Anxiety is common, screening tools are available, and treatment can be effective. Recently, anxiety screening has been recommended for adolescent girls beginning at 13 years of age.
To evaluate the evidence regarding anxiety screening test accuracy in primary care for children and adolescents and assess the effectiveness of treatment of individuals identified through screening.
We searched PubMed, the Cochrane library, and references to potentially eligible studies cited in other articles.
Screening studies were included if they were conducted in primary care or a similar population and employed a reference standard based on DSM criteria. Treatment studies were included if subjects were identified through screening and there was at least 1 comparator intervention or a placebo arm.
At least 2 reviewers evaluated each identified reference.
Two screening studies (1 with low risk of bias and 1 with high risk of bias) and 1 treatment study with a low risk of bias were included. The screening study with a low risk of bias reported a sensitivity of 56% and specificity of 80%. The treatment study found individual cognitive behavioral therapy to be effective for screen-detected adolescents with social phobia.
This review only included screening or treatment studies with clear evidence that the study populations were derived from an unselected population reflective of typical primary care. Relevant studies not indexed in PubMed or the Cochrane library could have been missed.
There are significant gaps in evidence related to anxiety screening in the primary care setting.
In 2020, the Women’s Preventive Services Initiative (WPSI) recommended “screening for anxiety in women and adolescent girls aged 13 years or older who are not currently diagnosed with anxiety disorders, including pregnant and postpartum women.”1 The WPSI recommendations are based on a systematic review and critical evaluation of the available clinical evidence and are linked to coverage through the Affordable Care Act,2 similar to the recommendations from the US Preventive Services Task Force and the American Academy of Pediatrics Bright Futures periodicity schedule. This anxiety screening recommendation raises important considerations for pediatric primary care, including the optimal instruments and frequency for screening, the appropriate clinical pathways needed after a positive screen result, whether screening should be restricted by sex, and if children <13 years of age should be included in screening.
These considerations are important because anxiety is common. In 2016, 7.1% of parents of children aged 3 to 17 years in the United States reported that their children currently had an anxiety problem.3 The reported proportion with anxiety increased with age, with adolescents 12 to 17 years of age having about a 60% increased risk of anxiety, compared with children 6 to 11 years of age, after adjusting for a wide variety of socioeconomic and health characteristics.3 One impact of the coronavirus disease 2019 pandemic is worsening mental health, including higher rates of anxiety.4 Anxiety often co-occurs with depression5 ; therefore, identifying the presence of anxiety raises the need to assess for depression or increased suicide risk. Because the temporal relationship between anxiety and depression is complex and can vary within individuals, screening for one condition will not necessarily identify individuals with the other.6 Anxiety can have significant impact on the individual (eg, alcohol or other substance use7 and school problems) and be associated with family dysfunction.8 Interventions are available to successfully treat anxiety, including cognitive behavioral therapy (CBT) and medications (eg, selective serotonin reuptake inhibitors and serotonin-norepinephrine reuptake inhibitors).9
To help inform anxiety screening in the pediatric setting, we conducted a systematic review with a key difference from the WPSI systematic review on the topic. Because screening studies enrolling subjects with a high risk of anxiety can overestimate test accuracy,10 we only considered screening studies that enrolled subjects being seen in the primary care setting for routine preventive care or enrolling subjects in similar setting (eg, school- or community-based screening of children or adolescents not known to have an increased risk of anxiety or with a condition associated with anxiety). Additionally, studies of the treatment effectiveness of subjects identified because of overt symptoms might not be generalizable to subjects with unrecognized anxiety who are identified through screening. For that reason, we solely included treatment studies that enrolled subjects similar to those who would be detected through screening as part of routine preventive care. Restrictive criteria were placed on the inclusion criteria because the goal of this review was not to evaluate the comparative effectiveness of different approaches to anxiety treatment of children and adolescents but instead focus on the potential benefits and harms (eg, labeling and overdiagnosis) of pediatric primary care anxiety screening with subsequent evaluation and treatment, either within the primary care setting or through referral.
Methods
Scope of the Review
The goal of this review was to assess whether screening for anxiety in children and adolescents ≤21 years of age in the primary care setting improves health or other important patient outcomes. To do this, we searched for direct evidence of the impact of screening (ie, randomized trials of anxiety screening in the primary care setting). We also considered indirect evidence regarding the benefits and harms of anxiety screening, including the accuracy of screening instruments in primary care for identifying anxiety, whether such screening is associated with harm, and the benefits and harms associated with treating anxiety among those identified through screening in primary care. This review was preregistered in the International Prospective Register of Systematic Reviews (PROSPERO, CRD42019142902).
Data Sources
We searched PubMed and the Cochrane library, including the Cochrane Database of Systematic Reviews and the Cochrane Central Register of Controlled Trials, for relevant studies. The PubMed search strategy (see Table 1) was developed by the investigators and then reviewed by our medical librarian. The Cochrane library was searched by using the keyword “anxiety.” The searches initially included the date of inception of each database through July 2, 2020, followed by an update to bring the search to May 4, 2021. Citations in included studies were reviewed to supplement the search. We also reviewed citations included in the WPSI anxiety systematic evidence review11 and citations included in a 2017 pediatric anxiety treatment comparative effectiveness review funded by the US Agency for Healthcare Research and Quality.9
PubMed Search Strategy
Search Number . | Terms . |
---|---|
1 | (((anxiety disorders OR phobias OR generalized anxiety disorder OR panic disorder) AND (screening OR questionnaire OR assessment))) |
2 | ((primary care OR pediatrics OR family practice)) |
3 | (Humans[Mesh] AND English[lang] AND ((infant[MeSH] OR child[MeSH] OR adolescent[MeSH]) OR young adult[MeSH])) |
4 | #1 AND #2 AND #3 |
Search Number . | Terms . |
---|---|
1 | (((anxiety disorders OR phobias OR generalized anxiety disorder OR panic disorder) AND (screening OR questionnaire OR assessment))) |
2 | ((primary care OR pediatrics OR family practice)) |
3 | (Humans[Mesh] AND English[lang] AND ((infant[MeSH] OR child[MeSH] OR adolescent[MeSH]) OR young adult[MeSH])) |
4 | #1 AND #2 AND #3 |
Study Selection
We included English-language reports from studies conducted in countries with high levels of human development.12 Study subjects must have been ≤21 years of age. Studies with an average age >21 years of age that did not have a subpopulation of subjects ≤21 years that could be separately analyzed were excluded.
For studies of screening, subjects must not have been known to have anxiety before screening, and the screening must have occurred in the primary care or similar setting (eg, school screening). Screening studies were included if they used (1) a screening instrument specifically designed to identify anxiety or (2) a more general mental health screening tool that had an anxiety module or was being used to identify anxiety. Screening studies were included if subjects with a positive screen result and at least a sample of subjects with a negative screen result had a standardized diagnostic evaluation for anxiety based on current Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, categories (ie, generalized anxiety disorder, panic disorder, phobias, agoraphobia, social anxiety disorder, and separation anxiety disorder) or previous versions of the Diagnostic and Statistical Manual of Mental Disorders, which included other disorders (eg, obsessive compulsive disorder, post-traumatic stress disorder).13 We excluded screening studies in which researchers selected subjects on the basis of symptoms (eg, anxiety symptoms and abdominal pain) or on the basis of having a specific health condition (eg, asthma and prematurity). We also excluded screening studies in which researchers only evaluated for 1 type of anxiety or if the test characteristics (ie, sensitivity, specificity) for anxiety could not be confirmed on the basis of information within the report. If sensitivity or specificity was not reported but could be determined on the basis of information in the report, the 95% confidence intervals (CIs) were calculated with Stata 17 (Stata Corp, College Station, TX).
To be included as a treatment study, all subjects must have been identified in a primary care or similar setting through screening. Randomized trials or observational studies of anxiety treatment were included if there was a comparator group (eg, placebo, treatment as usual, and active comparator) regardless of the intervention (eg, CBT, educational supports, medication, psychotherapy, complementary or alternative medicine, or combined interventions) except that medication studies had to be for an indication and age group approved by the US Food and Drug Administration. Treatment studies in which researchers targeted only 1 type of anxiety were included. Studies with a single-group design with no comparator or in which researchers only evaluated dose escalation were excluded.
Two reviewers independently reviewed all titles and abstracts of the identified articles, and either reviewer could move the article to full-text evaluation. Two reviewers independently evaluated those articles that moved to full-text evaluation, with agreement required to determine if the article should be included in the review. Disagreements were resolved by consensus.
Data Extraction and Quality Assessment
One reviewer independently extracted relevant data and assessed the methodologic quality from each included report. Screening study quality was assessed with the Quality Assessment of Diagnostic Accuracy Studies instrument.14 Treatment study quality was based on a 6-point scale that assessed the process for allocation (adequately generated, adequately concealed, knowledge of allocated intervention adequately prevented); whether incomplete data were addressed; if the report was free of possible selective reporting; and if there were any other problems that could put the study at high risk of bias. A second reviewer confirmed the accuracy of the data extraction and quality rating. Disagreements were resolved by consensus. No meta-analysis was planned because of expected gaps in the available research.
Results
A total of 2653 reports were reviewed for eligibility over the duration of the systematic review, of which 3 are included in this review (Fig 1).
Identification of studies for inclusion. aThe exclusion list is hierarchical. Studies were excluded for the first exclusion criterion identified.
Identification of studies for inclusion. aThe exclusion list is hierarchical. Studies were excluded for the first exclusion criterion identified.
Screening Studies
Two screening studies were included. In 1 study, researchers recruited 50 subjects 7 to 12 years of age attending a primary care clinic for preventive care and screened for anxiety with the Revised Children’s Manifest Anxiety Scale.15 On the basis of Diagnostic and Statistical Manual of Mental Disorders, Third Edition, criteria, 3 subjects had an anxiety diagnosis. Although not reported in the study, using this instrument to screen for anxiety had a sensitivity of 66.6% (n = 2 of 3; 95% CI: 9.4%–99.2%) and specificity of 100% (n = 47 of 47; 95% CI: 92.5%–100%). This study was considered to have a high risk of bias because it was unclear if the diagnostic evaluation was done without knowledge of the screening test results.
In the other study, researchers recruited subjects 2 through 5 years of age attending primary care clinics for routine preventive care or sick visits who were screened with the Child Behavior Checklist anxious and depressed scale, with comparison with Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, (DSM-IV) criteria.16 Overall, there were 3433 subjects screened, with 944 positive results. All who screened positive and a random sample (n = 189) with a negative screen result were selected for diagnostic follow-up. Ultimately, 917 (80.9%) subjects had a diagnostic follow-up, with dropouts primarily due to parent refusal. Most (80.6%) of the parent refusals were among those with a positive screen result. The study reported, “The screen performed adequately with a sensitivity of 56% and a specificity of 80%.”16 This study had a low risk of bias.
Neither study addressed harms related to screening.
Treatment Studies
One treatment study focusing on social phobia based on DSM-IV criteria with a low risk of bias (6 of 6 possible points) was included. In this randomized controlled trial, subjects were eighth- to 10th-grade students recruited from school-based screening with the Social Phobia and Anxiety Inventory for Children (SPAI-C).17 Insufficient information was provided to assess the screening test characteristics. There were 128 subjects randomly assigned to individual CBT (n = 36; 12 sessions [50 minutes each]; 21 subjects initially), group CBT (n = 58; 10 sessions [90 minutes each]; 20 subjects initially), and a group attentional placebo (n = 34; 10 sessions [90 minutes each]). The mean age was 14 years; 56.1% were female, and most (64.9%) subjects had other psychiatric comorbid conditions, including generalized anxiety disorder (28.1%) and/or depression (8.8%). On the basis of 2 main outcome measures, SPAI-C score and the Social Thoughts and Beliefs Scale (STABS), individual CBT demonstrated efficacy (Hedges-corrected Cohen’s d) from before treatment to immediately after treatment and from before treatment to 12-month follow-up (SPAI-C: 2.96 and 2.19; STABS: 1.92 and 1.62). The group CBT was not found to have efficacy immediately after treatment; however, there was improvement in the SPAI-C score at 12-month follow-up. Overall, the individual CBT was superior to the group CBT based on the SPAI-C score and STABS. At 12 months, 72.7% of those in the individual CBT group and 53.3% of those in the group CBT group no longer met the criteria for having social phobia, with no differences between the treatment groups. There was large differential attrition during treatment (16% individual CBT group, 41.2% group CBT, and 11.1% attentional placebo) that continued during the follow-up period. However, an intention-to-treat analysis for the 12-month follow-up for the CBT groups was similar to the efficacy evaluation. Harms related to treatment were not described.
Discussion
Although anxiety is common and an important cause of morbidity in children and adolescents, in this systematic evidence-review, we identified substantial gaps related to screening in the primary care pediatric setting. We were surprised that only 2 screening studies (1 in which researchers enrolled preschool-aged and other school-aged children and 1 treatment study in which researchers evaluated CBT for social phobia) met the inclusion criteria. Our search was limited to PubMed and the Cochrane library. Although we also reviewed the citations in reports considered for inclusion, there is a chance that we missed relevant articles. Screening for anxiety with subsequent treatment as part of routine preventive care might be beneficial. However, we seem to lack the evidence regarding screening accuracy, potential screening-related harms, or treatment outcomes to guide the effective implementation of anxiety screening. As described by the US Preventive Services Task Force, important considerations when deciding to implement screening include whether evidence-based interventions exist for the identified high-risk groups, if those interventions are readily available, if there are accurate screening tools to identify high-risk groups, and what harms may result from screening (eg, anxiety from positive screen results, the difficulty of getting behavioral health services for diagnosis and treatment, opportunity costs particularly for false-positives, and overtreatment).18,19 This evidence is necessary to inform recommendations for best practice in the primary care setting.
Many of the studies considered for this review were excluded because of the restriction of only including studies that were based in unselected primary care or similar populations. As with any systematic review, there are challenges about whether specific studies should be included. A total of 3 screening studies and 2 treatment studies that were excluded underscore the challenges of evaluating the impact of anxiety screening during preventive care visits.
One study of anxiety and depression screening targeted adolescents in primary care with asthma (n = 779) and then selected control subjects without asthma (n = 596).20 This study reported that, compared with DSM-IV criteria for anxiety with or without depression, the Childhood Anxiety Sensitivity Index had a sensitivity of 73% and specificity of 64% and the Mood and Feelings Questionnaire Short Form had a sensitivity of 76% and specificity of 70%. However, the test characteristics were not separately reported for the control subjects. As described in the report, adolescents with asthma have a higher risk of anxiety. Findings from a study with a large proportion of subjects with asthma, therefore, might not be generalizable to routine preventive care. Another study recruited subjects 8 to 15 years of age with clinic visits for preventive care, recurrent abdominal pain, or minor illnesses, who were screened with Pediatric Symptom Checklist-17 (PSC-17).21 This study was excluded because of the strong association between anxiety and recurrent abdominal pain.22 Follow-up, which oversampled subjects with behavioral health problems, included additional screening instruments and diagnostic evaluation with DSM-IV criteria. This study reported test characteristics for the PSC-17 internalizing scale at 2 cut-points (≥5 [sensitivity: 52%; specificity: 74%]), ≥4 [sensitivity: 65%; specificity: 62%]), the 5-question version of the Screen for Child Anxiety Related Emotional Disorders (subject-completed [sensitivity: 44%; specificity 89%]), parent-completed [sensitivity: 44%; specificity: 92%]). Another excluded screening study included 415 subjects 12–21 years of age from primary care with complete data regarding the Behavioral Health Screen, an Internet-based screening tool for anxiety and behavioral health concerns.23 Although the screening test characteristics were reported for anxiety (sensitivity: 88%; specificity: 67%), the diagnostic approach was not clearly described.
One of the treatment studies excluded from this review evaluated brief behavioral therapy conducted within primary care among subjects 8 to 17 years of age with anxiety or depression based on DSM-IV criteria, finding that it was effective compared with assisted referral to care.24 Of the 620 subjects, 90% were referred to the study from pediatric clinics, and 10% self-referred from flyers in the practices. This study was excluded because there was no information about how practices identified subjects for referral (ie, through screening versus previously diagnosed anxiety or depression). Additionally, the reported outcomes combined subjects with anxiety and/or depression in such a way that outcomes specific for those with anxiety could not be evaluated. Another excluded treatment study compared collaborative care with enhanced usual care, including psychoeducation and facilitated referral, for children with a wide array of behavioral problems.25 This study was excluded because subject enrollment was not based solely on screening and treatment effectiveness for anxiety could not be separately assessed.
The findings from this review raise important implementation issues for primary care pediatricians. Although screening is now recommended for women and adolescent girls 13 years and older, findings from this review cannot be used to evaluate the comparative effectiveness of any approach for either screening or follow-up care or whether screening should be expanded to other populations. Given the burden of anxiety and the rapid adoption of screening recommendations, there is an urgent need for pragmatic studies to inform practice.
This review did not consider the role of anxiety prevention. An important consideration for future research is whether the goal of screening should be to identify those with subclinical symptoms of anxiety to provide interventions to prevent progression to an anxiety disorder. For example, CBT can reduce anxiety symptoms in at-risk children and adolescents.26 Another important consideration is whether children or adolescents should be screened separately for each type of behavioral health problem (eg, anxiety, depression, and suicidality) or whether screening should be considered more holistically. For example, the PSC-1727 and the Behavioral Health Checklist28 are broadband screening instruments for internalizing, externalizing, and attentional concerns. These symptoms often co-occur and transdiagnostic approaches to screening and follow-up care might be effective29,30 and ultimately more feasible in the primary care setting.
Because of the frequency of recommended preventive care visits, pediatricians are uniquely positioned to address anxiety. This study highlights the significant gaps in knowledge related to routine comprehensive anxiety screening. It is important to recognize that these evidence gaps do not necessarily imply that anxiety screening is not beneficial. Pediatricians who choose to implement anxiety screening according to the WPSI recommendations or for other patients should ensure that strategies are in place to maximize the potential benefits (eg, ensure that patients and families know about the goal of screening, communicate the risk of false-positive and false-negative screen results, facilitate follow-up care, and monitor treatment outcomes).
Dr Kemper conceptualized the project, designed the systematic review, evaluated articles, drafted the initial manuscript, and reviewed and revised the manuscript; Drs Letostak, Hostutler, Stephenson, and Butter evaluated articles and reviewed and revised the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
FUNDING: No external funding.
This trial has been registered with PROSPERO (https://www.crd.york.ac.uk/prospero/) (identifier CRD42019142902).
- CBT
cognitive behavioral therapy
- CI
confidence interval
- DSM-IV
Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition
- PSC-17
Pediatric Symptom Checklist-17
- SNRI
Serotonin-norepinephrine reuptake inhibitors
- SPAI-C
Social Phobia and Anxiety Inventory for Children
- SSRI
Selective serotonin reuptake inhibitors
- STABS
Social Thoughts and Beliefs Scale
- USPSTF
United States Preventive Services Task Force
- WPSI
Women’s Preventive Services Initiative
References
Competing Interests
POTENTIAL CONFLICT OF INTEREST: Dr Kemper serves as an evidence-review consultant for the American Academy of Pediatrics Bright Futures. The other authors have indicated they have no potential conflicts of interest to disclose.
FINANCIAL DISCLOSURES: The authors have indicated they have no financial relationships relevant to this article to disclose.
Comments