Prescribing of controlled substances to children, adolescents, and young adults enrolled in Medicaid, 2001-2019

Abstract

Objective.

To examine nationwide trends in the prescribing of controlled medications to early adolescents, adolescents, and young adults enrolled in public insurance (Medicaid) from 2001-2019.

Methods.

The study utilized US Medicaid data covering publicly insured enrollees from 43 states (2001-2019). Early adolescents (10-12y), adolescents (13-17y), and young adults (18-24y, 25-29y) with ≥10 months enrollment in each calendar year were included. Filled prescription for opioids, stimulants, benzodiazepines, Z-hypnotics, barbiturates, and gabapentin. In each calendar year, annual proportions with 1+ controlled medication, 2+ classes of controlled medications, and each controlled medication were estimated.

Results.

In 2019, the sample included 17.9 million enrollees (53% female). The annual proportion prescribed any controlled medication peaked at 17.5% in early adolescents (2003), 20.6% in adolescents (2009), and 34.1% (18-24y) and 47.0% (25-29y) in young adults (2010). By 2019, the proportions declined to 11.7% (early adolescents), 12.6% (adolescents), 16.2% (18-24y), and 23.9% (25-29y). Trends varied by medication and age. The largest absolute decline was in the proportion with an opioid filled (2010=29.8%, 2019=11.2%, young adults 18-24y; 2003=14.3%, 2019=4.4%, adolescents). In contrast, the proportion with a stimulant fill increased, with eight-fold increases in young adults 25–29y (2001=0.3%, 2019=2.6%). Benzodiazepine and Z-hypnotic use peaked in 2010 and declined through 2019.

Conclusions.

In the past two decades, there were increases in stimulant prescriptions among young Medicaid enrollees. The declines in opioid, benzodiazepines, barbiturate and Z-hypnotic prescribing are encouraging and may indicate more cautious prescribing related to greater awareness of harms such as misuse and overdose, along with policy initiatives.

1.1. INTRODUCTION

The United States Drug Enforcement Administration (DEA) sets national regulations for controlled substances (United States Drug Enforcement Administration, 2024), defined as those with potential for misuse or dependency. Prescription medications such as opioids, stimulants, benzodiazepines (BZD), Z-hypnotics, and barbiturates are federally controlled prescription drugs (‘controlled medications’ hereafter). Further, individual state laws classify additional medications as controlled, such as gabapentin, which is scheduled in some states (Campbell, Coomer, Jacob, & Lenz, 2021; Karavolis, Su, & Peckham, 2022). Greater oversight and regulations of prescribing of controlled medications also aim to reduce harmful drug-related outcomes (Preuss, Kalava, & King, 2022). For example, prescription drug monitoring programs (PDMPs) are a regulatory means of tracking controlled medication prescriptions in each state (Smith et al., 2019).

Controlled medications are FDA approved for at least one clinical indication. Historically, many of these medications were considered appropriately prescribed to improve patient outcomes; however, the risks of misuse, dependency, and overdose have become increasingly prominent concerns (U.S. Food and Drug Administration, 2016, 2020, 2023). Controlled medications are frequently misused, particularly in young adults (McCabe et al., 2024; McCabe et al., 2019; Substance Abuse and Mental Health Services Administration, 2022). In 2021, the National Survey on Drug Use and Health reported that 14 million people aged 12 and older (5.1%) misused a prescription psychotherapeutic drug in the past year, with the highest prevalence in young adults (7.4%) (Substance Abuse and Mental Health Services Administration, 2022). While there are many ways to obtain controlled medications other than a prescription (e.g., friend, dealer), the most common source in 2022 of prescription opioid, stimulant, or BZD misuse in adolescents was through prescriptions received by users (37%) (McCabe et al., 2024). Both medical use and misuse of controlled medications have been associated with a risk for substance use disorders and controlled medications are involved in overdose events (McCabe, Veliz, Boyd, & Schulenberg, 2017; McCabe et al., 2019; Nguyen et al., 2024; Preuss et al., 2022), leading to calls for careful prescribing of controlled medications to young people with regular screening for misuse and substance use problems during treatment (McCabe et al., 2019).

The period from childhood to young adulthood is a particularly vulnerable period for risks from controlled medications. It is characterized by increasing risk-taking behavior, substance initiation, psychiatric illness onset and progression, and increasing misuse (Chambers, Taylor, & Potenza, 2003; Perlmutter et al., 2018; Sawyer et al., 2012). Half of mental disorders begin during adolescence (Kessler RC et al., 2005; Merikangas KR et al., 2010), for which stimulants, BZDs, Z-hypnotics, and barbiturates can be prescribed. The risk for prescription drug misuse often begins in adolescence and increases into young adulthood (Huang et al., 2006; McCabe, West, Morales, Cranford, & Boyd, 2007)(McCabe et al., 2019). Within young age groups, there is also extensive heterogeneity with regards to substance use, prescription misuse, and treatment practices; thus, it is important to examine prescribing trends in age subgroups.

Given safety concerns, trends in the prescribing of controlled medications are important to evaluate, especially in light of shifts in prescribing over the past two decades (Bykov, He, & Gagne, 2020; McCabe et al., 2024; Meadows et al., 2020; Paiva et al., 2023; Paulozzi et al., 2015). In a sample of privately insured adults, stimulant use increased from 2004-2019, while opioids, BZDs and Z-hypnotics declined since the early 2010s (Bykov et al., 2020). In adolescents, self-reported lifetime medical use of a stimulant, opioid, or BZD declined from 2009 to 2022 (24% to 16%) (McCabe et al., 2024). However, there has not been a collective examination of the prescribing of controlled medications in a national sample of publicly insured youth. Compared to the privately insured population, those enrolled in Medicaid have, on average, a higher prevalence of mental illness and substance use problems (Saunders H, 2022). We therefore aimed to examine nationwide trends in the prescribing of controlled medications to early adolescents, adolescents, and young adults enrolled in Medicaid from 2001 to 2019. We anticipated an overall decline in opioid prescribing to youth enrolled in Medicaid with effects extending to other sedating controlled substances given rising concerns around overdose.

1.2. METHODS

1.2.1. Data source.

Our analysis used Medicaid Analytic eXtracts (MAX) and Transformed Medicaid Statistical Information System Analytic Files (TAF) administrative fee-for-service claims and managed care encounter data from 1/1/2001-12/31/2019. Medicaid files include data for those insured by Medicaid or Children’s Health Insurance Program (CHIP) (Centers for Medicare & Medicaid Services, 2018), representing the largest collection of healthcare data for youth in the US. We utilized patient-level information on enrollment and demographics (MAX: Person Summary file; TAF: Annual Demographic and Eligibility file) and outpatient filled prescriptions (MAX: Prescription Drug file; TAF: Pharmacy file).

1.2.2. Study population.

The study cohort included participants from 43 states with Medicaid enrollment and prescription data available for all study years, 2001-2019; Arizona, Delaware, Nevada, Oregon, and Rhode Island were excluded due to lack of data available to the researchers prior to 2014 and Maine and Kansas were excluded due to missing prescription information in several study years. We included early adolescents (10-12y), adolescents (13-17y), and two groups of young adults (18-24y, 25-29y) with ≥10 months of Medicaid enrollment with full scope benefits in that calendar year. We excluded individuals with dual Medicaid-Medicare eligibility, in long-term care facilities, and enrolled in standalone state CHIP at any point in the calendar year due to concerns regarding data completeness.

1.2.3. Controlled medications.

We identified individuals with at least one prescription fill for an opioid, stimulant, BZD, Z-hypnotic, or barbiturate any time in that calendar year. We additionally included gabapentin, classified as a controlled substance in a growing number of states (Karavolis et al., 2022). Prescriptions were selected based on National Drug Codes (NDC), which were identified in First Databank based on agent (eTable 1); any relevant NDC from 2001-2019 was included. We focused on these six medication classes with a high prevalence of use in children, adolescents, or adults, involvement in overdoses, and/or use for treating common mental health or pain conditions (Centers for Disease Control and Prevention (CDC), 2025; Paulozzi et al., 2015; Stangeland et al., 2023; Substance Abuse and Mental Health Services Administration, 2024).

The primary controlled medication definition included any prescription product that contained a controlled substance, including combination products. Two secondary definitions were also considered: 1) opioid cold and cough suppressant agents were excluded and not considered controlled medications, and 2) all NDCs on the DEA controlled substance exclusion list as of 1/11/2024 (Drug Enforcement Administration, 2024) were excluded from our definition and not considered a ‘controlled medication’. The DEA list excludes products from being deemed a controlled substance (e.g., ‘butalbital, acetaminophen’ products); this exclusion primarily affected barbiturates.

1.2.4. Analysis.

For each calendar year, we identified individuals with 1+ prescription fill for an opioid, stimulant, BZD, Z-hypnotic, barbiturate, or gabapentin. The annual proportion was estimated from 2001 to 2019 by dividing the number of eligible enrollees with a prescription during that calendar year by all eligible enrollees during that calendar year. Data were pooled for the included 43 states and consistent methods were applied across age groups and medications. We estimated the annual proportion prescribed any controlled medication (opioid, stimulant, BZD, Z-hypnotic, barbiturate, gabapentin) and each controlled medication separately. We additionally estimated the annual proportion prescribed 2+ classes of controlled medications in that calendar year (e.g., opioid + stimulant; barbiturate + z-drug). Trends were stratified by age group: early adolescents (10-12y), adolescents (13-17y), and young adults (18-24y, 25-29y), and by sex. Within strata, we reported the absolute difference in proportions and the prevalence ratio (PR) from 2019 vs. 2001. The average annual proportion change (APC) was estimated for each medication class. Analyses were completed under primary and secondary controlled medication definitions (Section 1.2.3).

Sensitivity analyses excluded states with high fluctuation in year-to-year estimates to remove states with potentially inconsistent capture of prescription and enrollment data over the study period. Specifically, we excluded 14 states with ≥200% year-to-year shifts in the number of controlled medications filled and 1 state with ≥200% year-to-year shift in the denominator (enrolled population meeting inclusion criteria); only one state (NH) had a shift that corresponded with the year of Medicaid expansion. To determine if Medicaid expansion and the broadening of eligibility criteria for Medicaid enrollment impacted trends, we conducted a sensitivity analysis restricted to states without Medicaid expansion by 2019. Due to potential variability in the capture of services, we additionally stratified by Medicaid payor type, classified as a) fee-for-service, b) comprehensive managed care, or c) other managed care based on their last enrolled month in the year.

1.3. RESULTS

1.3.1. The study population included 9.0 to 18.4 million eligible individuals per calendar year (17,886,587 in 2019). In 2019, 53% of eligible individuals were female, 25% aged 10-12 years, 36% 13-17 years, 24% 18-24 years, and 15% 25-29 years.

1.3.2. Any controlled medication.

From 2001 to 2019, there was an overall decline in the proportion of young people with a prescription for a controlled medication, with substantial variation by age group (Figure 1) and a larger absolute decline in females than males (Table 1). In early adolescents and adolescents, the proportion prescribed a controlled medication was 16% (95%CI:16%-16%) and 18% (18%-18%), respectively, in 2001, and declined to 12% and 13% by 2019 (Table 1, Figure 1; CIs not displayed given precision due to large sample). In young adults, the proportion was higher but followed a similar trend; in 2001, 28% aged 18-24y had at least one controlled medication filled, peaking at 34% in 2010 and declining to 16% by 2019. Trends were consistent when excluding gabapentin, under the alternative secondary definition of controlled medications (eFigure 1), by Medicaid payor type (eFigure 2).

Figure 1.

a) Trends in the annual proportion of Medicaid enrollees with any controlled medication prescription by age group, 2001-2019

b) Trends in the annual proportion prescribed two or more classes of controlled medications by age group

Table 1.

Annual proportion of early adolescents, adolescents, and young adults enrolled in Medicaid prescribed a controlled medication

	2001		2007		2013		2019		2019 vs. 2001, absolute change, %	2019 vs. 2001, prevalence ratio (PR) 95% CI
	No. with Rx	Proportion, % *	No. with Rx	Proportion, % *	No. with Rx	Proportion, % *	No. with Rx	Proportion, % *
Any controlled substance
Age 10-12y	372,141	15.60	430,570	16.31	613,474	16.03	512,498	11.65	−3.94	0.75 (0.74-0.75)
13-17y	546,154	17.78	793,701	18.93	1,044,834	18.85	806,610	12.58	−5.19	0.71 (0.71-0.71)
18-24y	524,052	28.35	681,491	30.08	952,316	30.39	703,692	16.19	−12.16	0.57 (0.57-0.57)
25-29y	320,771	36.31	420,367	39.68	704,309	45.43	653,172	23.92	−12.39	0.66 (0.66-0.66)
Sex
Male	691,356	19.62	956,880	21.13	1,405,908	21.88	1,286,062	15.18	−4.43	0.77 (0.77-0.78)
Female	1,071,735	22.97	1,369,224	24.32	1,909,002	25.02	1,389,904	14.76	−8.21	0.64 (0.64-0.64)
2+ classes of controlled substances
Age 10-12y	32,797	1.37	38,462	1.46	42,563	1.11	20,778	0.47	−0.90	0.34 (0.34-0.35)
13-17y	52,395	1.71	85,762	2.05	109,142	1.97	60,688	0.95	−0.76	0.56 (0.55-0.56)
18-24y	64,072	3.47	100,992	4.46	149,588	4.77	89,626	2.06	−1.40	0.59 (0.59-0.60)
25-29y	60,995	6.90	97,797	9.23	178,758	11.53	131,058	4.80	−2.10	0.70 (0.69-0.70)
Sex
Male	78,242	2.22	119,032	2.63	166,563	2.59	116,921	1.38	−0.84	0.62 (0.61-0.62)
Female	132,014	2.83	203,980	3.62	313,486	4.11	185,229	1.97	−0.86	0.70 (0.69-0.70)
Opioids a
Age 10-12y	197,274	8.27	203,136	7.70	198,830	5.20	67,202	1.53	−6.74	0.18 (0.18-0.19)
13-17y	390,966	12.72	528,518	12.60	580,583	10.47	281,682	4.39	−8.33	0.35 (0.34-0.35)
18-24y	467,654	25.30	598,065	26.39	791,255	25.25	487,242	11.21	−14.09	0.44 (0.44-0.44)
25-29y	282,459	31.97	375,644	35.46	618,396	39.89	484,910	17.76	−14.21	0.56 (0.55-0.56)
Sex
Male	419,088	11.89	556,946	12.30	688,495	10.71	446,201	5.27	−6.62	0.44 (0.44-0.44)
Female	919,247	19.70	1,148,398	20.40	1,500,555	19.67	874,834	9.29	−10.41	0.47 (0.47-0.47)
Stimulants b
Age 10-12y	182,241	7.64	241,029	9.13	427,857	11.18	434,021	9.87	2.23	1.29 (1.29-1.30)
13-17y	147,307	4.79	278,036	6.63	480,391	8.66	501,591	7.82	3.03	1.63 (1.62-1.64)
18-24y	14,311	0.77	48,317	2.13	113,501	3.62	135,223	3.11	2.34	4.02 (3.95-4.09)
25-29y	2,554	0.29	9,731	0.92	36,027	2.32	70,605	2.59	2.30	8.94 (8.60-9.30)
Sex
Male	259,728	7.37	405,618	8.96	715,544	11.14	763,297	9.01	1.64	1.22 (1.22-1.23)
Female	86,674	1.86	171,490	3.05	342,225	4.49	378,138	4.02	2.16	2.16 (2.15-2.18)
Benzodiazepines c
Age 10-12y	10,985	0.46	15,962	0.60	23,686	0.62	26,484	0.60	0.14	1.31 (1.28-1.34)
13-17y	25,078	0.82	43,997	1.05	63,573	1.15	59,362	0.93	0.11	1.13 (1.12-1.15)
18-24y	56,367	3.05	88,772	3.92	131,967	4.21	106,991	2.46	−0.59	0.81 (0.80-0.82)
25-29y	57,836	6.55	92,106	8.69	158,758	10.24	136,830	5.01	−1.54	0.77 (0.76-0.77)
Sex
Male	47,188	1.34	77,644	1.71	124,580	1.94	133,714	1.58	0.24	1.18 (1.17-1.19)
Female	103,077	2.21	163,192	2.90	253,400	3.32	195,953	2.08	−0.13	0.94 (0.94-0.95)
Z-hypnotics d
Age 10-12y	1,026	0.04	966	0.04	532	0.01	166	<0.01	−0.04	0.09 (0.07-0.10)
13-17y	5,604	0.18	7,577	0.18	4,382	0.08	1,324	0.02	−0.16	0.11 (0.11-0.12)
18-24y	13,763	0.74	23,732	1.05	24,766	0.79	9,146	0.21	−0.53	0.28 (0.28-0.29)
25-29y	14,970	1.69	26,362	2.49	35,647	2.30	16,273	0.60	−1.10	0.35 (0.34-0.36)
Sex
Male	9,481	0.27	15,254	0.34	17,647	0.27	9,579	0.11	−0.16	0.42 (0.41-0.43)
Female	25,882	0.55	43,383	0.77	47,680	0.63	17,330	0.18	−0.37	0.33 (0.33-0.34)
Barbiturates e
Age 10-12y	10,420	0.44	8,057	0.31	4,041	0.11	1,892	0.04	−0.39	0.10 (0.09-0.10)
13-17y	22,379	0.73	22,665	0.54	22,604	0.41	10,759	0.17	−0.56	0.23 (0.23-0.24)
18-24y	33,242	1.80	29,650	1.31	39,329	1.26	24,066	0.55	−1.24	0.31 (0.30-0.31)
25-29y	25,872	2.93	23,952	2.26	37,466	2.42	26,788	0.98	−1.95	0.33 (0.33-0.34)
Sex
Male	23,129	0.66	20,844	0.46	19,511	0.30	12,478	0.15	−0.51	0.22 (0.22-0.23)
Female	68,784	1.47	63,479	1.13	83,929	1.10	51,027	0.54	−0.93	0.37 (0.36-0.37)
Gabapentin f
Age 10-12y	4,882	0.20	1,833	0.07	3,245	0.08	5,174	0.12	−0.09	0.57 (0.55-0.60)
13-17y	12,075	0.39	6,019	0.14	11,454	0.21	17,778	0.28	−0.12	0.71 (0.69-0.72)
18-24y	14,183	0.77	12,339	0.54	31,028	0.99	44,591	1.03	0.26	1.34 (1.31-1.36)
25-29y	13,426	1.52	15,632	1.48	51,819	3.34	81,596	2.99	1.47	1.97 (1.93-2.00)
Sex
Male	19,809	0.56	13,017	0.29	29,584	0.46	54,919	0.65	0.09	1.15 (1.14-1.17)
Female	24,757	0.53	22,806	0.41	67,962	0.89	94,220	1.00	0.47	1.89 (1.86-1.91)

^*CMS Medicaid data, 2001-2019; pooled 43 states; TAF years 2014-2018, data release 2, 2019, data release 1. Proportion out of the full eligible Medicaid population; Confidence intervals (CIs) for proportions are not depicted given the high precision offered by the large sample size. For examples, the 95% CIs for the proportions with smaller numerator or denominator counts include: a) age 10-12 years, 2019, Z-hypnotics: 166/4,398,249, 0.0038% (95%CI 0.0032% to 0.0043%), b) age 10-12 years, 2001, Z-hypnotics: 1026/2,385,053, 0.043% (95%CI: 0.040% to 0.046%), c) age 25-29 years, 2001, Stimulants: 2554/825,588, 0.29% (95%CI: 0.30% to 0.32%)

^aOpioids, average percent change, 2001 to 2019: 10-12y (−0.37), 13-17y (−0.46), 18-24y (−0.78), 25-29y (−0.79)

^bStimulants, average percent change, 2001 to 2019: 10-12y (0.12), 13-17y (0.17), 18-24y (0.13), 25-29y (0.13)

^cBenzodiazepines, average percent change, 2001 to 2019: 10-12y (0.01), 13-17y (0.01), 18-24y (−0.03), 25-29y (−0.09)

^dZ-hypnotics, average percent change, 2001 to 2019: 10-12y (0.00), 13-17y (−0.01), 18-24y (−0.03), 25-29y (−0.06)

^eBarbiturates, average percent change, 2001 to 2019: 10-12y (−0.02), 13-17y (−0.03), 18-24y (−0.07), 25-29y (−0.11)

^fGabapentin, average percent change, 2001 to 2019: 10-12y (0.00), 13-17y (−0.01), 18-24y (0.01), 25-29y (0.08)

The proportion filling 2+ classes of controlled medications in 2019 ranged from 0.5% in early adolescents to 4.8% in young adults (25-29y), (Figure 2; Table 1). The 2019 proportion was the lowest, after peaking at 1.6% in early adolescents in 2003 and 12.3% in young adults in 2010.

Figure 2.

Trends in the annual proportion of Medicaid enrollees prescribed a controlled medication by age group, 2001-2019

1.3.3. Trends by controlled medication type.

The proportions and trends in the prescribing of controlled medications varied by medication class (Figure 2; Table 1). Opioids, BZDs, and Z-hypnotics had overall declines in proportion prescribed by 2019, after increasing from 2001-2010. There was a steady decline in barbiturate prescribing from 2001 to 2019. In contrast, the proportion prescribed stimulants increased, while trends in gabapentin varied by age. Trends were consistent in the sensitivity analyses restricted to 28 states with lower year-to-year fluctuation in estimates and restricting to states that had not expanded Medicaid under the Affordable Care Act by 2019 (eFigure 3).

Early adolescents.

In early adolescents (10-12y), the proportion prescribed an opioid was highest in 2003 at 9.8% and declined to 1.5% by 2019 (Figure 2a; APC:−0.37%). Stimulant prescribing rose from 7.6% in 2001 to 11.2% in 2013, followed by a decline to 9.9% in 2019 (Figure 2b). Stimulant prescribing remained highest in early adolescents compared to adolescents and young adults throughout the study period. BZD (<0.7%), Z-hypnotic (<0.1%), barbiturate (<0.5%), and gabapentin (≤0.2%) prescribing remained low in early adolescents from 2001-2019.

Adolescents.

In adolescents (13-17y), the proportion prescribed an opioid peaked at 14.3% in 2003 and declined to 4.4% in 2019 (APC:−0.46%). From 2001 to 2019, there was an increase in proportion prescribed a stimulant (4.8% to 7.8%), with a peak in 2013 at 8.7%. The proportion prescribed a BZD (<1.2%), Z-hypnotic (<0.3%), barbiturate (<0.8%), or gabapentin (<0.4%) remained low from 2001-2019, with sharp declines in barbiturates and Z-hypnotics (0.2% in 2003-2006 to 0.02% in 2019, Figure 2d). Gabapentin prescribing was low (Figure 2f), while doubling from 2007 (0.14%) to 2019 (0.28%), the peak remained in 2001 (0.39%).

Young adults.

Except for stimulants, the proportion prescribed controlled medications was higher in young adults aged 25–29y than 18-24y, with similar overall trends (Figure 2; Table 1). From 2001 to 2019, stimulant prescribing increased four-fold in those aged 18-24y (0.8% to 3.1%; PR:4.0) and eight-fold in those aged 25–29y (0.3% to 2.6%), the largest relative change (PR:8.9) across ages and substances. Gabapentin prescribing increased from 2001 to 2019 in young adults aged 18-24y (0.8% to 1.0%;PR:1.3) and 25-29y (1.5% to 3.0%;PR:2.0). Z-hypnotic prescribing peaked in 2010 and declined through 2019 in young adults (18-24y:1.3% to 0.2%; 25-29y:3.3% to 0.6%). Similarly, BZD and opioid prescribing peaked in 2010 and declined through 2019. The largest decline in absolute change (−14 percentage-points) across all ages and substances was opioid prescribing in young adults (25-29y). When excluding opioid cough suppressants, opioid trends were consistent with lower proportions prescribed, as were trends excluding barbiturates on the DEA exclusion list (eFigure 4).

1.4. DISCUSSION

Over the past two decades, there was an overall decline in the proportion of young Medicaid enrollees with a controlled medication filled. In young adults (25-29y), for example, the prescribing of a controlled substance peaked at 47.0% in 2009 and declined to 23.9% by 2019. However, trends varied substantially by specific substance and age group. There were increases in stimulant prescriptions, steady declines in barbiturate prescriptions, and increases followed by declines for opioids, BZDs, and Z-hypnotics. As young people are a vulnerable age group with a heightened risk for drug misuse and overdose, it is important to monitor trends in the prescribing of controlled medications.

Our observed overall declines in controlled medication prescription fills align with declines in self-reported medical use of opioids, stimulants, and BZDs in adolescents from 2009-2010 vs. 2021-2022 (McCabe et al., 2024). However, we observed important differences across controlled medication classes, most notably an increase in stimulant prescriptions. The proportion with a stimulant prescription was highest in early adolescents but increased across age groups, with an 8-fold increase from 2001 to 2019 in Medicaid enrollees aged 25–29y. An increase in stimulant prescriptions has been reported across children and adults in the US (Bykov et al., 2020; Danielson et al., 2023; Han et al., 2025; IQVIA Government Solutions Inc, 2022; Meadows et al., 2020), and aligns with the increasing prevalence of attention-deficit/hyperactivity disorder (ADHD) diagnoses (Kazda et al., 2021), which has leveled-out in early adolescents in recent years (Lebrun-Harris, Ghandour, Kogan, & Warren, 2022; Li et al., 2023). Stimulant medications are the first-line medications for ADHD with demonstrated short-term efficacy for children and adults (Cortese et al., 2018; Posner, Polanczyk, & Sonuga-Barke, 2020). However, the misuse of stimulant medications is common in certain populations (Faraone et al., 2020; Han et al., 2025) and stimulants are involved in fatal overdoses (Kariisa, Scholl, Wilson, Seth, & Hoots, 2019). In one study, 39% of youth with a medically treated stimulant-involved overdose had a prior stimulant prescription (Bushnell GA, Samples H, Gerhard T, Calello DP, & Olfson M, 2022). Therefore, when prescribing stimulants, cautious prescribing, monitoring for misuse, and patient and caregiver education are recommended (Colaneri, Keim, & Adesman, 2017; Hadland & Marshall, 2021).

Almost 1 in 10 early adolescents and 1 in 3 young adults in Medicaid had an opioid prescription fill during the years of peak opioid prescribing in these data. The proportion with an opioid prescription then declined, which has been observed in the broader US population (Bykov et al., 2020; Guy et al., 2017; Meadows et al., 2020; Paiva et al., 2023; Renny, Yin, Jent, Hadland, & Cerda, 2021). The opioid overdose epidemic and corresponding safety concerns, along with changes in opioid prescribing policies, likely contributed to the declines in opioid prevalence in this young population. Of note, despite reductions in opioid prescriptions to early adolescents through 2019, the proportion of Medicaid enrollees with an opioid prescription (2019, 1.5% in early adolescents) remains substantially higher than other countries (e.g., Norway, opioid annual prevalence <0.06% in children) (Stangeland et al., 2023). The appropriate prescribing of opioids to young people continues to require detailed patient and family evaluation, treatment planning, and regular monitoring (Preuss et al., 2022).

We observed declining BZD and Z-hypnotic prescribing in youth enrolled in Medicaid beginning around 2010, consistent with prior literature on increases during the early 2000s followed by declines (Bertisch, Herzig, Winkelman, & Buettner, 2014; Bykov et al., 2020; Ford et al., 2014; Ma et al., 2023; Stangeland et al., 2023; Toce, Michelson, Hudgins, Olson, & Bourgeois, 2021). The decline in BZDs corresponded with declines in opioid prescribing and in opioid and BZD co-prescribing following the 2016 FDA boxed warnings on the risk of overdose with concurrent opioid and BZD use (U.S. Food and Drug Administration, 2016). The decline in Z-hypnotic prescriptions to publicly insured adolescents and young adults was substantial and likely related to increased awareness of safety concerns. In 2007, the FDA asked manufacturers to strengthen the language on Z-hypnotic product labels surrounding the risk of complex sleep-related behaviors; in 2019, a boxed warning was added (Stump E, 2007)(U.S. Food and Drug Administration, 2019). The declines in barbiturate prescribing were steady across the study period. The use of barbiturates began to decline with the introduction of BZDs and other sedative and hypnotic medications that offered safer alternatives to the misuse concerns surrounding barbiturates (Lader, 1991). Overall, BZD, Z-hypnotic, and barbiturate prescribing was low in Medicaid enrolled early adolescents and adolescents, likely indicative of more cautious prescribing for ages with limited evidence of treatment efficacy.

While gabapentin prescribing remained relatively low in early adolescent and adolescent Medicaid enrollees, the proportion of young adults prescribed gabapentin considerably increased, particularly in those aged 25-29y (1.5% to 3.0%). Increases in gabapentin (Peet et al., 2023; Stangeland et al., 2023) are related to expanding use for a variety of indications, including pain management and mental health conditions such as bipolar disorder, anxiety disorders, and insomnia, with varied evidence (Dauri et al., 2009; Hong et al., 2022; Wiffen et al., 2017). There are growing concerns surrounding the misuse of gabapentin (Campbell et al., 2021; Ispir, Danisman, & Katar, 2023). By 2019, 7 states classified gabapentin as a Scheduled V controlled substance and mandated gabapentin reporting to the state PDMP, and an additional 10 states mandated gabapentin reporting to the PDMP but did not classify gabapentin as a controlled substance (Karavolis et al., 2022).

Opioids, stimulants, BZDs, barbiturates, Z-hypnotics, and gabapentin can be prescribed in clinically appropriate situations in young populations. However, the use and misuse of controlled medications is associated with increased risk of substance use disorders (McCabe et al., 2019), particularly for misuse at younger ages (McCabe et al., 2007). A third (37%) of adolescents recently reported that the most common source of their prescription drug misuse was their own prescription (McCabe et al., 2024), highlighting the importance of educating clinicians and patients on risk assessment and minimization. This can include monitoring and screening for misuse and signs of substance use disorders at follow-up, educating the patient on the potential for addiction to the medications, and instructions for safe storage and disposal of medications (Hadland & Marshall, 2021; McCabe et al., 2019; Preuss et al., 2022). With clinical recommendations to only prescribe controlled medications to youth when necessary, it can be difficult to weigh the potential risks of misuse against the risks of undertreating a medical condition, as untreated psychiatric symptoms and pain can lead to serious long-term problems (de Zambotti, Goldstone, Colrain, & Baker, 2018; Pielech, Lunde, Becker, Vowles, & Sieberg, 2020; Preuss et al., 2022; Stangeland et al., 2023).

There are limitations to consider. 1) There were changes in the base Medicaid enrolled population during our study period (2001-2019), particularly an increase in eligible adults (19-64 years) following Medicaid expansion (Guth M, Garfield R, Rudowitz R, & Kaiser Family Foundation, 2020). However, results of our sensitivity analysis in states that did not expand Medicaid were consistent with trends shown in the larger group of states, for both adolescents and young adults. 2) Medicaid data quality varies by state-year. However, our sensitivity analysis limited to states with smaller year-to-year fluctuations for the full study period yielded results that were consistent with results from the full group of states. 3) Our definition required only one prescription in the calendar year; therefore, youth with one prescription were treated equivalently to those with longer medication use. Further, we did not examine trends in new use, long-term use, dosage, or days of supply. Such analyses were beyond the scope of the present study, but should be undertaken in future research. 4) The data end in 2019 due to delays in national Medicaid data releases and therefore do not inform trends during the COVID-19 era. Nevertheless, we provide two decades of trends in a 43-state Medicaid sample. 5) While we observed overall declines in the proportion of youth with 2+ classes of controlled medications filled within a year, further investigation is needed to consider whether prescriptions overlapped, a situation that may increase risks.

We present trends in prescription fills for controlled medication in early adolescents, adolescents, and young adults enrolled in Medicaid from 2001 to 2019. These findings highlight large shifts in prescribing during this period as well as considerable age heterogeneity. Multiple factors likely contributed to changes in opioid, stimulant, BZD, Z-hypnotic, barbiturate, and gabapentin prescribing including changes in state and federal policies, increased recognition of safety concerns, new treatments entering the market, broadening medical indications, and shifts in patient and clinician preferences. The prescribing of controlled medications to young populations continues to requires caution and monitoring given the potential risk for drug misuse and overdose.