Changes in Stimulant Dispensing to US Children Associated With the Stimulant Shortage

Stimulants are highly effective for controlling symptoms of attention-deficit/hyperactivity disorder (ADHD) in children.1 In October 2022, the Food and Drug Administration announced a shortage of the brand and generic formulations of immediate-release (IR) mixed amphetamine salts (Adderall IR), a commonly prescribed stimulant.2 In our prior analysis of aggregated prescription dispensing data, the monthly proportion of US children with dispensing of any stimulant did not change after this announcement, although the proportion with Adderall IR dispensing decreased and the proportion with dispensing of other stimulants increased.3These findings indirectly suggest the shortage did not increase stimulant discontinuation but did prompt some children to switch from Adderall IR to other stimulants. However, confirmation of this possibility requires patient-level rather than aggregated data. We evaluated whether the Adderall IR shortage was associated with changes in stimulant dispensing to children using this medication.We conducted a difference-in-differences analysis of the IQVIA Longitudinal Prescription Database, which captures 92% of prescriptions dispensed from US pharmacies across all payment types, including cash.4 Because data were deidentified, the University of Michigan Institutional Review Board exempted analyses from review.We included children aged 5 to 17 years with stimulant dispensing during quarter 3 of 2020, 2021, or 2022. Children with Adderall IR dispensing during these “index quarters” were assigned to the treatment group, even if they also had dispensing of other stimulants. Children who only had dispensing of stimulants other than Adderall IR were assigned to the control group.Primary outcomes were having no stimulant dispensing, any Adderall IR dispensing, and any dispensing of a stimulant other than Adderall IR during the 9 months after the index quarter (eg, October 2022 through June 2023 for the quarter 3 2022 cohort). For additional context, we also evaluated dispensing of any nonstimulant ADHD medication (atomoxetine or viloxazine) during this period. Using linear models, we compared changes in outcomes between the quarter 3 2020/2021 cohorts and quarter 3 2022 cohort among the treatment and control groups. We controlled for age, sex, payer type, state, and year. Models employed SEs clustered at the patient level.We conducted subgroup analysis among children aged 5 to 11 years vs 12 to 17 years and among Medicaid-insured vs commercially insured children. Analyses used R version 4.4.1 and 2-sided hypothesis tests with α = 0.05. Supplemental 1 includes additional details.The treatment group included 1 337 825 person-years of data from 871 841 children, of whom 486 282 (55.8%) were aged 12 to 17 years. The control group included 5 171 742 person-years of data from 2 975 586 children, of whom 1 482 132 (49.8%) were aged 12 to 17 years.Figure 1 shows unadjusted outcomes for each cohort. Compared with the control group, the risk of having no stimulant dispensing during the 9 months after the index quarter increased to a greater degree in the treatment group (1.2 percentage points [pp]; 95% CI, 1.0–1.3), while the risk of having Adderall IR dispensing decreased to a greater degree (−1.9 pp; 95% CI, −2.1 to −1.8). The risk of having dispensing of other stimulants increased to a greater degree (8.0 pp; 95% CI, 7.8–8.1), as did the risk of having dispensing of nonstimulants (0.08 pp; 95% CI, 0.03–0.1). There was a minor violation of the parallel trends assumption for 2 outcomes (Supplemental 2).5The direction of difference-in-differences estimates was unchanged in subgroup analyses (Table 1; Supplementals 3 and 4). The magnitude of these estimates was slightly larger in Medicaid-insured compared with commercially insured children. Estimates varied little by age group.Among children using Adderall IR before the shortage was announced, the proportion without stimulant dispensing during the next 9 months increased slightly more compared with children using other stimulants. The proportion with dispensing of Adderall IR decreased to a greater degree, while the proportion with dispensing of other stimulants increased to a greater degree. The proportion with dispensing of nonstimulants increased minimally.Findings directly demonstrate that the shortage prompted some children to switch from Adderall IR to other stimulants. The shortage also prompted some children using Adderall IR to discontinue stimulant therapy, particularly Medicaid-insured children. Fortunately, the increase in stimulant discontinuation was small.Limitations include minor violations of the parallel trends assumption. Future studies should assess the administrative burden on prescribers associated with the shortage and evaluate whether switching from Adderall IR to alternative treatments impacted ADHD symptomology and medication error rates.6,7