Prescription Stimulant Dispensing to US Children: 2017–2023

Attention-deficit hyperactivity disorder (ADHD) is the most diagnosed neurobehavioral condition among US children, with a lifetime prevalence of 11.4%.^1–3 Prescription stimulants are the first-line treatment for ADHD in children when pharmacotherapy is required.⁴ Due to the prevalence of ADHD and the importance of stimulant therapy in its management, stimulants are the most commonly prescribed controlled substance to US children.^5,6 

Two relatively recent events may have changed the rate of stimulant dispensing to US children. First, when the COVID-19 pandemic began in March 2020, many children transitioned to virtual learning, which may have changed the degree to which children used stimulants to manage ADHD.^7,8 Additionally, access to in-person health care was disrupted, potentially decreasing opportunities to diagnose and treat ADHD,^9,10 although such decreases may have been offset by a policy allowing stimulant prescribing via telehealth without an in-person visit.¹¹ Second, a nationwide shortage of one of the most commonly prescribed stimulants, immediate-release mixed amphetamine salts (Adderall or Adderall IR), was announced by the US Food and Drug Administration (FDA) in October 2022.¹² This shortage may have caused some patients to discontinue use of immediate-release mixed amphetamine salts or to seek treatment with different stimulant medications.

National data are lacking on the effect of these 2 events on stimulant dispensing to children. One analysis of commercially insured enrollees aged 5 to 64 years reported sharp increases in stimulant dispensing to adolescent girls and adult women during 2020 to 2021, but this study did not include publicly insured patients or capture the period during which the shortage of immediate-release mixed amphetamine salts began.¹³ According to an analysis of 2018 to 2022 data from a national prescription dispensing database, the number of stimulant prescriptions for patients new to therapy increased during the pandemic, but this analysis was limited to incident stimulant prescriptions and data from before October 2022.¹⁴ An analysis of the same national database documented large increases in stimulant dispensing from 2012 to 2022, but this analysis did not focus on children or evaluate the effect of stimulant shortages.¹⁵ Finally, a non–peer-reviewed analysis of a national electronic health record database found that the monthly prescription fill rate for mixed amphetamine salts decreased after October 2022, but this study did not examine whether the decrease in dispensing of mixed amphetamine salts was offset by increases in dispensing of other stimulants.¹⁶ 

Using 2017 to 2023 data from an all-payer national database, we examined changes in the prescription stimulant–dispensing rate to children aged 5 to 17 years after March 2020 and October 2022. To contextualize results, we repeated analyses among subgroups defined by age and sex and examined changes in dispensing of specific stimulant medications.

During March to April 2024, we analyzed the 2017 to 2023 IQVIA Longitudinal Prescription Database. This database reports unprojected, all-payer, prescription-level data from 92% of US retail pharmacies and approximately 70% of mail-order and long-term care pharmacies. Key data elements include an encrypted patient identifier, national drug codes, and patient sex, year of birth, and state of residence. Because the data are deidentified, the Institutional Review Board of the University of Michigan Medical School exempted analyses from human subjects review.

We identified prescriptions for immediate-release and extended-release stimulants dispensed to children aged 5 to 17 years from January 1, 2017, through December 31, 2023.^17,18 We excluded stimulant prescriptions dispensed to children aged 0 to 4 years, who rarely use stimulants. Stimulants included amphetamine, dexmethylphenidate, dextroamphetamine, lisdexamfetamine, methamphetamine, methylphenidate, and serdexmethylphenidate.

The primary outcome was the monthly stimulant-dispensing rate, defined as the number of children aged 5 to 17 years with 1 or more dispensed stimulant prescription per 100 000 US children. We obtained population denominators from the US Census Bureau (Supplemental 1).¹⁹ Secondary outcomes were the monthly dispensing rate for 6 commonly prescribed stimulants: immediate-release mixed amphetamine salts, extended-release mixed amphetamine salts, lisdexamfetamine, methylphenidate, dexmethylphenidate, and all other stimulants. Secondary outcomes also included the monthly dispensing rate for atomoxetine, viloxazine, and guanfacine, 3 commonly prescribed nonstimulant medications for ADHD.⁴ 

For all outcomes except the monthly viloxazine dispensing rate, we conducted an interrupted time series analysis in which we fitted linear segmented regression models assessing for level changes (ie, abrupt changes in the intercept of the fitted line) or slope changes (ie, changes in the trajectory of the fitted line) in outcomes during March 2020 and October 2022.²⁰ The counterfactual for the level and slope changes in March 2020 was based on the fitted line from January 2017 to February 2020, whereas the counterfactual for the level and slope changes in October 2022 was based on the fitted line from March 2020 to September 2022. To facilitate interpretation of results, we calculated the difference between the observed outcome in December 2023 from the outcome that would have occurred had pre–March 2020 trends continued through this month. The latter outcome was calculated by extrapolating the fitted line from January 2017 to February 2020 until December 2023. When assessing changes in the monthly viloxazine dispensing rate, we began analyses in April 2021, the date this drug was approved, and only assessed for level and slope changes in October 2022.

To control for monthly variation in dispensing and improve model fit, all segmented regression models included an indicator that equaled 1 for June or July and 0 otherwise. We included this indicator because stimulant dispensing consistently decreased during these 2 months, likely owing to the decreased need for stimulants when children are not attending school. To assess for autocorrelation (ie, temporal correlation between data points), we used the Cumby-Huizinga test. We accounted for any autocorrelation using robust Newey-West standard errors with the appropriate number of lags.²¹ 

To facilitate visualization of shifts in dispensing of specific stimulants, we calculated the unadjusted percentage change in the 6 stimulant-related secondary outcomes between each month during 2020 to 2023 and the mean outcome in the corresponding calendar month during 2017 to 2019. We also conducted subgroup analyses in which we repeated the interrupted time series analyses among young boys and girls aged 5 to 11 years and among adolescent boys and girls aged 12 to 17 years. In these subgroup analyses, we obtained subgroup-specific denominators from the US Census Bureau¹⁹ and excluded a small number of individuals with unknown sex. All analyses used R version 4.2.3 and 2-sided hypothesis tests with α = 0.05.

To evaluate whether conclusions regarding the effect of the shortage were robust to analytic decisions, we conducted 2 sensitivity analyses in which we limited to data from 2021 to 2023 or 2022 to 2023 and only assessed for level and slope changes in the monthly stimulant-dispensing rate to children in October 2022.

From January 2017 through December 2023, the monthly stimulant-dispensing rate decreased from 2989.2 to 2749.4 children per 100 000 (−8.0%; Figure 1). Table 1 shows coefficients from segmented regression models. In January 2017, the predicted monthly stimulant-dispensing rate (intercept) was 2422.2 children per 100 000 (95% CI, 2322.5–2521.9). Between then and March 2020, this rate changed little (slope, −0.8 children per 100 000 per month; 95% CI, −4.4 to 2.8). In March 2020, there was a level decrease (−454.9 children per 100 000; 95% CI, −572.6 to −337.2, or an 18.8% decrease relative to the intercept) and a slope increase (12.7 children per 100 000 per month; 95% CI, 6.6–18.8). In October 2022, there was no level change (−39.7 children per 100 000; 95% CI, −189.9 to 110.5) or slope change (−12.1 children per 100 000 per month; 95% CI, −27.5 to 3.3), although both point estimates were negative. In December 2023, the stimulant-dispensing rate was 6.6% lower than predicted had prepandemic trends continued.

Among young children aged 5 to 11 years, the monthly stimulant-dispensing rate was higher among boys than girls in January 2017 (Figure 2A). In March 2020, there was a level decrease and slope increase in both groups. In contrast, there was no level or slope change in either group in October 2022. In December 2023, the monthly stimulant-dispensing rate was 0.6% lower than predicted by the prepandemic trend among young boys and 9.0% higher than predicted among young girls.

Among adolescents aged 12 to 17 years, the monthly stimulant-dispensing rate was also higher among boys than girls in January 2017 (Figure 2B). In March 2020, there was a level decrease and slope in both groups. In October 2022, there was no level change in either group, but there was a slope decrease in both groups (boys: −35.4 children per 100 000 per month; 95% CI, −67.9 to −3.0; girls: −19.9 children per 100 000 per month; 95% CI, −32.1 to −7.8). In December 2023, the monthly stimulant-dispensing rate was 18.8% lower than predicted by the prepandemic trend among adolescent boys and 2.8% lower than predicted among adolescent girls.

Figure 3 shows the difference in the 6 stimulant-related secondary outcomes between each month in 2020 to 2023 and the corresponding calendar months in 2017 to 2019. Table 2 shows coefficients from segmented regression models for these 6 outcomes. For all 6 outcomes, there was a level decrease and slope increase in March 2020 (Figures 4A–4F).

For immediate-release mixed amphetamine salts, there was a large level decrease in the monthly dispensing rate in October 2022 (−64.0 children per 100 000; 95% CI, −100.1 to −27.9, a 13.2% decrease relative to the intercept), but there was no slope change. In December 2023, the dispensing rate for immediate-release mixed amphetamine salts was 18.7% lower than predicted had prepandemic trends continued.

For dexmethylphenidate, there was a level increase (34.8 children per 100 000; 95% CI, 8.8–60.8, a 11.7% increase relative to the intercept) and a slope decrease (−5.0 children per 100 000 per month; 95% CI, −7.7 to −2.4) in the monthly dispensing rate in October 2022. For extended-release mixed amphetamine salts, there was no level change in the monthly dispensing rate in October 2022, but there was a slope decrease (−6.6 children per 100 000 per month; 95% CI, −8.5 to −4.6). For lisdexamfetamine, methylphenidate, and all other stimulants, there were no level or slope changes in the monthly dispensing rate in October 2022.

For atomoxetine and guanfacine, there were level decreases and slope increases in the monthly dispensing rate in March 2020. For these drugs as well as for viloxazine, there was no level change in the monthly dispensing rate in October 2022, although there were modest slope changes for atomoxetine and viloxazine (Supplemental 2–4).

In both sensitivity analyses, the level and slope change in the monthly stimulant-dispensing rate to children in October 2022 was negative but not significant, similar to the main analysis (Supplemental 5–6).

Using 2017 to 2023 data from a comprehensive, all-payer national database, we evaluated changes in the monthly prescription stimulant–dispensing rate to children aged 5 to 17 years after March 2020 and October 2022. There were 3 main findings. First, the stimulant-dispensing rate to children in December 2023 was 6.6% lower than the rate predicted by prepandemic trends because this rate never fully recovered after the large decline in March 2020. Second, the monthly stimulant-dispensing rate changed in a heterogeneous fashion by age and sex, with a particularly large decrease among adolescent boys. Finally, the shortage of immediate-release mixed amphetamine salts was associated with a decline in dispensing of this medication but not with a decline in the overall stimulant-dispensing rate, partly because dispensing of other stimulants increased.

Our finding that the monthly prescription stimulant–dispensing rate to children declined sharply during March 2020 is consistent with prior studies.^6,14 This decline may represent disruptions in access to therapy and/or decreased need of prescribed stimulants to manage ADHD. After the pandemic started in March 2020, the monthly stimulant-dispensing rate began to recover but had not returned to the level predicted by prepandemic trends by December 2023. This finding stands in contrast to another study of IQVIA data, which found that the number of incident stimulant prescriptions dispensed to patients of all ages were above prepandemic levels by the end of the study period in March 2022.¹⁴ The discrepant findings may derive from our inclusion of all stimulant prescriptions, not just incident prescriptions, as well the fact that trends in stimulant dispensing to adults and children have diverged, with large increases among the former and much smaller changes in the latter.¹⁹ 

Changes in the stimulant-dispensing rate to children varied by age and sex. For boys aged 5 to 11 years, the stimulant-dispensing rate in December 2023 was similar to the levels predicted by prepandemic trends. In contrast, this rate was 9.0% higher than predicted among girls aged 5 to 11 years, consistent with a prior study showing that children with first-time diagnoses of ADHD were more likely to be girls during the pandemic compared with before the pandemic.¹⁰ A potentially optimistic explanation of these findings is that they represent a decrease in underdiagnosis and undertreatment of ADHD in girls.^22,23 Alternatively, it is possible that symptoms of inattention in girls increased during the pandemic, thus increasing their need for stimulants.

Among adolescent girls, the stimulant-dispensing rate was only slightly below the level predicted by prepandemic trends in December 2023, but it was 18.8% below the predicted level among adolescent boys. The latter finding is consistent with a prior analysis showing that antidepressant dispensing among adolescent boys fell after March 2020 and remained well below the prepandemic baseline afterwards.¹⁸ Decreases in dispensing of stimulants and antidepressants to adolescent boys are unlikely to reflect improved mental health, given evidence that the mental health of adolescent boys worsened during the pandemic.^19,24 A more plausible explanation is that adolescent boys were less likely to seek care for mental health symptoms during the pandemic, potentially increasing their risk of adverse events associated with poor mental health. This possibility should be urgently explored.

The emergence of a nationwide shortage of immediate-release mixed amphetamine salts in October 2022 was not accompanied by a significant level or slope change in the monthly stimulant-dispensing rate to children aged 5 to 17 years. This finding suggests that the shortage did not substantially prevent children from starting stimulant therapy or prompt existing patients from stopping stimulant therapy.

However, any reassurance from this finding must be tempered for several reasons. First, although not significant, the level and slope changes in the monthly stimulant-dispensing rate in October 2022 were negative, suggesting the shortages may have inhibited stimulant dispensing to some degree. Second, there was a slope decrease in the monthly stimulant-dispensing rate among adolescent boys and girls after October 2022, suggesting the shortages may have disrupted stimulant therapy in certain subgroups of patients. Third, we lacked information on the degree to which the shortage may have prevented children who needed stimulants from obtaining them. If the number of such children increased in October 2022, the lack of change in the monthly stimulant-dispensing rate would imply that the shortage increased stimulant underuse. Finally, there was a level increase in the monthly dispensing rate of dexmethylphenidate and a slope increase in the monthly dispensing rate of atomoxetine in October 2022, suggesting that some children may have switched from immediate-release mixed amphetamine salts to other stimulant medications or to nonstimulants. These switches could be concerning if children did not respond as well to the alternative medications, if the switches imposed a substantial time burden on families and clinicians,²⁵ or if the switches increased the risk of medication errors, a possibility supported by prior research on the effects of drug shortages.^26,27 

Our findings have clinical and policy implications. In its announcement of the shortage of immediate-release mixed amphetamine salts in October 2022, the FDA pointed both to production issues and to unprecedented demand for stimulants in the United States.¹² Our finding that stimulant dispensing to children was below levels predicted by prepandemic trends in December 2023, coupled with prior studies showing large recent increases in stimulant dispensing to adults,¹³ suggests that the increased demand for stimulants was concentrated among adults. Thus, our findings indirectly demonstrate how increased demand for a medication class in adult patients may adversely affect access to the medication class in children.

Ameliorating shortages of stimulants among children is critical given the large number of children who use these medications daily and the deleterious consequences of uncontrolled or undertreated ADHD symptoms for child health and educational outcomes.²⁸ To achieve this goal, it is important to ensure the appropriateness of stimulant prescribing to adults, including developing guidelines for the diagnosis and management of ADHD in adults.²³ Moreover, the US Drug Enforcement Administration, which limits stimulant production by setting quotas,¹² might consider increasing quotas to ensure that children do not experience barriers to accessing stimulants even if demand for stimulants among adults continues to surge. Quotas should not only be increased for specific stimulants, such as immediate-release mixed amphetamine salts. For example, our study demonstrated a large slope decrease in the monthly dispensing rate of extended-release mixed amphetamine salts in October 2022, potentially because this medication was also in shortage at the same time.²⁹ 

This study has limitations. First, we assumed the effect of the shortage of immediate-release mixed amphetamine salts occurred in October 2022, even though this effect likely occurred more gradually. Second, analyses did not examine the effects of shortages of stimulants other than immediate-release mixed amphetamine salts. Third, the IQVIA database does not report race and ethnicity. As such, we could not assess whether the COVID-19 pandemic or shortage of mixed amphetamine salts exacerbated preexisting disparities in the treatment of ADHD.³⁰ 

This national analysis suggests that the monthly stimulant-dispensing rate to US children markedly declined after the COVID-19 pandemic began. This rate may also have declined after the shortage of immediate-release mixed amphetamine salts was announced in October 2022, but estimates of this change were not significant, in part because the decline in dispensing of immediate-release mixed amphetamine salts was offset by increased dispensing of other stimulants. Findings suggest the shortage may have prompted some children to switch from immediate-release mixed amphetamine salts to alternative medications. Future studies evaluating whether any such switches worsened ADHD-related outcomes might help motivate policy action to prevent future shortages of stimulants.