Lowenstein · Pediatrics 2026 · interrupted time-series study · n=?

Trends in Youth Mental Health Among Patients in Family Medicine Practices: 2017-2023.

Level 3 - non-randomized controlled study

Interrupted time-series study using electronic health record registry data

PubMed 41967839 · doi:10.1542/peds.2025-072305 · record verified 2026-08-26

What was done

Using electronic health record data from a US national primary care registry spanning January 2017 through December 2023, authors conducted an interrupted time-series analysis evaluating monthly visit rates for mental health diagnoses (anxiety, affective mood disorders, ADHD, trauma/stressor-related disorders, behavior/conduct problems, and eating disorders). They compared the level and slope of diagnosis rates across three periods: prepandemic (January 2017–February 2020), the first pandemic year (March 2020–February 2021), and post-first year pandemic (March 2021–December 2023).

What was found

Average monthly visit rates with anxiety diagnoses increased significantly from pre- to postpandemic periods, rising by 71% among females aged 12 to 17 years and by 62% among males aged 12 to 17 years. In older females, rates of mood disorder diagnoses remained 35% higher and ADHD diagnoses remained 8.5% higher one year after the pandemic onset. However, despite these sustained increases, overall monthly rates began decreasing significantly during the postpandemic period.

Why it matters

This study shows that pandemic-associated increases in youth primary care mental health diagnoses disproportionately affected older adolescents—particularly females—and provides longitudinal evidence that diagnostic visit rates have started to revert toward prepandemic baselines.

Limits

The total sample size, patient count, and number of contributing practices are not reported in the abstract. The data rely on clinical diagnostic coding in primary care electronic health records, which reflects healthcare-seeking behavior and clinician documentation rather than population-level symptom prevalence, and unmeasured shifts in telemedicine or coding practices during the pandemic could confound trends.