Healthcare Costs Associated with Adverse Events in Pediatric Patients with Attention-Deficit/Hyperactivity Disorder (ADHD): A Claims-Based Study

Abstract

Introduction

Adverse events (AEs) are common in pediatric patients receiving attention-deficit/hyperactivity disorder (ADHD) treatment; however, real-world studies on their costs from a payer’s perspective are lacking. Therefore, this study investigated the healthcare costs associated with selected AEs among pediatric patients receiving ADHD treatment in the United States.

Methods

Using a retrospective cohort design, patients aged 6–17 years who received pharmacologic treatment for ADHD were identified from US claims data (October 1, 2015–September 30, 2023) and were categorized into AE and AE-free cohorts, separately for each studied AE. The eight selected AEs had statistically significant risk differences in a matching-adjusted indirect comparison of ADHD treatments and were identifiable from claims with ICD-10-CM codes. Entropy balancing was used to create cohorts with similar characteristics. Total excess healthcare costs and costs associated with AE-specific claims per patient per month (PPPM) were compared across balanced cohorts with vs. without a given AE.

Results

Overall, 393,919 patients (mean age: 12.5 years; male: 65.4%; stimulant monotherapy: 71.8%) were included, among whom 13.6% had ≥ 1 studied AE that resulted in a medical encounter during their treatment episode. The most prevalent AEs were upper abdominal pain (5.2%), vomiting (3.4%), and insomnia (3.2%). All AEs were associated with substantial AE-specific costs PPPM (asthenia: $196; somnolence: $171; insomnia: $169; vomiting: $106; dizziness: $92; upper abdominal pain: $91; irritability: $75; decreased weight: $46) and total excess healthcare costs PPPM (asthenia: $1178; somnolence: $821; vomiting: $427; insomnia: $404; dizziness: $380; upper abdominal pain: $336; irritability: $231; decreased weight: $219; all p < 0.01).

Conclusions

AEs were common during ADHD treatment episodes in pediatric patients and were associated with significant healthcare costs. ADHD treatments with a favorable safety profile could help alleviate the economic burden of AEs.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12325-025-03345-x.

Key Summary Points

Why carry out this study?

Although adverse events (AEs) are common in pediatric patients receiving attention-deficit/hyperactivity disorder (ADHD) treatment, real-world studies on their costs from a payer’s perspective are lacking.

To bridge this gap in knowledge, this retrospective cohort study used claims data (October 1, 2015–September 30, 2023) to investigate the healthcare costs associated with eight selected AEs among 393,919 pediatric patients (6–17 years old) receiving pharmacologic treatment for ADHD in the United States.

What was learned from this study?

More than one in eight patients had at least one AE that resulted in a medical encounter during their treatment episode, with abdominal pain, vomiting, and insomnia being the most common.

AEs experienced during treatment episodes were associated with significant healthcare costs, with asthenia and somnolence accounting for the highest total excess healthcare costs as well as the highest AE-specific medical costs.

This study quantifies the substantial economic burden associated with AEs experienced during ADHD treatment in pediatric patients, underscoring the importance of ADHD treatment options that can help reduce these AEs.

Introduction

Attention-deficit/hyperactivity disorder (ADHD) is characterized by long-lasting symptoms of inattention, hyperactivity, and impulsivity that affect executive function, emotional regulation, and motivation [1]. With an estimated prevalence of 11.4% among individuals aged 3–17 years in 2022, ADHD represents the most common neurodevelopmental disorder in children in the United States [2]. Among those affected, nearly 60% are diagnosed with moderate or severe ADHD, and just over half receive pharmacologic treatment for the condition [2].

Treatment for pediatric patients with ADHD is multidisciplinary, encompassing both pharmacologic and non-pharmacologic approaches [3]. Pharmacologic approaches include stimulants (e.g., amphetamines or methylphenidates) and non-stimulants (e.g., atomoxetine or guanfacine) [4], both of which are associated with adverse events (AEs) such as decreased appetite, abdominal pain, headaches, sleep disturbances, dizziness, and irritability [3]. The implications of AEs for the management of ADHD may be far-reaching as previous literature has shown that they can negatively impact the quality of life outcomes in both pediatric patients and their parents/caregivers [5] and commonly lead to treatment changes [6]. Furthermore, patients requiring treatment modifications may incur additional healthcare resource utilization and costs [7]. Given these implications, novel treatments that have a more tolerable safety profile compared to existing options may help to alleviate the burden of AEs associated with ADHD treatment.

Although real-world insights into treatment patterns, healthcare resource utilization, and costs associated with ADHD in children exist [6–9], evidence on healthcare costs resulting from AEs during treatment episodes remains limited. A recent matching-adjusted indirect comparison (MAIC) study compared the incidence of AEs between centanafadine, an investigational norepinephrine-dopamine-serotonin reuptake inhibitor, and lisdexamfetamine dimesylate (Vyvanse^®), atomoxetine hydrochloride (Strattera^®), viloxazine extended-release (Qelbree^®), guanfacine extended-release (ER) (Intuniv^®), methylphenidate ER tablet (Concerta^®), and methylphenidate ER capsule (Jornay^®); however, information on the costs associated with these AEs has yet to be described [10]. Given the availability of numerous treatment options, an evaluation of the impact of AEs across ADHD treatments is warranted to provide stakeholders with ample evidence to make informed decisions on a patient- and healthcare-system level. Previously, a retrospective cohort study in adult patients receiving pharmacologic treatment for ADHD found that AEs were associated with significant healthcare costs, including inpatient, outpatient, and emergency room services [11]. Using a similar methodology, this study sought to investigate the healthcare costs from a payer’s perspective associated with selected AEs that resulted in medical care among pediatric patients receiving treatment for ADHD in the United States.

Methods

Data Source

This study used data from the IQVIA PharMetrics^® Plus database from October 1, 2015, to September 30, 2023. The database contains fully adjudicated medical and pharmacy claims for > 210 million unique enrollees, including 38 million pediatric patients. Variables available in the database included demographic information, inpatient and outpatient diagnoses and procedures, inpatient stays, prescription fills, plan type, monthly indicators of health plan enrollment, and the amount paid by health plans to the provider for rendered services. IQVIA data were de-identified and complied with the patient requirements of the Health Insurance Portability and Accountability Act (HIPAA), specifically, 45 CFR § 164.514. No institutional review board exemption was requested.

Study Design

This retrospective cohort study evaluated the incremental costs associated with AEs during ADHD treatment. Specifically, healthcare costs incurred among patients with a given AE during an ADHD treatment episode were compared to costs incurred among patients without that AE during a similar ADHD treatment episode (Fig. 1). For each patient, a newly initiated treatment regimen with ≥ 1 US Food and Drug Administration (FDA)-approved agents for the treatment of ADHD was randomly selected. The date of treatment initiation was defined as the index date. Capturing a randomly selected treatment initiation date allowed for the selection of a representative real-world sample of patients across different disease durations and severities. The baseline period was defined as the 12 months preceding the index date. This period served as a washout period to identify treatment initiation and to assess patient characteristics. The study period covered the entirety of the randomly selected treatment episode, from the index date to the earliest of (1) a treatment regimen change (i.e., discontinuation, drop, switch, or add-on), (2) end of continuous health plan enrollment, or (3) end of data availability. Healthcare costs were assessed during this period. Although the study period was not required to have a minimum duration, patients were required to have continuous health plan enrollment during the 12-month baseline period and for a minimum of 6 months following the index date. This requirement ensured a sufficiently long period of observation to capture any AEs reported by patients at follow-up appointments after treatment initiation and thus helped to reduce the chances of misclassification of patients in a given AE cohort.

Fig. 1.

Study design. ADHD attention-deficit/hyperactivity disorder

AEs investigated in this study were upper abdominal pain, vomiting, insomnia, decreased weight, dizziness, irritability, asthenia, and somnolence. These eight AEs were sourced from an MAIC of centanafadine and common pediatric ADHD treatments (i.e., lisdexamfetamine dimesylate atomoxetine hydrochloride, viloxazine extended-release, guanfacine ER, methylphenidate ER tablet, and methylphenidate ER capsule) [10]. Specifically, AEs were selected if there was a statistically significant risk difference between centanafadine and any of the comparators in the MAIC and if they were identifiable using International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes. Based on these criteria, decreased appetite and sedation were excluded because they were not identifiable with ICD-10-CM codes despite a statistically significant difference across treatments in the MAIC analyses. Although dry mouth fulfilled the AE selection criteria, the sample of patients identified was very small (n = 107). As pediatric patients may have difficulty identifying and expressing discomfort from dry mouth, it is likely that medical care would not be sought for this AE in most cases. Moreover, a portion of dry mouth cases identified from claims may have been driven by other conditions or factors, such as chromosomal anomalies (e.g., Down syndrome [12]), which were more than 15 times higher among those with vs. without dry mouth in our sample. Considering this and the small sample, dry mouth was excluded from the analysis as the results may not have been robust.

Sample Selection and Study Cohorts

Patients who met the eligibility criteria (1) had ≥ 2 recorded ADHD diagnoses on distinct dates, (2) had ≥ 1 prescription fill for an FDA-approved pharmacologic ADHD treatment on or following the first observed diagnosis date for ADHD, (3) had continuous health plan enrollment for ≥ 12 months before the index date (i.e., baseline period) and ≥ 6 months following the index date, and (4) were aged 6–17 years old as of the index date (Fig. 2). Patients included in the study were categorized into AE or AE-free cohorts for each selected AE. Patients in a given AE cohort had ≥ 1 diagnosis for the studied AE during the study period and no diagnosis for the AE during the baseline period. Patients in the AE-free cohort had no diagnoses for the studied AE during both the baseline and study periods. Patients with study periods < 6 months were additionally required to have no reported AE during the first 6 months following the index date to ensure no AEs reported immediately after the treatment change were missed. Cohorts were not mutually exclusive such that a patient in the AE cohort for a given AE could be in the AE-free cohort for another AE. For instance, if a patient experienced insomnia during their treatment episode, this patient was placed in the AE cohort for insomnia but could also be placed in the AE-free cohort for another studied AE, such as irritability. Likewise, cohorts were not mutually exclusive among AEs; patients in the AE cohort for insomnia could also be in the AE cohort for another AE (e.g., if they have both insomnia and irritability, they would be in the AE cohort for both AEs). Doing so allowed for cohorts to be more representative of real-world situations where patients may experience multiple AEs at the same time.

Fig. 2.

Sample selection flow chart. ADHD attention-deficit/hyperactivity disorder, FDA Food and Drug Administration, ICD-10-CM International Classification of Diseases, Tenth Revision, Clinical Modification. ¹ADHD was identified using ICD-10-CM code F90.x. ²Patients could have initiated > 1 treatment (i.e., multiple candidate index dates). Treatment initiation occurring during inpatient stays were not considered. ³Only the year of birth is available in the IQVIA database; therefore, when determining patients' eligibility based on age (≥ 6 years old and < 18), all dates of birth were imputed as the mid-year point (July 1)

Measures, Outcomes, and Statistical Analyses

Descriptive statistics were used to summarize patient demographic and clinical characteristics both before and after reweighting for the AE and AE-free cohorts for each of the selected AEs separately. Means, medians, and standard deviations were reported for continuous variables, while frequency counts and percentages were reported for categorical variables. Standardized differences between the AE and AE-free cohorts were calculated before and after balancing, with standardized differences < 0.2 considered well balanced between cohorts [13].

Healthcare costs were adjusted for inflation to 2023 US dollars based on the medical component of the consumer price index and were reported from the payer’s perspective as the amount paid to the healthcare provider by the payer [14]. As the length of treatment regimen duration varied among patients, costs were reported as the average costs per patient per month (PPPM), defined as the sum of costs for a patient during the study period divided by the length of the study period, reported in months. Total excess healthcare costs comprised all-cause medical inpatient, outpatient, and emergency room visits and non-ADHD treatment-related pharmacy costs. In addition, AE-specific medical costs were reported and included inpatient, outpatient, and emergency room costs for medical claims for which there was a recorded diagnosis of the given AE. Costs associated with AEs may exceed AE-specific medical costs; therefore, this study compared similar treatment episodes, with and without AEs, to describe the total burden of a given AE (i.e., total excess healthcare costs). For example, although costs associated with a diagnosis of dehydration would not be captured under the vomiting-specific medical costs, it may be a consequence of severe or untreated vomiting, and as such would be captured under total excess healthcare costs.

Reweighting was conducted using entropy balancing to ensure that key characteristics of patients in the AE and AE-free cohorts had identical means and standard deviations [15], specifically age, sex, region, health plan type, calendar year of the index date, line of therapy at the index date, duration of the index treatment regimen, pediatric comorbidity index score [16], specialist visits, and number of other AEs excluding the studied AE. As this study sought to evaluate the incremental burden of AEs, comorbidities associated with a given AE were not balanced because they are expected to be part of the disease burden for that AE.

A weighted two-part generalized linear regression model (GLM) was used to compare healthcare costs between balanced AE and AE-free cohorts [11, 17]. The first part was a logistic regression model with a binomial distribution that estimated the probability of detecting positive costs for a given cost component (e.g., inpatient costs), and the second part was a GLM with a log link and gamma distribution that estimated positive costs. A separate regression analysis was performed for each of the selected AEs; healthcare costs were the dependent variable and an indicator variable for having the given AE was the independent variable. Robust standard errors were used to estimate P-values and 95% confidence intervals (CIs). All analyses were conducted using SAS Enterprise Guide, version 7.1 (SAS Institute, Cary, NC) and STATA Statistical Software, Release 16.

Results

Patient Characteristics

A total of 393,919 eligible patients met the study's inclusion criteria (Fig. 2). The mean age in the study sample was 12.5 years, nearly two thirds (65.4%) were male, and roughly half (44.0%) lived in the south (Table 1). About one in five (22.7%) patients had seen a specialist (i.e., psychiatrist or neurologist) over the course of the 12-month baseline period, with a mean of 1.2 visits per patient. Stimulant monotherapy was the most commonly initiated ADHD treatment (71.8%), and 41.7% of patients initiated their first observed ADHD treatment as of the index date. More than one in eight patients (13.6%) had a medical encounter for ≥ 1 of the studied AEs during their ADHD treatment episode, with a mean of 0.2 AEs per patient. The prevalence of AEs was estimated from the proportion of patients diagnosed with a given AE recorded on a medical claim during their index treatment episode. Based on this, upper abdominal pain was the most frequent AE (5.2%) followed by vomiting (3.4%), insomnia (3.2%), decreased weight (2.3%), dizziness (1.5%), irritability (0.8%), asthenia (0.5%), and somnolence (0.1%) (Fig. 3).

Table 1.

Baseline characteristics, overall sample

	Overall
Number of patients, N	393,919
Patient characteristics at the index date
Age (years), mean ± SD [median]	12.5 ± 3.4 [12.6]
Male, n (%)	257,785 (65.4%)
Region, n (%)
South	173,388	(44.0%)
Midwest	117,995	(30.0%)
Northeast	61,747	(15.7%)
West	40,604	(10.3%)
Unknown	185	(0.0%)
Health plan type, n (%)
Preferred provider organization	311,342	(79.0%)
Health maintenance organization	41,098	(10.4%)
Point of service	25,788	(6.5%)
Consumer directed health care	11,545	(2.9%)
Indemnity/traditional	4023	(1.0%)
Unknown	123	(0.0%)
Calendar year of index date, n (%)
2016	19,667	(5.0%)
2017	60,053	(15.2%)
2018	51,505	(13.1%)
2019	52,936	(13.4%)
2020	54,854	(13.9%)
2021	63,643	(16.2%)
2022	67,388	(17.1%)
2023	23,873	(6.1%)
Clinical characteristics during the baseline period
Pediatric comorbidity index, mean ± SD [median]	2.2 ± 2.8 [1.0]
Selected physical comorbidities, n (%)
Anxiety	114,597	(29.1%)
Pain conditions	90,372	(22.9%)
Depression	69,543	(17.7%)
Developmental delays	40,565	(10.3%)
Asthma	34,562	(8.8%)
Specialist visits, n (%)	89,412	(22.7%)
Number of specialist visits, mean ± SD [median]	1.2 ± 4.4 [0.0]
Index treatment characteristics
ADHD treatment initiated, n (%)
Stimulants monotherapy	283,005	(71.8%)
Non-stimulants monotherapy	38,587	(9.8%)
Combination therapy	72,327	(18.4%)
Line of therapy, n (%)
1	164,086	(41.7%)
2	115,128	(29.2%)
≥ 3	114,705	(29.1%)
Number of studied AEs during the treatment episode, mean ± SD [median]	0.2 ± 0.5 [0.0]
≥ 1 studied AE, n (%)	53,444	(13.6%)
Duration of line of therapy, days, mean ± SD [median]	325.3 ± 369.1 [209.0]
End of regimen type, n (%)
Remained	120,778	(30.7%)
Switch	71,504	(18.2%)
Drop	31,812	(8.1%)
Add-on	29,367	(7.5%)
Discontinuation (all lines)1	88,910	(22.6%)
Discontinuation2	51,548	(13.1%)

¹Discontinuation of the studied treatment regimen with no other ADHD treatment observed after the studied regimen

² Discontinuation of the studied treatment regimen with ≥ 1 other ADHD treatment observed after the studied treatment regimen

ADHD attention-deficit/hyperactivity disorder, AE adverse event, N number, SD standard deviation

Fig. 3.

Adverse event prevalence. ¹Adverse events were identified based on recorded diagnosis on a medical claim; prevalence was estimated from the proportion of patients from the total sample (N = 393,919) with a diagnosis for a given adverse event recorded on a medical claim during their index treatment episode

Study Cohorts

AE cohorts ranged from 502 to 16,191 patients. The upper abdominal pain cohort was the largest with 16,161 patients, followed by vomiting (11,459), decreased weight (7887), insomnia (7447), dizziness (5319), irritability (2322), asthenia (1520), and somnolence (502). Overall, baseline patient characteristics were consistent across AE cohorts (Supplementary Tables S1–S8). Small differences were noted; for instance, there were more females in the dizziness cohort (56.3%), while all other cohorts comprised more males, and these patients were slightly older (mean age: 14 years old). In the irritability cohort, there were substantially more males (73.8%) than in other cohorts. The mean number of specialist visits varied between 1.1 and 3.0 and was highest in the somnolence (3.0), dizziness (2.0), and asthenia (1.9) cohorts.

Healthcare Costs

Each of the studied AEs was associated with statistically significant increased healthcare costs PPPM (all p < 0.01; Fig. 4). For total excess medical and non-ADHD treatment-related pharmacy costs, asthenia was associated with the highest costs ($1178) followed by somnolence ($821), vomiting ($427), insomnia ($404), dizziness ($380), upper abdominal pain ($336), irritability ($231), and decreased weight ($219). For costs directly related to claims for a diagnosed AE, asthenia was associated with the highest costs ($196) followed by somnolence ($171), insomnia ($169), vomiting ($106), dizziness ($92), upper abdominal pain ($91), irritability ($75), and decreased weight ($46). Although the magnitude of total excess medical and pharmacy costs was greater than AE-specific medical costs, the costs largely followed the same order.

Fig. 4.

Mean difference in total healthcare costs between the AE and AE-free cohorts (2023 USD PPPM)¹. AE adverse event, PPPM per patient per month, USD US dollar. ¹All differences in total healthcare costs were statistically significant at the 5% confidence level

Discussion

In this real-world retrospective cohort study of pediatric patients receiving treatment for ADHD, more than one in eight patients had a medical encounter for ≥ 1 of the studied AEs during their ADHD treatment episode, with abdominal pain, vomiting, and insomnia being the most common. Moreover, AEs experienced during treatment episodes were associated with significant healthcare costs. Asthenia and somnolence accounted for the highest total excess healthcare costs, as well as the highest AE-specific medical costs. To contextualize these findings from a payer’s perspective, accounting for both the prevalence and magnitude of incremental healthcare costs associated with each AE, we calculated the expected annual incremental healthcare costs associated with AEs during ADHD treatment that would be incurred for a hypothetical health plan of 1 million pediatric patients. Specifically, multiplying the costs of each individual AE by the expected number of cases in the hypothetical health plan, the estimated annual incremental healthcare costs would be the highest for upper abdominal pain ($12.8 million) followed by vomiting ($10.8 million) and insomnia ($9.4 million) (Fig. 5). Collectively, the results of this study provide insights into the substantial economic burden associated with AEs experienced during ADHD treatment in pediatric patients and emphasize the importance of reducing these AEs in alleviating this cost burden.

Fig. 5.

Estimated annual incremental healthcare costs associated with AEs (per 1 million pediatric patients; 2023 USD million)^1, 2. AE adverse event, USD US dollar. ¹Assuming an ADHD prevalence of 11.4% and a treatment rate of 53.6% for non-institutionalized children aged 3–17 living in the United States (based on the 2022 National Survey of Children’s Health published in Danielson et al. [2]). ²AE prevalence was based on findings from the current study and AEs lasting for the entire 12-month period. AE-specific medical costs were defined based on a medical claim with a recorded diagnosis for that given AE (e.g., insomnia); additional excess AE-related costs and other healthcare costs included any other incremental healthcare costs in the AE cohort compared to the AE-free cohort that did not have a recorded diagnosis for that given AE (e.g., insomnia)

Although studies investigating the costs associated with AEs in the pediatric population are lacking, a study among adult patients receiving treatment for ADHD similarly demonstrated that AEs experienced during treatment episodes are associated with significant AE-specific and total excess healthcare costs [11]. In that study, roughly half of adult patients had ≥ 1 AE documented through a medical encounter, compared to one in eight in the present study, and the prevalence of AEs was generally higher [11]. The observed differences may partially result from a tendency of reduced tolerability to ADHD medications in adults compared to pediatric patients [18], leading to higher AE prevalence in adults. The lower prevalence of AEs in the pediatric population—relative to the adult population—may also be related to a reduced tendency of children/adolescents to report these AEs to parents or other caregivers or of parents’ ability to communicate these AEs to healthcare professionals. Nonetheless, both studies highlight the substantial impact of AEs experienced during ADHD treatment, which spans pediatric and adult populations.

The prevalence of AEs is the primary driver of their cumulative costs. For instance, although upper abdominal pain was not among the most costly AEs in this study, it had the highest annual costs for the hypothetical health plan given that it was the most prevalent AE (5.2%). Based on the estimated proportion of ADHD-treated patients and AE prevalence in this study, and supposing that no more than one AE is recorded on a given medical claim, annual AE-specific medical costs arising from the eight studied AEs in a hypothetical plan of 1 million pediatric enrollees would amount to approximately $13 million (Fig. 6). Furthermore, while this amount is reported on an annual basis, it is important to note that certain treatment-related AEs may not be experienced for an entire year; therefore, annual costs may be overstated in this example. On the other hand, these AE-specific medical costs may not capture the total cost of AEs experienced during treatment episodes, thus understating the actual costs incurred. For instance, costs incurred in a medical claim with a recorded diagnosis of vomiting would be categorized as AE-specific medical costs, while costs associated with medical claims for potential complications—such as dehydration—would not be considered AE-specific. These examples emphasize the need for a cost model that incorporates ADHD prevalence, proportion of ADHD patients receiving treatment, and AEs among patients receiving treatment, which would enable health plans to calculate more specific estimates of their costs. Nevertheless, our findings suggest that the estimated AE-specific medical costs remain non-negligible; thus, strategies that lessen the toll of AEs experienced during ADHD treatment may ultimately lead to a reduction in the burden of ADHD overall.

Fig. 6.

Estimated annual AE-specific medical costs in a hypothetical health plan with 1 million members. ADHD attention-deficit and hyperactivity disorder, AE adverse event. ¹Assuming an ADHD prevalence of 11.4% and a treatment rate of 53.6% for non-institutionalized children aged 3–17 living in the United States (based on the 2022 National Survey of Children’s Health published in Danielson et al. [2]). ²Prevalence was estimated from the proportion of patients diagnosed with a given AE recorded on a medical claim during their index treatment episode (based on the total sample of 393,919 patients). Calculations are based on all AE-specific medical costs being independent—i.e., no more than one AE is recorded on an individual claim and AEs lasting the entire 12-month period. Example: vomiting had a prevalence of 3.4% in this study. In a hypothetical population of 1 million pediatric patients, this prevalence translates to 2104 treated patients with ADHD experiencing vomiting (3.4% of 61,104 patients). At the population level, the total annual AE-specific medical costs are: 2104 treated patients with ADHD experiencing vomiting × $106 AE-specific costs PPPM × 12 months = $2,669,311 (with rounding)

AEs experienced during ADHD treatment have been reported as a common cause of treatment modification [6, 19], which is associated with additional healthcare costs not examined in this study [7]. Indeed, a retrospective chart review study conducted in children and adolescents reported that among patients who discontinued an ADHD treatment, 25% of patients cited treatment-related complications as a reason for discontinuation [6]. In addition, a previous US-based claims study of pediatric patients receiving treatment for ADHD reported that around two thirds of children and adolescents experienced ≥ 1 treatment change, such as discontinuation, switch, add-ons, and drop, over the course of the 12 months following the first observed treatment initiated [7]. These treatment changes were associated with significant excess annual healthcare costs among pediatric patients and increased with each subsequent treatment modification [7]. Hence, the economic impact of AEs extends beyond the healthcare costs observed in this study, and these AEs contribute to the clinical burden of ADHD management.

Apart from healthcare costs, AEs may affect treatment adherence as well as the daily functioning and quality of life of patients and their parents/caregivers. Previous studies have shown that AEs, most commonly sleep disturbances, decreased appetite, and emotional impulsivity, are a common cause of treatment interruption and reduced adherence [5, 20], which may negatively affect the effectiveness of treatment. Regarding quality of life, a recent parents/caregivers survey showed that pediatric patients with ADHD who experienced AEs had substantially reduced emotional, social, and school functioning than those without AEs [5]. In addition, parents/caregivers of pediatric patients with AEs tend to have increased work and activity impairment, worse mental health outcomes, and increased outpatient visits for any reason [5]. Together, these studies suggest that the burden of AEs may be far-reaching, particularly when considering the quality of life implications on patients and their parents/caregivers alike. Additional research is needed to further explore these impacts.

The AEs included in this study were limited to those that had a statistically significant risk difference in the pediatric MAIC study [10], which could be identified in claims using ICD-10-CM codes. Hence, this study does not encompass all possible AEs associated with ADHD treatment, including those for which patients did not receive medical care, and may therefore underestimate the prevalence and number of AEs experienced. Indeed, compared to the present study, where around 14% of pediatric patients had ≥ 1 AE, previous real-world studies reported a higher prevalence of AEs (42%–67%) [5, 6]. Importantly, commonly reported AEs not included in this study may still be of concern. For example, decreased appetite, which was reported to be experienced by 12% of pediatric patients in a parent/caregiver survey [5], has been associated with delayed or suppressed growth in children [21–23]. However, the lack of associated ICD-10-CM code prevents cost estimations using claims. Dry mouth, reported in 10% of patients in the same parent/caregiver survey [5], may be captured in claims through an ICD-10-CM code [24], but it may be challenging to capture its costs as pediatric patients may not specifically seek care for it. Instead, patients may show non-specific signs and symptoms of dry mouth, such as difficulty swallowing, increased fluid intake, and dental caries [25, 26]. Similarly, as noted by clinical experts, patients may experience bad breath (halitosis), a frequent consequence of dry mouth, and may not recognize the two to be linked [27]. Lastly, headaches, reported in 8% of patients in the parent/caregiver survey [5], may have limited healthcare costs as they do not tend to require medical attention but can still have important consequences on attendance and functioning at school and quality of life [28, 29]. Therefore, these studies further suggest that impacts of AEs may exceed what was reported in this study.

Given the prevalence and burden of AEs, it is important to highlight the role of shared decision-making between physicians and patients when selecting a treatment for ADHD. Studies that provide insights into patient preferences and the role of AEs in adherence and treatment changes may help to guide these decisions [6]. Future research, such as cost modelling studies that integrate drivers of treatment change, may provide a deeper understanding of population-level costs with greater specificity and further inform healthcare and policy decisions. Overall, there is the potential for treatments with a favorable safety profile to lessen the burden associated with AEs experienced during ADHD treatment for both patients and the healthcare system [10].

Limitations

Some limitations should be considered when interpreting the results of this study. First, the use of claims data is associated with a risk of data omissions, coding errors, and the presence of rule-out diagnoses, which may have led to the misclassification of patients in a given cohort. Second, this study captured AEs for which medical care was sought. As these AEs may be more severe and costly, the results may not be representative of milder AEs and the prevalence of AEs may have been underestimated. In addition, information on whether AEs were related to treatment was unavailable in the claims database. Third, while cohorts were weighted based on observable characteristics, the possibility of residual confounding due to unobservable confounders cannot be ruled out. Hence, the results of this retrospective observational study should be viewed as measures of association rather than causal inferences. Fourth, this study does not encompass all possible AEs; only those available in claims data and for which there were differences across treatments in a prior MAIC study were included [10]. Finally, the results of this study may not be generalizable to all US pediatric patients with ADHD given that the study sample only included commercially insured patients.

Conclusions

This retrospective cohort study of pediatric patients treated for ADHD found that AEs experienced during ADHD treatment episodes were common and associated with substantial total and AE-specific healthcare costs. Collectively, the findings of this study shed light on the burden associated with AEs in ADHD treatment, which should be considered when making decisions related to ADHD management. ADHD treatment options with a more favorable safety profile may help alleviate the burden experienced by patients and the healthcare system.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Jeff Schein: Conceptualization, Resources, Supervision, Methodology, Writing – Reviewing and Editing. Ann Childress: Conceptualization, Methodology, Writing – Reviewing and Editing. Maryaline Catillon: Conceptualization, Methodology, Project administration, Formal analysis, Visualization, Writing—Original draft. Marjolaine Gauthier-Loiselle: Conceptualization, Methodology, Project administration, Formal analysis, Visualization, Writing—Original draft. Martin Cloutier: Conceptualization, Supervision, Methodology, Formal analysis, Visualization, Writing—Original draft. Anaïs Lemyre: Methodology, Formal analysis, Visualization, Writing—Original draft. Alice Qu: Methodology, Formal analysis, Visualization, Writing—Reviewing and Editing. Frederic Kinkead: Methodology, Formal analysis, Visualization, Writing—Reviewing and Editing.

Funding

This study and the journal’s Rapid Service and Open Access fees was funded by Otsuka Pharmaceutical Development & Commercialization, Inc.

Data Availability

The findings of this study are based on information licensed from IQVIA: IQVIA PharMetrics^® Plus for the period from October 1, 2015, to September 30, 2023, reflecting estimates of real-world activity. All rights reserved.

Declarations

Conflict of Interest

Jeff Schein is an employee of Otsuka Pharmaceutical Development & Commercialization. Martin Cloutier, Marjolaine Gauthier-Loiselle, Maryaline Catillon, Anaïs Lemyre, Alice Qu, and Frederic Kinkead are employees of Analysis Group, Inc., a consulting company that has provided paid consulting services to Otsuka Pharmaceutical Development & Commercialization, Inc, which funded the development and conduct of this study and manuscript. Ann Childress received research support from Aardvark, Allergan, Axsome, Emalex, Akili, Cingulate, Corium, Ironshore, Les Laboratoires Servier, Lumos, Neurocentria, Otsuka, Purdue, Adlon, Sunovion, Tris, KemPharm, and Supernus; was on the advisory board of Corium, Otsuka, Tris, and Supernus; received consulting fees from Aardvark, Alora, Axsome, Aytu, Cingulate, Corium, Lumos, Medison Pharma, Neurocentria, Noven, Otsuka, Sky, Tris, KemPharm, Supernus, and Tulex; received speaker fees from Takeda, Corium, Ironshore, Tris, and Supernus; and received writing support from Otsuka, Takeda, Corium, Ironshore, Purdue, and Tris.

Ethical Approval

The study was considered exempt research under 45 CFR § 46.104(d)(4) as it involved only the secondary use of data that were de-identified in compliance with the Health Insurance Portability and Accountability Act (HIPAA), specifically, 45 CFR § 164.514.