Trends in Office-Based Anxiety Treatment Among US Children, Youth, and Young Adults: 2006–2018

We conducted a serial cross-sectional analysis of the NAMCS, a yearly, nationally representative survey of office-based physician visits in the United States. Patients 4 to 24 years of age were included in the current study and stratified by age group (4–11 years, 12–17 years, and 18–24 years). Visits with an anxiety disorder diagnosis were defined as having the following diagnostic codes: International Classification of Diseases, 9^th Edition 300.0, 300.2, 309.21, 313.23; International Classification of Diseases, 10^th Edition F40, F41, F930, and F940. Treatment categories were created on the basis of the NAMCS survey collection tool: (1) psychotherapy or counseling (hereafter, “therapy alone”), (2) therapy and medications, (3) medications alone, and (4) neither therapy nor medication (hereafter, “neither”). Medications included were those commonly used to treat anxiety disorder symptoms for short- or long-term use: antidepressants (selective serotonin reuptake inhibitors [SSRIs], serotonin and norepinephrine reuptake inhibitors [SNRIs], tricyclic antidepressants), benzodiazepines, atypical anxiolytics (buspirone), and antihistamines.^14,15  We included all anxiety disorder diagnoses that may be observed among children and youth. Although types of medications used to treat specific diagnoses vary, psychotherapeutic treatment is the recommended first-line treatment of most conditions. SSRIs are to be considered for the treatment of severe social, generalized, separation anxiety, or panic disorder in patients >6 years of age, preferentially in combination with cognitive behavioral therapy. We also explore an alternative definition of the four treatment categories described above using a more limited list of medications (only including SSRI and SNRI).

Consistent with previous longitudinal trend studies using the NAMCS,^12,13  data from multiple years were pooled together because of small numbers of visits with anxiety disorder diagnosis for individual years, creating 3 time periods: (1) 2006 to 2009, (2) 2010 to 2013, and (3) 2014 to 2016 and 2018 (NAMCS not released for 2017). We compared differences in treatment categories using multinomial logistic regression, adjusting for age group, sex, and race/ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic, and other) to account for differences over time in sample demographics associated with mental health treatment.¹⁶  We included categorical indicator variables for the periods and compared the second and third periods to the first period (referent group) to assess temporal changes in the use of treatments; we considered P values for coefficients to be significant at α <.05. The resulting proportions were weighted to reflect national estimates. The primary analyses were repeated as a sensitivity analysis with the alternative treatment category outcome on the basis of the more limited medication list. All analyses were performed by using StataIC 16 (StataCorp, College Station, TX). This study used publicly available nonidentifiable data and was exempt from institutional review board approval.

There were 1734 visits for anxiety disorder over the study period in the sample, representing a total of 46.4 million office visits with an anxiety disorder diagnosis. The overall proportion of office visits with an anxiety disorder diagnosis significantly increased from 1.4% (95% confidence interval [CI] 1.2–1.7; n = 9 246 921 visits) in 2006 to 2009 to 4.2% (95% CI 3.4–5.2; n = 23 120 958 visits) in 2014 to 2018. The proportion of visits with anxiety disorder diagnosis increased for all age groups, particularly for adolescents (1.4% [95% CI 1.1–1.9] to 4.6% [95% CI 3.3–6.4]) and young adults (2.5% [95% CI 2.1–3.1] to 7.1% [95% CI 5.7–8.9]). The increase in the proportion of child visits with anxiety disorder diagnoses was smaller, from 0.6% (95% CI 0.4–0.9) to 1.5% (95% CI 1.0–2.2). The distribution of patient demographic characteristics across visits with an anxiety disorder diagnosis, including age group, sex, and race/ethnicity, were similar across periods, except for a higher proportion of Hispanic patients in the last period (Table 1). The type of providers for anxiety visits was also similar over time (Table 1).

The overall proportion of visits with therapy alone or therapy and medications decreased from 48.8% (95% CI 40.1–57.6; n = 4 513 468 visits) to 32.6% (95% CI 24.5–41.8; n = 7 536 619 visits) in the first to last periods, respectively (Table 1). However, the overall proportion of visits with medication prescribed did not differ by period, and in the last period, medications were prescribed in 61.8% (95% CI 54.1–68.9; n = 14 286 916) of all visits. Among those prescribed medications, the most prescribed medications were SSRIs (46.3%), benzodiazepines (13.2%), SNRIs (8.1%), and antihistamines (3.6%). Types of medications prescribed were also consistent over time, except for antihistamines, which increased from 0.1% (95% CI 0.0–0.8) in 2006 to 2009 to 3.6% (95% CI 1.8–7.0) in 2014 to 2018.

When examining differences by period in the 4 treatment categories, the proportion of visits receiving neither medication nor therapy increased over time for all age groups, as did the proportion of visits receiving medications alone (Fig 1). Relative to therapy only, the likelihood of having neither therapy nor medications was significantly greater for the last period compared with the first period (relative risk ratio = 2.70, 95% CI 1.33–5.47; P <.01), as was having medication only (relative risk ratio = 2.42, 95% CI 1.24–4.72; P = .01). However, there were no significant differences for the second period compared with the first period. In addition, there were no differences for any period in the likelihood of combined therapy relative to therapy only. The differences by period were similar when the alternative definition of medications (including SSRI and SNRI only) was used to derive treatment categories (Supplemental Fig 2).

Anxiety disorder was more frequently diagnosed over time among a representative sample of US office-based visits for children, adolescents, and young adults. In addition, the proportion cof office visits for anxiety disorders with any type of treatment at the visit decreased over time. However, overall medication prescribing was largely stable, and there was little hange in the types of medications prescribed, except for changes in antihistamine prescribing. As a result, a larger proportion of visits received either medication alone or no treatment. Medications are recommended as either first-line or adjunct treatment for those receiving therapy for anxiety disorders.¹⁴ 

Previously reported declines in receipt of office-based therapy for mental health conditions among adults^12,17  extend to young patients with anxiety disorder through 2018. The low use of psychosocial care with pharmacotherapy for mental health conditions has been documented previously among children.^18–20  Findings from the current study reveal a gap in the quality of care for patients with anxiety disorders, which could worsen with the increasing prevalence of anxiety.^9,13,21  If the volume of children and adolescents with anxiety disorders continues to increase, the health system, in which most young patients are already receiving medications alone or neither medication nor therapy during office-based visits, would be further taxed.

There are several potential explanations for the trends in anxiety disorder treatments observed in the study with important implications for policy and practice. First, the absolute number of visits with a diagnosis of anxiety disorders increased, meaning that office-based settings treated patients with an anxiety disorder at levels that may have exceeded their capacities to deliver in-office counseling. If they could not counsel these patients, providers could give them medications alone or no treatment. Alternatively, there could be secular trends in behavioral health prescribing practices that favor medication, which could be due to marketing,²²  incentives or reimbursement practices,²³  or caregiver/patient preferences.²⁴  Recent changes in prescribing for mental health conditions have been documented elsewhere, with rates of antidepressant prescribing increasing in recent years in Medicaid and commercially insured children and youth.²⁵  With the recent recommendations for universal anxiety screening in primary care settings by the US Preventive Services Task Force,²⁶  the prevalence of clinically recognized anxiety disorder may increase even further and exacerbate quality concerns for anxiety disorder treatment if not implemented with additional integrated treatment resources.

This study relied on provider reports and diagnoses, which are likely to under-identify patients with anxiety disorders, and we lack information about the care patients received after or before these visits. Second, we do not have information on the specific type of psychotherapy or counseling recommended or delivered. Thus, we cannot comment on whether the therapies delivered were evidence-based. In addition, the medications prescribed may not be the preferred medications for treating anxiety disorder in all cases. However, our description captures routine practice in office-based settings. Third, although we adjusted for demographic characteristics in our regression model, we cannot rule out the possibility that variation in other unobserved patient characteristics in the cross-sectional patient visit samples over time may contribute to the changes in the treatment categories we observed. Finally, we used a broad age range of children, adolescents, and young adults, and age-specific guidelines do differ. However, our goal was to broadly describe the treatment of anxiety disorders across a range of ages to describe trends in office-based settings. Future research could examine how prescribing varied over time in delivering age-specific treatment guidelines.

The reduction in therapy during office visits and the greater reliance on medications for anxiety disorders may reflect growing resource constraints in office settings in the context of a child and youth mental health crisis that has been building over time.¹  Further research is needed to investigate how reliance on medications has evolved during the coronavirus disease 2019 pandemic and whether these trends have continued or worsened.

CI

confidence interval

NAMCS

National Ambulatory Care Medical Survey

SNRI

serotonin and norepinephrine reuptake inhibitor

SSRI

selective serotonin reuptake inhibitor