Purpose: To review the evidence on screening (benefits and harms of screening, accuracy of screening, benefits and harms of treatment) for suicide risk, anxiety, and depression in children and adolescents in settings relevant to primary care in the United States for the U.S. Preventive Services Task Force.
Data Sources: PubMed, the Cochrane Library, PsycINFO, CINAHL and trial registries through July 19, 2021; bibliographies from retrieved articles, outside experts, and surveillance of the literature through June 1, 2022.
Study Selection: Two investigators independently selected English-language studies using a priori defined criteria. We included trials that evaluated the benefits or harms of screening for suicide risk, anxiety, or depression compared with no screening or usual care. We included studies of screening with instruments feasible in primary care settings. For treatment benefits and harms, we included drugs approved for pediatric use by the Food and Drug Administration. For suicide and depression treatment studies, we included any eligible psychotherapy or collaborative care interventions. For anxiety, we restricted nonpharmacological interventions to cognitive behavioral therapy (CBT). Eligible outcomes included test accuracy, symptoms, response, remission, loss of diagnosis, all-cause mortality, functioning, suicide-related symptoms or events, withdrawal due to adverse events, serious adverse events, and harms from screening. We also included systematic reviews reporting on harms of treatment. We excluded studies with poor methodological quality.
Data Extraction and Analysis: One investigator extracted data and a second checked accuracy. Two reviewers independently rated methodological quality for all included studies. When at least three similar studies were available, we conducted meta-analyses.
Data Synthesis: We included 80 studies (in 106 publications). No studies evaluated the direct benefits of screening compared with no screening or usual care. Seventeen studies reported on accuracy of screening instruments for one or more conditions; of these, one reported on suicide (N=580), 10 on anxiety (N=3,260), seven on depression (N=3,316), and two on anxiety or depression (N=695). Studies reported a wide range for sensitivity and specificity across a variety of instruments, with no more than one or two studies on each instrument. For suicide, sensitivity ranged from 0.87 to 0.91, and specificity was 0.60. For anxiety, sensitivity generally ranged from 0.34 to 1.00, and specificity from 0.47 to 0.98. For depression, sensitivity ranged from 0.59 to 0.94, and specificity from 0.38 to 0.96. Two RCTs (N=2,675) compared short-term distress from screening for suicide risk and reported no significant differences between those screened and those who were not screened.
Sixty randomized, controlled trials (RCTs) addressed benefits of treatment; of these, 16 reported on suicide risk interventions (N=3,034), 29 on anxiety treatment (N=2,970), 13 on depression treatment (N=2,156), and two on depression or anxiety treatment (N=236). Interventions addressing suicide risk or self-harm reported lower scores for the Beck Hopelessness Scale (pooled mean difference: −2.35 [95% confidence interval [CI], −4.06 to −0.65]; N=644; k=4) for intervention arms when compared with control arms. Findings for other measures were mixed or not statistically significantly different.
Of the 29 RCTs on anxiety treatment, 22 were on CBT; six were on pharmacotherapy; and one had multiple arms evaluating CBT, sertraline, and CBT plus sertraline. The evidence suggests CBT was associated with gains on several pooled measures of symptom improvement (magnitude of change varies by outcome measure), response (pooled relative risk [RR]: 1.89 [95% CI, 1.17 to 3.05]; N=606; k=6; I2=64%), remission (RR: 2.68 [95% CI, 1.48 to 4.88]; N=321; k=4), and loss of diagnosis (RRs range from 3.02 to 3.09), when compared with usual care or wait-list. The evidence on functioning for CBT was mixed. The evidence suggests pharmacotherapy, when compared with placebo, was associated with gains on two pooled measures of symptom improvement (mean difference Pediatric Anxiety Rating Scale: −4.0 [95% CI, −5.5 to −2.5], N=726, k=5 and mean difference Clinical Global Impressions-Severity: −0.84 [95% CI, −1.13 to −0.55]; N=550, k=4) and response (RR: 2.11 [95% CI, 1.58 to 2.98]; N=370; k=5) but was mixed on measures of functioning.
Of the 13 RCTs on depression treatment, eight were on psychotherapy; two on pharmacotherapy; one on CBT, fluoxetine, and their combination; and one on collaborative care. Results for psychotherapy varied by measure. Two pooled estimates suggested that psychotherapy is associated with improved symptoms (Beck Depression Inventory [BDI] or BDI-II standardized mean difference: −0.58 [95% CI, −0.83 to −0.34]; N=471; k=4 and Hamilton Depression Scale mean difference: −2.25 [95% CI,−4.09 to −0.41]; N=262; k=3), clinical response (3 studies with statistically significant results using varying thresholds), and loss of diagnosis (RR: 1.73 [95% CI, 1.00 to 3.00; N=395; k=4) but no statistically significant differences for other measures. The evidence suggested statistically pharmacotherapy was associated with improvement for one measure of symptoms (Children’s Depression Rating Scale-Revised [CDRS-R] mean difference −3.76 [95% CI, −5.95 to −1.57, N=793; k=3), and pharmacotherapy was associated with improvement for remission, but the pooled differences were not statistically significant. The single collaborative care trial (N=101) found that collaborative care was associated with improved symptoms at 6 months (CDRS-R change: 8.5 [95% CI, 13.4 to −3.6]), response by 12 months (odds ratio [OR] for ≥50% reduction in CDRS-R score: 3.3 [95% CI, 1.4 to 8.2], and remission (OR for Patient Health Questionnaire-9 <5 at 6 months: 5.2 [95% CI, 1.6 to 17.3]). The study reported no statistically significant benefits on measures of functioning.
Twenty studies (19 randomized controlled trials and 1 meta-analysis) addressed harms. Of these, two reported on suicide risk interventions (N=885), 11 on anxiety treatment (N=1,293), and seven on depression treatment (N=1,352).
Two RCTs of interventions to reduce suicide risk or self-harm reported no statistically significant differences in adverse events.
Of the 11 RCTs reporting harms of anxiety treatments, four evaluated CBT; six evaluated pharmacotherapy; and one evaluated CBT, sertraline, and their combination. The evidence from CBT studies yielded inconsistent results on suicide-related events; these studies also suggested lower rates of withdrawal due to adverse events and serious adverse events in the CBT arms. Suicide-related events and withdrawals due to adverse events in pharmacotherapy studies were rare and not statistically significant; however, they were more commonly reported in pharmacotherapy arms when compared with placebo arms.
Of the seven studies reporting harms of depression treatment, three evaluated pharmacotherapy; two evaluated psychotherapy; one evaluated CBT, fluoxetine, and their combination; and one evaluated collaborative care (1,276 from trials). Suicide-related outcomes, withdrawal as a result of adverse events, and serious adverse events were not statistically significant between study arms but were more frequent for pharmacotherapy when compared with placebo; inconsistencies in the evidence further reduced certainty. The evidence from the collaborative care study was inconsistent.
Limitations: No studies were available that reported benefits of screening compared with no screening. Limited evidence was available on harms of screening, long-term outcomes, test accuracy, and suicide risk and depression treatment in children. Treatment-as-usual comparators for suicide risk interventions included active treatments. The review was limited to drugs approved for pediatric use by the Food and Drug Administration (FDA). For anxiety, psychotherapy was limited to CBT.
Conclusions: We found no eligible studies that reported on benefits directly arising from screening when compared with usual care or no screening. Limited direct evidence suggests no short-term harms from screening for suicide risk. The evidence for screening for suicide risk, anxiety, and depression in children and adolescents relied on indirect evidence on the accuracy of screening and the benefits and harms of treatment. The evidence suggests that some screening instruments are reasonably accurate for anxiety and depression, but the evidence is limited for suicide risk screening instruments. Both pharmacotherapy and psychotherapy treatments have some benefit for some depression and anxiety outcomes (specifically, CBT for anxiety alone was reviewed); the evidence is limited for suicide risk interventions. Harms are rare in treatment studies but more frequent in pharmacotherapy arms when compared with placebo. Evidence gaps persist in children younger than age 11 years for test accuracy; depression and suicide risk interventions; and screening and treatment differences by sex, race/ethnicity, sexual orientation, and gender identity.