Prescription Medication Use Among Children and Adolescents in the United States

NHANES is a nationally representative survey sampled from the US civilian, noninstitutionalized population conducted by the National Center for Health Statistics. Sampling methods are described elsewhere.²⁰ We restricted our sample to children and adolescents ages 0 to 19 years who responded to the prescription medication questionnaires. A parent or caregiver provided information for survey participants who were <16 years of age and for those who could not answer themselves. We used the 6 most recent NHANES cycles (2003–2004, 2005–2006, 2007–2008, 2009–2010, 2011–2012, and 2013–2014). A total of 23 179 children and adolescent participants ages 0 to 19 years were sampled over the 10-year period examined. Twenty-seven participants were excluded because of missing medication information, yielding an analytic subsample of 23 152 participants.

During the household interview participants were asked whether they had taken a prescription medication in the past 30 days.²¹ Those who answered “yes” were asked to show the interviewer the medication containers for all the medications used. If no container was available, the participant was asked to name the prescription medication they used. Prescription medication names and therapeutic classes were coded by using Lexicon Plus (Cerner Multum, Inc, Kansas City, MO), a proprietary database of Cerner Multum, Inc.

We differentiated acute versus chronic prescription medication use on the basis of duration of use. For each medication, participants were asked “How long was medication taken?” We defined acute use as the use of a prescription medicine for ≤30 days and all other use as chronic use. For prescription medications used chronically, approximately three-fourths (73%) were used for >365 days. We defined concurrent use as the simultaneous use of at least 2 prescription medications during the past 30 days. We also identified therapeutic classes most commonly used together, overall and by age group and gender.

We used Micromedex to identify potential DDIs for all prescription medications used in children and adolescents during 2013–2014. Micromedex defines contraindicated combinations as “drugs [that] are contraindicated for concurrent use” and major DDIs as “the interaction may be life threatening, require medical intervention to minimize or prevent serious adverse events, or both.”

We used descriptive statistics to estimate the prevalence of prescription medication and concurrent use for each of the 6 NHANES cycles examined. For analyses stratified by age group and gender, we focused on the most recent cycle (2013–2014); to increase sample size, however, we pooled the most recent 3 cycles (2009–2010, 2011–2012, and 2013–2014) for the analyses of commonly used combinations of therapeutic drug classes and DDIs. We used Taylor linearization methods to incorporate sample weights to adjust for the complex sampling methods in NHANES for estimate prevalence and conduct statistical tests. For most analyses, we used logistic regression to test for significance of differences by gender within age groups; to examine gender differences in the concurrent use of therapeutic drug classes and DDIs, the Pearson’s χ² test was used. We tested for the statistical significance of trends across cycles using logistic regression. We used Stata (Stata Corp, College Station, TX) version 14 to perform all analyses. All P values reported are 2-sided. The study was considered exempt by a University of Illinois at Chicago Institutional Review Board.

Table 1 reports prevalence of the use of prescription medications overall and by population characteristics for 2013–2014. Nearly one-fifth (19.8%) of children and adolescents ≤19 years old used at least 1 prescription medication in the previous 30 days; 13.9% used chronic medications, and 7.1% used acute medications. Prescription medication use increased with age, ranging from 14.7% in children ages 0 to 5 years to 22.8% among adolescents (13–19 years). Acute medication use was highest in younger children (9.1%), whereas the use of chronic medications was lowest (6.4%).

Figure 1 depicts the prevalence of the use of prescription medications over time (Fig 1A) and by age group and gender (Fig 1B). The most commonly used prescription medications included respiratory agents, especially bronchodilators (the most common type was albuterol) and psychotherapeutic agents, especially central nervous system (CNS) stimulants (methylphenidate was most common) and antidepressants (fluoxetine was most common) (Table 2). There was a notable decline in the use of acute medications between 2003–2004 and 2013–2014 (10.9% [95% confidence interval (CI), 9.3%–12.9%] to 7.1% [CI, 6.0%–8.4%]; P < .01), driven largely by a decrease in the use of antibiotics (7.9% [CI, 6.6%–9.5%] to 4.8% [CI, 3.7%–6.1%]; P < .01; Supplemental Table 5).

The use of prescription medications was highest among adolescent girls (28.0% [CI, 21.3%–35.8%]) and boys ages 6 to 12 (26.5% [CI, 22.5%–30.9%]; Fig 1). Gender differences were also most pronounced in these 2 age groups. Among children ages 6 to 12, boys were nearly twice as likely to use prescription medications compared with girls (26.5% [CI, 22.5%–30.9%] vs 15.1% [CI, 11.8%–19.0%]; P < .01); this difference was primarily due to a higher rate of chronic medication use (22.3% [CI, 19.1%–25.9%] vs 10.1% [CI, 7.7%–13.0%]; P < .01), especially CNS stimulants (9.4% [CI, 7.2%–12.0%] vs 2.2% [CI, 1.2%–4.1%]; P < .01), α-adrenergic agonists (3.3% [CI, 1.9%–5.7%] vs 1.1% [CI, 0.3%–3.5%]; P < .01), and leukotriene modifiers (4.0% [CI, 2.5%–6.4%] vs 0.9% [CI, 0.4%–1.9%], respectively; P < .01) (Table 2).

During adolescence, however, girls were more likely to use prescription medications than boys; use was 28.0% (CI, 21.3%–35.8%) among girls versus 17.6% (CI, 13.3%–22.9%) among boys (P < .01). This difference was due to a higher rate of both acute (9.3% [CI, 5.9%–14.4%] vs 4.8% [CI, 3.2%–7.2%]; P < .05) and chronic (21.3% [CI, 15.0%–29.4%] vs 13.8% [CI, 10.2%–18.5%]; P < .05) medication use, specifically antidepressants (3.8% [CI, 1.9%–7.6%] vs 1.3% [CI, 0.5%–3.7%]; P < .01), antibiotics (6.6% [CI, 3.6%–11.7%] vs 3.5% [CI, 2.0%–6.0%]; P < .01), analgesics (3.2% [CI, 1.6%–6.4%] vs 1.2% [CI, 0.6%–2.3%]; P < .01), and antiemetics (1.6% [CI, 0.8%–3.3%] vs 0.1% [CI, 0.0%–1.2%]; P = .06) (Table 2). Rates of CNS stimulant use, in contrast, were lower in adolescent girls than in boys (1.6% [CI, 0.9%–3.0%] vs 4.3% [CI, 2.3%–7.9%]; P < .01).

Figure 2 depicts the prevalence of the use of concurrent prescription medications over time (Fig 2A) and by age group and gender (Fig 2B). During 2013 and 2014, 7.5% (CI, 6.6%–8.6%) of children and adolescents concurrently used prescription medications, a decrease from 2003–2004 (9.9% [CI, 9.1%–10.8%]; P < .01). The concurrent use of prescription medications varied across age groups for both boys and girls and was substantially more common among boys ages 6 to 12 than among girls in this age group (11.9% [CI, 8.8%–15.8%] vs 4.7% [CI, 3.4%–6.7%]; P < .01).

Among concurrent users of prescription medications, combinations that involved respiratory agents, such as bronchodilators and leukotriene modifiers, and psychotropic medications, including antidepressants, atypical antipsychotics, and CNS stimulants, were the most prevalent (Table 3). More than 20% (23.7% [CI, 11.1%–43.8%]) of boys ages 6 to 12 and more than half (57.9% [CI, 35.1%–77.7%]) of adolescent boys who used CNS stimulants concurrently used ≥2 other psychotropic medications. Combinations that involved antidepressants were most prevalent among adolescent girls; among these antidepressant users, 51.2% [CI, 19.4%–82.6%] were concurrently using at least 2 psychotropic medications.

Among concurrent users of ≥2 prescription medications during 2013–2014 (N = 301), a total of 156 unique prescription medications were used (Supplemental Table 6). Using Micromedex, a total of 826 potential DDIs (contraindicated [n = 29] or potentially major [n = 797]) for these 156 medications were identified. In the pooled 2003–2014 sample, 101 unique combinations of these potential DDIs were used by 158 people (Supplemental Table 7). In the pooled 2009–2014 sample, 63 unique DDI combinations were used by 78 people, making the overall rate 0.7% (CI, 0.5%–1.0%). We found only 1 individual using a contraindicated DDI (aripiprazole–metoclopramide).

Among concurrent users of prescription medications, 8.2% [CI, 6.0%–11.1%] were at risk for potential major DDIs (Table 4). Nearly half of interacting regimens used involved psychotropic agents, primarily antidepressants, the most common adverse interaction effect being QT prolongation. The majority (68.0% [CI, 55.9%–78.0%]) of these interacting regimens involved the use of at least 1 acute medication.

The use of combinations with potential major DDIs was higher among adolescent girls (18.1% [CI, 11.0%–28.4%]) than among boys (6.6% [CI, 3.6%–11.9%], P < .05). This difference was largely due to the disproportionate use of tricyclic antidepressants in interacting combinations that involved acute medications, specifically nonsteroidal anti-inflammatory drugs (NSAIDs) (most commonly diclofenac–amitriptyline), antiemetics (ondansetron–nortriptyline), and albuterol (albuterol–amitriptyline) by adolescent girls. Atypical antipsychotics accounted for the vast majority of interacting drug combinations in boys ages 6 to 12, whereas antidepressants accounted for the majority in adolescent boys.

Whereas overall prevalence of potential major DDIs among concurrent users of prescription medications is similar to that reported during 2003–2008 (8.7% [95% CI, 6.5%–11.5%]), there was a significant increase among adolescent girls (11.8% [CI, 6.9%–19.7%] during 2003–2008 vs 18.1% [CI, 11.0%–28.4%] during 2009–2014; P < .05; Supplemental Table 8).

We used directly observed medication data from nationally representative samples to examine use and concurrent use of prescription medications among children and adolescents in the United States. During 2013–2014, one-fifth of children and adolescents used at least 1 prescription medication and ∼1 in 10 concurrently used ≥2 prescription medications. Among children and adolescents concurrently using ≥2 prescription medications, 1 in 12 was at risk for a major DDI. Our findings are important because prescription medications are commonly used by children and adolescents in the United States, and population-level monitoring is needed to detect modifiable risk and optimize safe use.

Since 2003–2004, we observed a decline in the use of acute medications, corroborating findings from a previous analysis of pharmacy claims.²² This decline was mainly attributable to lower rates of antibiotic use, which was also observed in previous research. We also found, however, that acute medications were commonly used in potentially life-threatening drug combinations, particularly among adolescent girls. More than three-quarters of potential major DDIs we identified involved prescription medications used acutely (eg, macrolide antibiotics, antiemetics, and albuterol), and nearly half of these interacting combinations also involved an antidepressant. The potential for QT prolongation is especially noteworthy given that it is often asymptomatic and unpredictable yet can develop quickly into a serious arrhythmia or sudden cardiac death,^23,24 a serious yet underreported problem in children and adolescents.^25,26 Instability in insurance coverage for children and fragmentation of health care (eg, use of retail clinics and emergency departments for episodic care) may increase the risk that a physician prescribing medication for an acute illness is unaware of a child’s chronic medication regimen. Systemic strategies, including patient and provider access to accurate medication use data, are needed to reduce the risk of such major DDIs among children.

Highlighted in our findings is also an important opportunity to improve the safe use of medications among adolescent girls given that nearly 1 in 5 adolescent girls who concurrently used prescription medications was at risk for a DDI. The threefold higher prevalence of these interacting drug regimens among adolescent girls, when compared with adolescent boys, was largely due to a higher rate of tricyclic antidepressant use in combination with acute medications, most commonly macrolide antibiotics, antiemetics, NSAIDs, and proton pump inhibitors (PPIs). These acute medications may be prescribed to treat sexually transmitted infections (eg, azithromycin) or gastrointestinal symptoms (eg, ondansetron, omeprazole) from eating disorders.^27,28 Some are also available over-the-counter (OTC), specifically NSAIDs and PPIs. Product labeling for OTC medications does not always include comprehensive information on adverse effects, including DDIs. For example, in the labeling for OTC omeprazole, the DDI between citalopram and omeprazole is not mentioned.²⁹ Many adolescent girls and their parents may therefore not be aware of the cardiovascular risks associated with use.

We also found that prescription medications associated with an increased risk of suicidality are commonly used in children and adolescents and are often used together. For example, more than half of adolescent girls taking antidepressants concurrently use at least 2 additional psychotropic medications or hormonal contraceptives. Although there is some evidence that the combined use of these drugs may increase the onset and severity of suicidal thoughts and behavior,^19,30 we found no cases of DDIs associated with suicidality. This finding is expected because suicidality, in contrast to serotonin syndrome and QT prolongation, is not captured as a potential DDI in existing drug interaction software,³¹ including Micromedex. Therefore, health care professionals, including psychiatrists, may not be fully aware of the suicidal risks associated with the concurrent use of prescription medications in the evaluation and treatment of depressive symptoms in younger patients.

With these findings, it is suggested that preventive efforts used to improve the safe use of medications by children should promote public and health professional awareness of the increased risks associated with concurrent use of prescription and OTC medications, particularly suicidality and QT prolongation. Such efforts may include incorporation of a list of commonly used medications and interacting combinations associated with increases in QT prolongation and suicidal risks in treatment guidelines and screening tools for depression. Efforts should be used to target both consumers and health professionals, including those involved in acute or episodic care. Such efforts are particularly important among adolescents considering suicide and cardiovascular events, as well as unintentional drug overdoses, are persistently leading causes of death.^16,32 Nonmedical use of prescription drugs, including opioids, is also more prevalent in adolescents than in younger children³³ and is known to increase risk for both suicide and sudden death.

This study has several limitations. We examined the potential for DDIs rather than actual adverse drug events. Prescribers might have been aware of the risks and benefits of coprescribing the medications, established an appropriate monitoring plan, and determined in discussion with the patient and/or caregiver that the potentially interacting regimen may be the best option for the patient. Second, our medication data were limited to prescription medications and were not used to capture information on OTC products. Therefore, we likely underestimate the rates of use, and potential DDIs, for several therapeutic classes (eg, antihistamines, PPIs, NSAIDs) that are widely available without a prescription. Moreover, OTC products, including acetaminophen and cough and cold medicines, are key contributors to adverse drug events in children.³⁴ Third, we used the Micromedex drug interaction software to identify potentially major DDIs in our sample, and other drug interaction software may yield different estimates. However, the accuracy, including sensitivity and specificity, comprehensiveness, and usefulness of Micromedex has been previously established.^35,36For example, authors of a study in which the various software for DDIs were evaluated ranked Micromedex the highest in accuracy.³⁵ Fourth, the estimates for DDIs are for 30-day point prevalence. Therefore, it is possible that over a longer period of time, the concurrent use of potential DDIs is greater than our estimates.

In addition, the distinction between acute and chronic is not absolute; some medications initiated <30 days before data collection, yet intended for chronic use, may have been misclassified as acute. Last, for children <16 years old, a parent or guardian answered questions on their child’s use of prescription medications. Therefore, underreporting is possible, particularly for prescription medications surreptitiously used by the child. Finally, NHANES may be underpowered to examine the prevalence of medication use, resulting in unreliable nationwide estimates.

Using nationally representative data, we found that many children and adolescents use and concurrently use prescription medications in the United States. Among concurrent users of prescription medications, nearly 1 in 12 was at risk for a major DDI. Largely because of their higher rate of acute medication use, adolescent girls were at a higher risk of using interacting drug regimens than other subgroups. Treatment and prevention efforts to reduce the burden of adverse drug events in younger populations should be used to consider the role of interacting drug combinations, especially among these individuals.

CI

confidence interval

CNS

central nervous system

DDI

drug–drug interaction

NSAID

nonsteroidal anti-inflammatory drug

OTC

over-the-counter

PPI

proton pump inhibitor