BACKGROUND AND OBJECTIVES:

Despite their medical vulnerability, youth with chronic medical conditions (YCMCs) drink at levels commensurate with healthy youth. However, information about the prevalence of alcohol use among YCMCs who take alcohol-interactive (AI) medications is scant. To address gaps and inform interventions, we quantified simultaneous exposure to alcohol use and AI medications among YCMCs, hypothesizing that AI exposure would be associated with lower alcohol consumption and mediated by perceptions of alcohol-medication interference.

METHODS:

Adolescents with type 1 diabetes, juvenile idiopathic arthritis, moderate persistent asthma, cystic fibrosis, attention-deficit/hyperactivity disorder, or inflammatory bowel disease completed an electronic survey. We measured the prevalence of exposure to AI medications and the associations with past-year alcohol use as well as binge drinking and total consumption volume in the past 3 months using multivariate regression to estimate the odds of alcohol use given AI medication exposure and perceptions of interference.

RESULTS:

Of 396 youth, 86.4% were on AI medications, of whom, 35.4% reported past-year alcohol use (46.3% among those who were not on AI medications). AI medication use was associated with 43% lower odds of past-year alcohol use (adjusted odds ratio: 0.57; 95% confidence interval: 0.39–0.85) and lower total consumption (β = .43; SE = 0.11; P < .001). Perceptions of alcohol-medication interference partially mediated the relationship between AI medication exposure and past-year alcohol use (Sobel test P = .05).

CONCLUSIONS:

Many YCMCs reported using alcohol; however, drinking was less likely among those who were taking AI medications. Perceptions about alcohol-medication interference mediated the association between drinking and AI medication exposure, suggesting the potential salience of interventions that emphasize alcohol-related risks.

What’s Known on This Subject:

Use of alcohol while on alcohol-interactive medications is common among adults despite risks to health, treatment efficacy, and safety. This issue has not been investigated for youth, including those with a chronic illness, although many use alcohol and binge drink.

What This Study Adds:

In this study, we quantify simultaneous exposure to alcohol use and alcohol-interactive medications among youth with a chronic illness and find that it is common. Medication use is protective of drinking, in part, because of awareness of the potential for medication interference and harm.

Although alcohol use poses a significant risk for all adolescents (contributing to the top 4 causes of morbidity and mortality in this age group),1 for the 12% of youth globally and 25% of US youth with a chronic illness,2,3 drinking alcohol may amplify risks for medical complications, treatment nonadherence, and exposure to behaviors and lifestyle factors that undermine health and trigger disease flares.4,8 An emerging literature on the intersection of alcohol use and chronic illness suggests that despite their medical vulnerability, youth with a chronic illness drink at levels commensurate with their healthy peers.4,5 Researchers of clinical epidemiologic studies have found that approximately one-third of high school–aged youth with chronic medical conditions (YCMCs) report past-year drinking, of whom, one-third drink at a binge level.6 Although similar percentages of YCMCs and healthy youth initiate alcohol use in early adolescence, in population-based studies, YCMCs are disproportionately likely to progress to heavy and problem drinking by older adolescence and young adulthood.4 In addition, YCMCs who drink are nearly twice as likely as those who do not drink to report regular treatment nonadherence, including among youth whose treatment regimens require the use of daily life-sustaining medication and careful dosing, as with type 1 diabetes.6 

When drinking, YCMCs risk harm from exposure to prescription and over-the-counter medications (or alcohol-interactive [AI] medications) that can react negatively with alcohol. Alcohol can alter the effects of such medications, creating toxicities and/or reducing efficacy.9,10 Alcohol use can also affect laboratory test values,11 leading to unnecessarily aggressive therapy. For youth who drink and take AI medications, there is potential for enduring damage to the liver, kidneys, and other organs.9 This is especially so if alcohol use follows the heavy and high-volume pattern that is typical of youth.12 Adults exhibit high levels of simultaneous exposure to AI medications and alcohol use; among Americans age ≥20 years, 41.5% report both past-year exposure to alcohol use and AI medications, whereas 77.8% of those age ≥65 years report this dual exposure.13 To date, there are no published reports of the prevalence of exposure to alcohol use and AI medications among adolescents.

We sought to measure the prevalence of and associations between alcohol use and exposure to AI medications among adolescents with a chronic medical condition. In addition, we sought to test whether youth perceptions about alcohol-medication interference mediate observed associations. Such findings would reveal a target for health-protective interventions for the large and growing population of youth who are medically vulnerable and at risk of using alcohol.

As part of a larger survey study to validate a youth alcohol use screening tool,5 we measured the quantity and frequency of alcohol use by a clinical cohort of adolescents with a pediatric onset chronic disease, their use of medications with alcohol use contraindications, and their perceptions about the interference between alcohol and their medications. The study was conducted among youth presenting for routine subspecialty care between June 2013 and July 2015 at clinics affiliated with a large teaching hospital in the Northeastern United States. Some participants were recontacted to participate in qualitative interviews about substance use. Youth were given a waiver of parental consent under the approval of the Boston Children’s Hospital Institutional Review Board. All survey items were administered by using an online structured assessment that was delivered on a tablet computer configured with a polarizing screen for privacy; medication use was assessed via a medical chart review. A detailed report of methods has been published previously.6 

Clinics included those offering subspecialty care for type 1 diabetes, juvenile idiopathic arthritis, moderate persistent asthma or cystic fibrosis, attention-deficit/hyperactivity disorder, or inflammatory bowel disease (including ulcerative colitis or Crohn disease).

We recruited a convenience sample of ∼100 English-speaking patients ages 9 to 18 years from each subspecialty clinic or program (some patients had turned 19 when they were recontacted for qualitative interviews). Eligibility included having the condition for at least 1 year and being able to manipulate a tablet computer.6 In total, 662 youth were approached for this study, of whom, 508 consented (76.7% consent rate). For this analysis, we excluded participants in eighth grade or below (n = 99) because younger adolescents, who are less likely to use alcohol,1 may not understand medication interactions. We further excluded participants (n = 13) with missing survey data for measures of alcohol use behaviors and perceptions of alcohol interfering with medications, leaving a final analytic sample of N = 396.

Demographic Characteristics

Participants reported their age in years, current grade in school, sex, race and/or ethnicity, number of parents or guardians in the household, and highest level of education attained by a parent as a proxy for socioeconomic status. Ten participants were missing data on the number of parents or guardians in the household, and for these participants, the sample mode (≥2 parents) was used. For 9 respondents who were missing data on race and/or ethnicity, the sample mode (white, non-Hispanic) was used, and for 12 participants who were missing data on the level of education attained by a parent, the sample mode (college graduate) was used.

Alcohol Use Behaviors

A battery of questions was used to measure alcohol use, including past-year alcohol use and, for youth reporting past-year alcohol use, binge drinking and total alcohol volume in the past 3 months. Binge drinking was defined by using the National Institute on Alcohol Abuse and Alcoholism age and sex cutoffs14; for example, 14- to 15-year-old boys would be placed in the binge category for consuming ≥4 drinks on 1 occasion. Total alcohol volume in the past 3 months (hereafter referred to as total alcohol volume) was obtained by multiplying respondents’ numeric responses to the following questions: “In the last 3 months, on how many days did you have something to drink?” and “When you drank in the last 3 months, how many drinks did you usually have on 1 occasion?” For total alcohol volume, a winsorized mean (trimming values above the 99th percentile ) was used. Binge drinking and total volume measures are reported only for participants who reported past-year drinking and for whom complete data on these measures are available (N = 131).

Exposure to AI Medications

Exposure to AI medications was determined via electronic medical record review. For each participant, trained research assistants reviewed the patient chart for the study enrollment visit and recorded all prescription and over-the-counter medications that were listed. Medications were coded as AI or non-AI by using published contraindications.15 Examples of AI medications that were used by this cohort include methotrexate, insulin, and acetaminophen. Interactions with alcohol ranged from major (eg, acetaminophen) to moderate (eg, insulin, methotrexate). Confirmatory coding of AI or non-AI for each medication was checked by a pediatrician.

Perceptions About Alcohol Interfering With Medications

To assess participants’ perceptions of alcohol interfering with their medications, participants were asked the following question: “Can alcohol interfere with or get in the way of any of the medications you take?” Response options were “yes,” “no,” or “I don’t know.”

Statistical Analyses

Analyses were conducted by using SAS 9.4 software (SAS Institute, Inc, Cary, NC). Statistical significance was considered at P < .05. Summary statistics were computed to characterize the study sample overall and by exposure to AI medications. Differences in demographics and alcohol use behaviors by AI medication exposure were compared by using Wilcoxon rank-sum, Kruskal-Wallis, or χ2 tests, as appropriate. Generalized estimating equations (GEEs) were used to examine associations between AI medication exposure and alcohol use behaviors, accounting for clustering within subspecialty clinics. Past-year alcohol use and binge drinking were modeled by using GEEs that followed a binary distribution (logit link function) and total alcohol volume was modeled by using GEEs that followed a Poisson distribution. Multinomial mixed models were used to examine the association between AI medication exposure and perceptions of alcohol-medication interference, controlling for random effects within subspecialty clinics. All multivariate models were adjusted for age, sex, race and/or ethnicity, parental educational attainment, and number of parents living in the home. Mediation of alcohol use behaviors by AI medication exposure was evaluated with stepwise regression and application of the Sobel test.16 First, models that adjusted for age, sex, race and/or ethnicity, parental education, number of parents in the home, and AI medication exposure were run (model 1) followed by models that additionally adjusted for perceptions about alcohol interfering with medications (model 2). The Sobel test was used to evaluate if and to what extent perceptions of alcohol-medication interference mediated the relationship between AI exposure and alcohol use behaviors.16 

The total sample was predominantly non-Hispanic white (74.7%), and the majority had at least 1 parent with a college degree or higher (74.5%). On average, the sample was 16.4 years of age (range: 13–18; SD = 1.3), with the majority of participants in high school (84.3%) and a smaller proportion (15.7%) in college. The majority of the sample (86.4%) was on AI medications, with AI medication exposure differing by subspecialty clinic type (P < .001; Table 1).

TABLE 1

Demographic Characteristics of the Sample in Aggregate and by Exposure to AI Medications

VariableTotalOn AI MedicationNot on AI MedicationP
Total, N (%) 396 (100) 342 (86.4) 54 (13.6) — 
Age, y, mean (SD) 16.4 (1.3) 16.4 (1.3) 16.4 (1.3) .82 
Grade, n (%)    .24 
 9th 55 (13.9) 44 (12.9) 11 (20.4) — 
 10th 84 (21.2) 77 (22.5) 7 (13.0) — 
 11th 102 (25.8) 87 (25.4) 15 (27.8) — 
 12th 93 (23.5) 83 (24.3) 10 (18.5) — 
 College 62 (15.7) 51 (14.9) 11 (20.4) — 
Sex, n (%)    .68 
 Female 201 (50.8) 175 (51.2) 26 (48.1) — 
 Male 195 (49.2) 167 (48.8) 28 (51.9) — 
Race and/or ethnicity, n (%)    .14 
 White, non-Hispanic 296 (74.7) 260 (76.0) 36 (66.7) — 
 Other race and/or ethnicity 100 (25.3) 82 (24.0) 18 (33.3) — 
Parental education, n (%)    .16 
 Less than a college degree 101 (25.5) 83 (24.3) 18 (33.3) — 
 College graduate 295 (74.5) 259 (75.7) 36 (66.7) — 
Parents living in home, n (%)    .37 
 <2 or foster care 98 (24.7) 82 (24.0) 16 (29.6) — 
 ≥2 298 (75.3) 260 (76.0) 38 (70.4) — 
Clinic or disease group, n (%)    <.001 
 Pulmonary medicine 81 (20.5) 63 (18.4) 18 (33.3) — 
 Endocrinology 79 (19.9) 79 (23.1) 0 (0.0) — 
 Gastroenterology 84 (21.2) 64 (18.7) 20 (37.0) — 
 Rheumatology 81 (20.5) 69 (20.2) 12 (22.2) — 
 Developmental medicine 71 (17.9) 67 (19.6) 4 (7.4) — 
VariableTotalOn AI MedicationNot on AI MedicationP
Total, N (%) 396 (100) 342 (86.4) 54 (13.6) — 
Age, y, mean (SD) 16.4 (1.3) 16.4 (1.3) 16.4 (1.3) .82 
Grade, n (%)    .24 
 9th 55 (13.9) 44 (12.9) 11 (20.4) — 
 10th 84 (21.2) 77 (22.5) 7 (13.0) — 
 11th 102 (25.8) 87 (25.4) 15 (27.8) — 
 12th 93 (23.5) 83 (24.3) 10 (18.5) — 
 College 62 (15.7) 51 (14.9) 11 (20.4) — 
Sex, n (%)    .68 
 Female 201 (50.8) 175 (51.2) 26 (48.1) — 
 Male 195 (49.2) 167 (48.8) 28 (51.9) — 
Race and/or ethnicity, n (%)    .14 
 White, non-Hispanic 296 (74.7) 260 (76.0) 36 (66.7) — 
 Other race and/or ethnicity 100 (25.3) 82 (24.0) 18 (33.3) — 
Parental education, n (%)    .16 
 Less than a college degree 101 (25.5) 83 (24.3) 18 (33.3) — 
 College graduate 295 (74.5) 259 (75.7) 36 (66.7) — 
Parents living in home, n (%)    .37 
 <2 or foster care 98 (24.7) 82 (24.0) 16 (29.6) — 
 ≥2 298 (75.3) 260 (76.0) 38 (70.4) — 
Clinic or disease group, n (%)    <.001 
 Pulmonary medicine 81 (20.5) 63 (18.4) 18 (33.3) — 
 Endocrinology 79 (19.9) 79 (23.1) 0 (0.0) — 
 Gastroenterology 84 (21.2) 64 (18.7) 20 (37.0) — 
 Rheumatology 81 (20.5) 69 (20.2) 12 (22.2) — 
 Developmental medicine 71 (17.9) 67 (19.6) 4 (7.4) — 

Row percentages are shown in the ‘Total’ column, column percentages are displayed elsewhere. Because of rounding, the percentages presented above may not add up exactly to 100%. P values were derived from the χ2 or Wilcoxon rank test. —, not applicable.

Overall, more than one-third (36.9%) of participants reported past-year alcohol use, of whom, 38.2% reported binge drinking (Table 2). Among those on AI medications, 35.4% reported past-year alcohol use, of whom, 35.1% reported binge drinking. Past-year drinkers who were on AI medications reported an average of 10.8 drinks consumed during the previous 3 months compared with an average of 16.9 drinks reported by those who were not on AI medications (Table 2).

TABLE 2

Associations Between Alcohol Use Behaviors, AI Medication Exposure, and Perceptions of Alcohol-Medication Interference

VariableTotalAI Medication ExposurePCan Alcohol Interfere With or Get in the Way of Any of the Medications You Take?P
On AI MedicationNot on AI MedicationNoI Don’t KnowYes
Total, N (%) 396 (100) 342 (86.4) 54 (13.6) — 29 (7.3) 176 (44.4) 191 (48.2) — 
Past-year alcohol use, n (%)    .12    <.001 
 Yes 146 (36.9) 121 (35.4) 25 (46.3) — 24 (82.8) 53 (30.1) 69 (36.1) — 
 No 250 (63.1) 221 (64.6) 29 (53.7) — 5 (17.2) 123 (69.9) 122 (63.9) — 
Binge drinking in the past 3 mo, n (%)a    .09    .04 
 Yes 50 (38.2) 39 (35.1) 11 (55.0) — 13 (56.5) 12 (26.1) 25 (40.3) — 
 No 81 (61.8) 72 (64.9) 9 (45.0) — 10 (43.5) 34 (73.9) 37 (59.7) — 
Total alcohol volume in the past 3 mo, mean (SD)a 11.7 (21.7) 10.8 (20.5) 16.9 (27.5) .25 24.3 (34.5) 7.3 (12.1) 10.4 (19.7) .02 
AI medication exposure, n (%)    —    .008 
 On AI medication 342 (86.4) — — — 21 (72.4) 147 (83.5) 174 (91.1) — 
 Not on AI medication 54 (13.6) — — — 8 (27.6) 29 (16.5) 17 (8.9) — 
VariableTotalAI Medication ExposurePCan Alcohol Interfere With or Get in the Way of Any of the Medications You Take?P
On AI MedicationNot on AI MedicationNoI Don’t KnowYes
Total, N (%) 396 (100) 342 (86.4) 54 (13.6) — 29 (7.3) 176 (44.4) 191 (48.2) — 
Past-year alcohol use, n (%)    .12    <.001 
 Yes 146 (36.9) 121 (35.4) 25 (46.3) — 24 (82.8) 53 (30.1) 69 (36.1) — 
 No 250 (63.1) 221 (64.6) 29 (53.7) — 5 (17.2) 123 (69.9) 122 (63.9) — 
Binge drinking in the past 3 mo, n (%)a    .09    .04 
 Yes 50 (38.2) 39 (35.1) 11 (55.0) — 13 (56.5) 12 (26.1) 25 (40.3) — 
 No 81 (61.8) 72 (64.9) 9 (45.0) — 10 (43.5) 34 (73.9) 37 (59.7) — 
Total alcohol volume in the past 3 mo, mean (SD)a 11.7 (21.7) 10.8 (20.5) 16.9 (27.5) .25 24.3 (34.5) 7.3 (12.1) 10.4 (19.7) .02 
AI medication exposure, n (%)    —    .008 
 On AI medication 342 (86.4) — — — 21 (72.4) 147 (83.5) 174 (91.1) — 
 Not on AI medication 54 (13.6) — — — 8 (27.6) 29 (16.5) 17 (8.9) — 

Because of rounding, the percentages presented above may not add up exactly to 100%. P values were derived from the χ2, Wilcoxon rank, or Kruskal-Wallis test. —, not applicable.

a

Among past-year drinkers with complete data on binge drinking and total volume (N = 131).

Nearly half of the sample (48.2%) endorsed that alcohol can interfere with their medications, whereas 44.4% were unsure of such interference, and 7.3% believed that alcohol does not interfere with their medications (Table 2). Past-year alcohol use was most prevalent among youth who believed that alcohol did not interfere with their medication (82.8%), significantly higher than among youth who endorsed interference (36.1%) and those who were unsure of the interference (30.1%; P < .001). On average, youth who reported believing that alcohol does not interfere with their medications also reported consuming more alcohol in the past 3 months (the average number of drinks consumed in the past 3 months was 24.3 for youth who answered, “No,” 7.3 for youth who answered, “I don’t know,” and 10.4 for youth who answered, “Yes”; P = .02). In adjusted analyses, compared with youth who were not on AI medications, those who were taking AI medications were less likely to report believing that alcohol does not interfere with their medications (adjusted odds ratio [aOR]: 0.28; 95% confidence interval [CI]: 0.10–0.76) or report that they were unsure about the interference of alcohol (aOR: 0.49; 95% CI: 0.25–0.96; Supplemental Table 5).

Youth who were on AI medications had 43% lower odds of reporting that they consumed alcohol in the past year compared with youth who were not taking AI medications (aOR: 0.57; 95% CI: 0.39–0.85 for model 1; Table 3), and this association was significantly attenuated after adjusting for perceptions of alcohol-medication interference (Sobel z score = −1.95; P = .05). Among youth who drank in the past year, those who were on AI medications consumed fewer drinks in the past 3 months than those who were not on AI medications (P < .001 for model 1; Table 3); this association was partially attenuated by perceptions of alcohol-medication interference (Sobel z score = −1.83; P = .07). This finding is bolstered by qualitative data that reveal how perceptions of alcohol-medication interference affect decisions about health behaviors (Table 4).

TABLE 3

Associations Between AI Medication Use and Perceptions and Alcohol Use Behaviors

VariablePast-Year Alcohol UseBinge Drinking in the Past 3 mo Among Past-Year DrinkersaTotal Volume in the Past 3 mo Among Past-Year Drinkers, β (SE)a
Model 1, AI medication use, OR (95% CI)    
 On AI medication 0.57 (0.39–0.85) 0.36 (0.10–1.28) −.43 (0.11)b 
 Not on AI medication 1.00 (Reference) 1.00 (Reference) .00 (Reference) 
Model 2, AI medication exposure, OR (95% CI)    
 On AI medication 0.66 (0.36–1.19) 0.37 (0.09–1.53) −.29 (0.14)c 
 Not on AI medication 1.00 (Reference) 1.00 (Reference) .00 (Reference) 
Model 2, perceptions of alcohol-medication interference, OR (95% CI)    
 No (does not interfere with medications) 6.74 (2.05–22.14) 1.89 (0.87–4.12) .73 (0.27)c 
 I don’t know 0.96 (0.61–1.52) 0.58 (0.25–1.37) −.20 (0.14) 
 Yes (does interfere with medications) 1.00 (Reference) 1.00 (Reference) .00 (Reference) 
Sobel test    
 Change in β, % −23.8 –1.5 −33.0 
z scored −1.95 –1.35 −1.83 
Pe .05 .18 .07 
VariablePast-Year Alcohol UseBinge Drinking in the Past 3 mo Among Past-Year DrinkersaTotal Volume in the Past 3 mo Among Past-Year Drinkers, β (SE)a
Model 1, AI medication use, OR (95% CI)    
 On AI medication 0.57 (0.39–0.85) 0.36 (0.10–1.28) −.43 (0.11)b 
 Not on AI medication 1.00 (Reference) 1.00 (Reference) .00 (Reference) 
Model 2, AI medication exposure, OR (95% CI)    
 On AI medication 0.66 (0.36–1.19) 0.37 (0.09–1.53) −.29 (0.14)c 
 Not on AI medication 1.00 (Reference) 1.00 (Reference) .00 (Reference) 
Model 2, perceptions of alcohol-medication interference, OR (95% CI)    
 No (does not interfere with medications) 6.74 (2.05–22.14) 1.89 (0.87–4.12) .73 (0.27)c 
 I don’t know 0.96 (0.61–1.52) 0.58 (0.25–1.37) −.20 (0.14) 
 Yes (does interfere with medications) 1.00 (Reference) 1.00 (Reference) .00 (Reference) 
Sobel test    
 Change in β, % −23.8 –1.5 −33.0 
z scored −1.95 –1.35 −1.83 
Pe .05 .18 .07 

Model 1 was adjusted for AI medication status, age, sex, race and/or ethnicity, parental education, and number of parents in the home while accounting for clustering within clinics. Additionally, model 2 was adjusted for perceptions of alcohol-medication interference. OR, odds ratio.

a

Among past-year drinkers with complete data on binge drinking and total volume (N = 131).

b

β coefficient is statistically significant at P < .001.

c

β coefficient is statistically significant at P < .05.

d

z score for Sobel’s test of mediation.

e

P value for test of mediation.

TABLE 4

Select Illustrative Quotes From YCMCs About Alcohol-Medication Interference Shared During Qualitative Interviews About Alcohol Use

Quote
Avoidance of alcohol to remain on medications “Well, I wouldn’t [drink on my medication]. I would be concerned if I was drinking and I had my pill. …I don’t think it would go well…based on the side effects I’ve had and before. I would get headaches before. …And just to think that if I was drinking alcohol and drank while doing that, it would be horrible.” (17-year-old female patient with ADHD) 
“I Googled if there were any super adverse side effects [from drinking on my medication] because I remember my doctor, he told me 1 or 2 drinks should be fine…but just to make sure because obviously I am not looking to mess anything up.” (19-year-old female patient with Crohn disease) 
“I do not want it [alcohol] to worsen my arthritis or make my medication not work. I mostly worry about the long-term damage to my liver or kidney; medication is already rough on the organs, and I know alcohol can be as well.” (18-year-old female patient with arthritis) 
“I read, like, before my bottles, like, of medicine. It’s like, ‘Do not take with alcohol.’ And I just, like, went online, if it’s, like, good to drink alcohol with medicine, and basically it just said, like, ‘No, don’t because it won’t work,’ like, the medicine.” (16-year-old female patient with polymyositis) 
“I know I can’t really skip [my medication] because then my knees are going to hurt really bad. So I guess if I ever drink in the future, I’ll just barely have any because I’m afraid that it’s going to, like, cancel it out or something.” (16-year-old female patient with JIA) 
“And if you’re really intoxicated, your body can’t function and release the sugar at the same time. And so even if you have a good glucagon, which is—it’s, like, this really long needle that you push all this liquid into, like a little bottle, and there’s a tablet in it that dissolves—I don’t know what the tablet is, it might just be pure sugar—and then you draw it all out and then you put it in someone’s thigh …So it’s for use in emergency situations, but, like, if you’re really drunk it doesn’t work because your liver can’t process it; so if you’re diabetic and you go low you can be kind of, you know, in a bad situation.” (18-year-old female patient with type 1 diabetes) 
Avoidance of medications to drink alcohol “In all honesty, if I do, like, drink a lot that night I honestly won’t even take my medications sometimes just ’cause I’m afraid that, like, it won’t work or something will, like, react bad.” (19-year-old female patient with IBD-associated arthritis) 
“[My rheumatologist] brought me off the medication because he asked if I was going to continue drinking or not, and I said that I probably would and, therefore, I can no longer be on the medication.” (19-year-old male patient with arthritis) 
“I don’t think I’m supposed to drink with my medications…the one that’s absorbed by my liver, anyway. But sometimes, I’ll just skip it that day and the next day and just make sure that my liver doesn’t shut down. I don’t really know how that works.” (18-year-old female patient with arthritis) 
Quote
Avoidance of alcohol to remain on medications “Well, I wouldn’t [drink on my medication]. I would be concerned if I was drinking and I had my pill. …I don’t think it would go well…based on the side effects I’ve had and before. I would get headaches before. …And just to think that if I was drinking alcohol and drank while doing that, it would be horrible.” (17-year-old female patient with ADHD) 
“I Googled if there were any super adverse side effects [from drinking on my medication] because I remember my doctor, he told me 1 or 2 drinks should be fine…but just to make sure because obviously I am not looking to mess anything up.” (19-year-old female patient with Crohn disease) 
“I do not want it [alcohol] to worsen my arthritis or make my medication not work. I mostly worry about the long-term damage to my liver or kidney; medication is already rough on the organs, and I know alcohol can be as well.” (18-year-old female patient with arthritis) 
“I read, like, before my bottles, like, of medicine. It’s like, ‘Do not take with alcohol.’ And I just, like, went online, if it’s, like, good to drink alcohol with medicine, and basically it just said, like, ‘No, don’t because it won’t work,’ like, the medicine.” (16-year-old female patient with polymyositis) 
“I know I can’t really skip [my medication] because then my knees are going to hurt really bad. So I guess if I ever drink in the future, I’ll just barely have any because I’m afraid that it’s going to, like, cancel it out or something.” (16-year-old female patient with JIA) 
“And if you’re really intoxicated, your body can’t function and release the sugar at the same time. And so even if you have a good glucagon, which is—it’s, like, this really long needle that you push all this liquid into, like a little bottle, and there’s a tablet in it that dissolves—I don’t know what the tablet is, it might just be pure sugar—and then you draw it all out and then you put it in someone’s thigh …So it’s for use in emergency situations, but, like, if you’re really drunk it doesn’t work because your liver can’t process it; so if you’re diabetic and you go low you can be kind of, you know, in a bad situation.” (18-year-old female patient with type 1 diabetes) 
Avoidance of medications to drink alcohol “In all honesty, if I do, like, drink a lot that night I honestly won’t even take my medications sometimes just ’cause I’m afraid that, like, it won’t work or something will, like, react bad.” (19-year-old female patient with IBD-associated arthritis) 
“[My rheumatologist] brought me off the medication because he asked if I was going to continue drinking or not, and I said that I probably would and, therefore, I can no longer be on the medication.” (19-year-old male patient with arthritis) 
“I don’t think I’m supposed to drink with my medications…the one that’s absorbed by my liver, anyway. But sometimes, I’ll just skip it that day and the next day and just make sure that my liver doesn’t shut down. I don’t really know how that works.” (18-year-old female patient with arthritis) 

ADHD, attention-deficit/hyperactivity disorder; IBD, inflammatory bowel disease; JIA, juvenile idiopathic arthritis.

Alcohol use was prevalent among a cohort of adolescents who were medically vulnerable and in care for a chronic illness, including among a portion of the cohort taking AI medications. This is worrisome given the potential for alcohol use to adversely impact the efficacy and safety of treatments, the accuracy of tests to monitor disease, and, more generally, undermine self-care and health. Nevertheless, the prevalence of alcohol use was lower among youth who were taking AI medications compared with that measured among their peers who were not taking AI medications, consistent with findings from studies with older adults.13 The protective effect of being on an AI medication was apparent for the odds of past-year alcohol use and with respect to the quantity of alcohol consumed among youth who reported that they drink; among drinkers, those who were on AI medications consumed less alcohol (measured as total alcohol volume in the past 3 months) than their peers who were not taking AI medications.

Our results suggest that among youth, the protective association between being on an AI medication and alcohol use may operate through knowledge, awareness, or concern about the potential for alcohol to interfere with the efficacy or safety of medications that are being used to treat a chronic condition. Awareness of alcohol’s potential to interfere with medications or uncertainty about this (not medication use alone) may support health-protecting behaviors and choices, as evidenced in narrative data. This finding may be used to inform the design of interventions used to target reduced drinking among the large percentage of youth who are growing up with a chronic disease; for them, interventions may be impactful if they are centered on explaining the potential for reduced treatment efficacy and harm related to drinking when taking medications with alcohol use contraindications.

Findings reveal the strong health-promoting potential of delivering to medically vulnerable youth tailored and targeted health guidance about alcohol-related risks. This is good news and aligns with the American Academy of Pediatrics recommendations that screening and brief intervention, including personalized brief advice on the impact of substance use on health, be incorporated into routine health care for all youth.17 Importantly, brief alcohol use screening tools have high validity for identifying the risk of alcohol use among youth with chronic conditions.5 Screening, coupled with brief interventions, has revealed promise for reducing alcohol use among adults18,19 and may be a highly efficient and impactful intervention strategy for youth in care for a chronic illness. These youth have strong ties to their medical providers,20 regular health care encounters,21 and a strong motivation to protect their health.22 Findings augur well for the potential that careful health messaging may drive protective interventions. Notably, youth with chronic illness who drink are more likely than their peers who do not drink to report treatment nonadherence.6 It is likely that for some youth, nonadherence reflects an attempt to reduce the risk of alcohol-medication interactions, which is also evident in qualitative data. Here, the development of nuanced interventions and messages that direct youth toward the preservation of treatment adherence and reduction of alcohol use may be merited.

Importantly, nearly half of the study sample reported not knowing if alcohol could interfere with their medications. Although poorly informed, this group was not at an elevated risk for drinking; uncertainty may be protective and possibly moderated by risk tolerance. In contrast, a small fraction of the sample did not perceive that alcohol could interfere with their medications, and the odds of drinking and of a high total volume of alcohol consumption were inordinately high for this group. This especially high-risk group may benefit from additional interventions that are focused specifically on improving their understanding of alcohol-related health risks in the setting of their chronic disease and its treatment.

Findings of the protective effect of knowledge or beliefs on drinking behaviors for youth who are medically vulnerable contrast with results of studies that were undertaken with healthy youth whose drinking behaviors were powerfully influenced by social and environmental factors (including patterns of alcohol accessibility, availability, and promotion), policy and parental controls,1,23,24 and less clearly influenced by knowledge.25,26 For healthy youth, the pull of social factors and the push of alcohol marketing and promotion may outweigh concern for health harm. They may consider health risks that stem from alcohol use to be abstract, temporally remote, and not applicable to them.27,28 In contrast, drinking-related decisions for YCMCs may be filtered through real experiences with having a disease, managing symptoms, and engaging with treatment. Youth may be primed to attend to messages not to drink that are framed in patient-centered concerns to avoid flares and complications and stay attuned to sensorial cues of disease activity.22 Although YCMCs are not immune to social pressures to drink, their decisions about drinking may be balanced by consideration of the real potential for significant near-term health harm from drinking, whether harms pertain to all youth (eg, accident or injury) or to those who are medically vulnerable (eg, toxicity or disease exacerbation). In this context, delivering health messages that speak to the risks for alcohol-medication interactions and the negative impacts of drinking on treatment efficacy may comprise important and impactful themes for preventive interventions.

This report is the first investigation of the exposure to AI medications and alcohol use behaviors among YCMCs. Strengths include the use of a large medically heterogeneous sample with a confirmed clinical diagnosis and the high study participation rate. We used validated screening measures to assess alcohol use behaviors and rigorous chart review with clinical confirmation to determine AI medication status. Nevertheless, use of a cross-sectional study design precludes understanding about temporal ordering or causality among exposure and behaviors. There could be a temporal offset between reports about drinking (which reflect patterns in the past year and the past 3 months) and AI medication use (assessed for the time period coterminous to the study visit). Because all participants were recruited on the basis of a disease duration of at least 1 year, we think that misclassification of AI medication use is likely small for the period of time about which drinking is reported.6 It is possible that participants who were not taking AI medications were generally healthier and more able to participate in social activities that included alcohol consumption, and this may partially account for the findings. Although the participation rate for this study was high, findings may not be generalized to all youth with chronic illness, including those who were diagnosed with conditions not included in the current study. A chart review may not capture the use of over-the-counter medications with alcohol use contraindications. Self-reports of alcohol use are subject to recall and reporting bias, although general prevalence levels are consistent with national survey self-reports.1,29 

Large percentages of adolescents in the United States and globally are growing up with a chronic medical condition,2,3 and interventions to protect their health are anticipated to yield considerable return on investment, with health care resources saved and suffering averted. Alcohol screening and prevention programs that are used to target YCMCs are needed in health care settings. Patient-centered preventive interventions and safety campaigns that are tailored to adolescents and used to focus attention on potential harms from alcohol use, including while taking AI medications, could ameliorate this safety issue for youth who are medically vulnerable. Ongoing research is warranted to understand this group’s alcohol-related knowledge, attitudes, and behaviors and to develop and test strategies for fostering sustained health-protecting behaviors to support their health.

     
  • AI

    alcohol-interactive

  •  
  • aOR

    adjusted odds ratio

  •  
  • CI

    confidence interval

  •  
  • GEE

    generalized estimating equation

  •  
  • YCMC

    youth with chronic medical conditions

Drs Weitzman and Levy conceptualized and designed the study, designed data collection measures and protocols, supervised data collections, directed data analyses, drafted the initial manuscript, and reviewed and revised the manuscript; Ms Magane assisted with data analyses, drafted the initial manuscript, and reviewed and revised the manuscript; Dr Wisk directed data analyses and assisted with drafting and revising the manuscript; Mr Allario conducted data analyses and reviewed and revised the manuscript; Dr Harstad collected data and critically reviewed the manuscript; and all authors approved the final manuscript as submitted.

FUNDING: Supported by National Institutes of Health grant 1R01AA021913-01. Dr Wisk was supported by the Agency for Healthcare Research and Quality (K12HS022986). Funded by the National Institutes of Health (NIH).

We thank Boston Children’s Hospital for providing access to study clinic sites and Drs Fatma Dedeoglu, Laurie Fishman, Katharine Garvey, Andrew MacGinnitie, Paul Rufo, and Joseph Wolfsdorf for facilitating access to patients and enabling this research.

1
Kann
L
,
Kinchen
S
,
Shanklin
SL
, et al;
Centers for Disease Control and Prevention (CDC)
.
Youth risk behavior surveillance–United States, 2013 [published correction appears in MMWR Morb Wkly Rep. 2014;63(26):576].
MMWR Suppl
.
2014
;
63
(
4
):
1
168
2
Sawyer
SM
,
Drew
S
,
Yeo
MS
,
Britto
MT
.
Adolescents with a chronic condition: challenges living, challenges treating.
Lancet
.
2007
;
369
(
9571
):
1481
1489
3
Van Cleave
J
,
Gortmaker
SL
,
Perrin
JM
.
Dynamics of obesity and chronic health conditions among children and youth.
JAMA
.
2010
;
303
(
7
):
623
630
[PubMed]
4
Wisk
LE
,
Weitzman
ER
.
Substance use patterns through early adulthood: results for youth with and without chronic conditions.
Am J Prev Med
.
2016
;
51
(
1
):
33
45
[PubMed]
5
Levy
S
,
Dedeoglu
F
,
Gaffin
JM
, et al
.
A screening tool for assessing alcohol use risk among medically vulnerable youth.
PLoS One
.
2016
;
11
(
5
):
e0156240
6
Weitzman
ER
,
Ziemnik
RE
,
Huang
Q
,
Levy
S
.
Alcohol and marijuana use and treatment nonadherence among medically vulnerable youth.
Pediatrics
.
2015
;
136
(
3
):
450
457
[PubMed]
7
Suris
J-C
,
Parera
N
.
Sex, drugs and chronic illness: health behaviours among chronically ill youth.
Eur J Public Health
.
2005
;
15
(
5
):
484
488
[PubMed]
8
Erickson
JD
,
Patterson
JM
,
Wall
M
,
Neumark-Sztainer
D
.
Risk behaviors and emotional well-being in youth with chronic health conditions.
Child Health Care
.
2005
;
34
(
3
):
181
192
9
Weathermon
R
,
Crabb
DW
.
Alcohol and medication interactions.
Alcohol Res Health
.
1999
;
23
(
1
):
40
54
[PubMed]
10
National Institute on Alcohol Abuse and Alcoholism
. Harmful interactions: mixing alcohol with medicines. 2014. Available at: https://pubs.niaaa.nih.gov/publications/medicine/medicine.htm. Accessed November 11, 2017
11
Jang
ES
,
Jeong
SH
,
Hwang
SH
, et al
.
Effects of coffee, smoking, and alcohol on liver function tests: a comprehensive cross-sectional study.
BMC Gastroenterol
.
2012
;
12
:
145
[PubMed]
12
Hingson
R
,
Zha
W
,
Smyth
D
.
Magnitude and trends in heavy episodic drinking, alcohol-impaired driving, and alcohol-related mortality and overdose hospitalizations among emerging adults of college ages 18-24 in the United States, 1998-2014.
J Stud Alcohol Drugs
.
2017
;
78
(
4
):
540
548
[PubMed]
13
Breslow
RA
,
Dong
C
,
White
A
.
Prevalence of alcohol-interactive prescription medication use among current drinkers: United States, 1999 to 2010.
Alcohol Clin Exp Res
.
2015
;
39
(
2
):
371
379
[PubMed]
14
National Institute on Alcohol Abuse and Alcoholism
.
Alcohol Screening and Brief Intervention for Youth: A Practitioner’s Guide
.
Rockville, MD
:
National Institute on Alcohol Abuse and Alcoholism
;
2015
15
Wolters Kluwer Clinical Drug Information, Inc
.
Lexicomp Online, Pediatric and Neonatal Lexi-Drugs Online
.
Hudson, OH
:
Wolters Kluwer Clinical Drug Information, Inc
;
2015
16
Mackinnon
DP
,
Warsi
G
,
Dwyer
JH
.
A simulation study of mediated effect measures.
Multivariate Behav Res
.
1995
;
30
(
1
):
41
[PubMed]
17
Levy
SJ
,
Williams
JF
;
Committee on Substance Use and Prevention
.
Substance use screening, brief intervention, and referral to treatment.
Pediatrics
.
2016
;
138
(
1
):
e20161211
[PubMed]
18
Donoghue
K
,
Patton
R
,
Phillips
T
,
Deluca
P
,
Drummond
C
.
The effectiveness of electronic screening and brief intervention for reducing levels of alcohol consumption: a systematic review and meta-analysis.
J Med Internet Res
.
2014
;
16
(
6
):
e142
[PubMed]
19
Glass
JE
,
Hamilton
AM
,
Powell
BJ
,
Perron
BE
,
Brown
RT
,
Ilgen
MA
.
Revisiting our review of Screening, Brief Intervention and Referral to Treatment (SBIRT): meta-analytical results still point to no efficacy in increasing the use of substance use disorder services.
Addiction
.
2016
;
111
(
1
):
181
183
[PubMed]
20
Huang
JS
,
Gottschalk
M
,
Pian
M
,
Dillon
L
,
Barajas
D
,
Bartholomew
LK
.
Transition to adult care: systematic assessment of adolescents with chronic illnesses and their medical teams.
J Pediatr
.
2011
;
159
(
6
):
994
998.e2
[PubMed]
21
Newacheck
PW
,
Kim
SE
.
A national profile of health care utilization and expenditures for children with special health care needs [published correction appears in Arch Pediatr Adolesc Med. 2005;159(4):318].
Arch Pediatr Adolesc Med
.
2005
;
159
(
1
):
10
17
[PubMed]
22
Weitzman
ER
,
Salimian
P
,
Rabinow
L
,
Levy
S
.
Shaped by my disease: perspectives on substance use shared by youth with chronic medical conditions and implications for preventive interventions and clinical guidance
. In:
Society for Adolescent Health and Medicine Annual Meeting
;
March 9–12, 2016
;
Washington, DC
23
Patrick
ME
,
Schulenberg
JE
.
Prevalence and predictors of adolescent alcohol use and binge drinking in the United States.
Alcohol Res
.
2013
;
35
(
2
):
193
200
[PubMed]
24
Bendtsen
P
,
Damsgaard
MT
,
Huckle
T
, et al
.
Adolescent alcohol use: a reflection of national drinking patterns and policy?
Addiction
.
2014
;
109
(
11
):
1857
1868
[PubMed]
25
Abrams
GB
,
Kolligian
J
 Jr
,
Mills
DL
,
DeJong
W
.
Failure of college students to complete an online alcohol education course as a predictor of high-risk drinking that requires medical attention.
Am J Drug Alcohol Abuse
.
2011
;
37
(
6
):
515
519
[PubMed]
26
Hasking
P
,
Schofield
L
.
Examining alcohol consumption with the theory of planned behaviour: do health and alcohol knowledge play a role?
Psychol Health Med
.
2015
;
20
(
7
):
838
845
[PubMed]
27
Lapsley
DK
,
Hill
PL
.
Subjective invulnerability, optimism bias and adjustment in emerging adulthood.
J Youth Adolesc
.
2010
;
39
(
8
):
847
857
[PubMed]
28
Milam
JE
,
Sussman
S
,
Ritt-Olson
A
,
Dent
CW
.
Perceived invulnerability and cigarette smoking among adolescents.
Addict Behav
.
2000
;
25
(
1
):
71
80
29
Johnston
LD
,
O’Malley
PM
,
Miech
RA
,
Bachman
JG
,
Schulenberg
JE
. Monitoring the future: national survey results on drug use, 1975-2016: overview, key findings on adolescent drug use. 2017. Available at: www.monitoringthefuture.org/pubs/monographs/mtf-overview2016.pdf. Accessed July 27, 2017

Competing Interests

POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.

Supplementary data