlike many previous studies, the results of this study did not indicate that higher levels of mental health difficulties had statistically higher alcohol use across all students. This was true across 3 sets of questions: binge drinking, average number of days alcohol was consumed, and number of drinks consumed on each occasion. However, students' self-perceptions of drinking problems as measured by the CAGE screening questionnaire indicated that at scores greater than or equal to 2 (the suggested cutoff), students reporting greater mental health difficulties reported higher levels of problem drinking behaviors. One limitation of this study was that the CAGE questionnaire was an approximation, varying by one question, rather than an exact copy of the CAGE. However, the content of the alternative question is sufficiently similar that it is probably unlikely the results would have changed substantially if the original question had been available.    

Previous studies have shown an association between SMI and alcohol use, specifically in women, and although our results showed some trends in female college students that agreed with this, it did not hold true for all college students.  When looking at the CAGE questionnaire, college students with SMI as defined by K6 scores were shown to be screening positive for excessive drinking and potential alcoholism. However, when evaluating binge drinking, no clear evidence suggested SMI was associated with increased binge drinking behavior. Even when students were separated by gender, no clear evidence suggested that females with SMI were more likely to binge drink except in a narrow band of higher-level frequencies. The population in the Texas College Survey has some differences when compared with Tsai,8 who did find a significant association. However, these differences could have accounted for the different outcomes. The Texas College Survey included men and women whereas Tsai included only women. The age range was also narrower in the Texas College Survey (18-26 years) compared with Tsai (18-44 years). It is possible that these demographic differences led to ambiguous results when drinking behavior was analyzed as the average age of students for the Texas College Survey is slightly below the legal drinking age of 21 years.

Consistent with previous tobacco use studies, students in the SMI group were more likely to use cigarettes than students with less severe or no mental health problems. Students with higher levels of mental health difficulties were significantly more likely to have ever used cigarettes and also were more likely to have recent use, defined by use within the last 30 days. This is consistent with Hagman12 who found that people with SMI had higher lifetime and recent (within the last month) cigarette use.

Marijuana use among students reporting more mental health difficulties was significantly higher than among other students, similar to that from the National Survey on Drug Use and Health for general population adults.8 Shi11 found that marijuana use was significantly more likely to occur in adults with depression or serious psychological distress. The current results in college students are consistent with this finding. Other drugs followed a similar pattern. Herman-Stahl9 found that psychological distress was a risk factor associated with nonmedical uses of stimulants, and the results from this study confirm this in college age students. Students with any type of psychological distress, moderate or severe, had significantly more stimulant use than students with no mental health problems. Scott10 also found higher levels of psychological distress in poly-substance users who used ecstasy with either methamphetamine or marijuana. This study further strengthens previous evidence that identified a correlation between substance use and psychological distress, and provides evidence that other drugs not previously noted also have an association with psychological distress among Texas college students.

Some limitations of this study include the self-reporting nature of the Texas College Survey. Self-report can cause recall bias, and under-reporting of substance use and psychological distress could occur. In addition, the K10, which is very similar in nature to the K6, has been shown to screen well for anxiety and depression but has shown deficits when screening for psychosis.15 The results from Scott10 highlight this concern with results that failed to show elevated K10 scores in marijuana and ecstasy users despite higher numbers of self-reported mental health problems.  Missing psychosis symptoms with the K6 potentially underestimates how many students have negative psychological effects associated with substance use.  The retrospective, cross-sectional design of this study also prevents causation to be drawn between mental health problems and substance use. Further studies could potentially address this issue. Finally, although the K6 has shown strong evidence as a screening tool for serious psychological distress, it does not provide a diagnosis, which is necessary to determine appropriate treatment.

Overall, these results indicate that college age students who reported more mental health difficulties and problems used more tobacco and more drugs, and also reported more problematic alcohol behaviors on the CAGE test. Students' exposure to and use of substances could affect mental illness manifestation. Their proximity to university mental health resources and involvement in the college community makes it possible for interventions to be targeted to positively affect their recovery from mental illness. Further exploration of the association between substance use and occurrence or severity of mental illness will help guide future initiatives to mitigate their risk among students. If substance use is found at a higher rate among college and university students with diagnosed mental illness, these groups should be identified so resources can be allocated to help limit risks. Because substance use may contribute to mental illness, limiting exposure to substances would also be a positive initiative to help reduce mental illness and its consequences on Texas campuses.  


  1. Kessler RC, Berglund P, Demler O, Jin R, Merikangas KR, Walters EE. Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatry. 2005;62(6):593-602.
  2. Bagalman E, Napili A. Prevalence of Mental Illness in the United States: Data Sources and Estimates. Washington, DC: Congressional Research Service; 2013. 
  3. Kessler RC, Green JG, Gruber MJ, et al. Screening for serious mental illness in the general population with the K6 screening scale: results from the WHO World Mental Health (WMH) survey initiative. Int J Methods Psychiatr Res. 2010;19(suppl 1):4-22. 
  4. Kessler RC, Andrews G, Colpe LJ, Short screening scales to monitor population prevalences and trends in non-specific psychological distress. Psychol Med. 2002;32(6):959-976. 
  5. Kessler RC, Barker PR, Colpe LJ, et al. Screening for serious mental illness in the general population. Arch Gen Psychiatry. 2003;60(2):184-189.
  6. Prochaska JJ, Sung HY, Max W, Shi Y, Ong M. Validity study of the K6 scale as a measure of moderate mental distress based on mental health treatment need and utilization. Int J Methods Psychiatr Res. 2012;21(2):88-97. 
  7. Stallman H. Psychological distress in university students: a comparison with general population data. Aust Psychol. 2010;45(4):249-257.
  8. Tsai J, Floyd RL, O'Connor MJ, Velasquez MM. Alcohol use and serious psychological distress among women of childbearing age. Addict Behav. 2009;34(2):146-153.  
  9. Herman-Stahl MA, Krebs CP, Kroutil LA, Heller DC. Risk and protective factors for methamphetamine use and nonmedical use of prescription stimulants among young adults aged 18 to 25. Addict Behav. 2007;32(5):1003-1015. 
  10. Scott LA, Roxburgh A, Bruno R, Matthews A, Burns L. The impact of comorbid cannabis and methamphetamine use on mental health among regular ecstasy users. Addict Behav. 2012;37(9):1058-1062. 
  11. Shi Y. At high risk and want to quit: marijuana use among adults with depression or serious psychological distress. Addict Behav. 2014;39(4):761-767. 
  12. Hagman BT, Delnevo CD, Hrywna M, Williams JM. Tobacco use among those with serious psychological distress: results from the national survey of drug use and health, 2002. Addict Behav. 2008;33(4):582-592. 
  13. Public Policy Research Institute. Methodology Report for the 2013 Texas Survey of Substance Use Among College Students. College Station, TX: Texas A&M University; 2013. 
  14. Ewing JA. Detecting alcoholism: the CAGE questionnaire. JAMA. 1984;252(14):1905-1907.
  15. Hides L, Lubman DI, Devlin H, et al. Reliability and validity of the Kessler 10 and Patient Health Questionnaire among injecting drug users. Aust N Z J Psychiatry. 2007;41(2):166-168.

April 2017 Texas Medicine Contents
Texas Medicine Main Page


Last Updated On

April 03, 2017

Originally Published On

March 28, 2017

Related Content

Mental Health | Substance abuse