387K Behavioral Health Diagnoses: What Post-COVID Data Tells Us

The Dataset

The analysis covers 387,000 behavioral health diagnoses across a regional health system network spanning 14 outpatient behavioral health clinics, 8 primary care practices with integrated behavioral health, and 3 inpatient psychiatric units. The observation period runs January 2019 through December 2024, providing a complete pre-COVID baseline (2019) and four post-COVID years.

Diagnoses are drawn from encounter-level ICD-10 data and categorized using the DSM-5 aligned groupings. A patient with multiple behavioral health diagnoses contributes multiple records — the dataset counts diagnoses, not unique patients. This distinction matters: the growth in diagnosis counts reflects both new patients entering behavioral health services and existing patients receiving expanded or changed diagnoses.

Anxiety Disorders: The Dominant Clinical Story

Anxiety disorder diagnoses (ICD-10 F41.x: generalized anxiety disorder, panic disorder, specific phobia, social anxiety disorder, and related conditions) showed the single largest absolute increase of any diagnostic category: 67% growth from 2019 to 2023, with the rate stabilizing in 2024 at approximately 58% above the 2019 baseline.

The growth pattern is not uniform across the anxiety subcategories. Generalized anxiety disorder (F41.1) accounts for the largest share of growth, consistent with the sustained, diffuse nature of pandemic-related economic and health stressors. Panic disorder (F41.0) showed the sharpest early spike (2020-2021) and has partially retreated. Social anxiety disorder (F40.10-F40.19) remains elevated and, interestingly, has not declined to pre-pandemic levels despite the removal of most social distancing measures — suggesting that prolonged social isolation created lasting anxiety patterns for a subset of patients.

Who Is Seeking Behavioral Health Services Now

The demographic shift is substantial. Pre-COVID (2019), the 25-45 age group accounted for 28% of all behavioral health diagnoses in this dataset. By 2024, that figure is 42% — a 14-percentage-point shift that represents hundreds of thousands of patients nationally entering behavioral health services for the first time in mid-adulthood.

This shift has implications that extend beyond behavioral health specialty capacity. The 25-45 age group has historically been underserved by behavioral health infrastructure: too old for child and adolescent mental health services, too economically active to rely on public mental health programs, too stigma-averse (historically) to seek outpatient psychiatric care. The pandemic appears to have reduced the stigma barrier sufficiently that this demographic is now presenting to primary care and behavioral health services at rates the system was not designed to handle.

The Telehealth Shift: 3% to 71%

The most dramatic operational change in behavioral health delivery is the telehealth shift. In 2019, 3% of behavioral health encounters in this dataset were delivered via telehealth. By 2021 (the first full post-waiver year), 74% of encounters were via telehealth. By 2024, the number has stabilized at 71%.

This is not a temporary accommodation that will reverse as waivers expire. The behavioral health patient population has voted clearly for telehealth delivery, and the clinical evidence supports the choice: systematic reviews of telehealth behavioral health outcomes published through 2024 find comparable patient satisfaction and non-inferior clinical outcomes for most anxiety and depression presentations, with the significant advantage of dramatically higher appointment completion rates (telehealth no-show rates average 11% vs. 26% for in-person behavioral health appointments in this dataset).

For practice analytics, the telehealth shift creates new data infrastructure requirements. Telehealth encounters generate different encounter types in most EHR systems, require different quality measure denominator handling, and involve different billing codes (95 modifier vs. 02 place of service for most payers). Practices that haven't updated their analytics infrastructure to handle telehealth encounters correctly are reporting behavioral health quality measures with systematically miscounted denominators.

The Follow-Up Compliance Crisis

The Follow-Up After Hospitalization for Mental Illness (FUH) measure tracks the percentage of patients with psychiatric hospitalizations who have an outpatient mental health follow-up within 7 days of discharge. It is a HEDIS measure, a MIPS quality measure, and a CMS value-based care metric.

In this dataset, only 54% of patients discharged from inpatient psychiatric care completed the 7-day FUH follow-up. The national HEDIS benchmark for 7-day FUH follow-up is 38.2% (2024 NCQA data), which means this health system performs above the national average — but the absolute rate of 54% still means nearly half of all psychiatrically hospitalized patients are leaving without a confirmed follow-up appointment in the week most critical for preventing readmission and stabilizing their outpatient care.

PHQ-9 Severity Distribution: The Acuity Shift

PHQ-9 scores at initial assessment have shifted substantially toward higher acuity. In 2019, 22% of patients presenting to behavioral health services scored in the moderate-to-severe range (PHQ-9 10+). By 2024, that figure is 31% — a 9-percentage-point increase in the proportion of patients who are significantly symptomatic at presentation.

This acuity shift reflects two interacting phenomena. First, the behavioral health system absorbed a surge of new patients beginning in 2020, many of whom were presenting with more acute symptoms than the pre-pandemic baseline. Second, the system capacity did not expand proportionally — wait times for behavioral health appointments lengthened, which means patients who finally reach care are further along in their illness trajectory than they would have been with earlier access.

Practice Implications: What This Data Demands

Primary care practices cannot manage behavioral health at the scale this data describes without analytic infrastructure. A primary care panel of 2,000 patients now includes approximately 840 patients with a behavioral health diagnosis — 42% of the panel, if the population matches national trends. Managing depression follow-up, anxiety outcomes tracking, PHQ-9 trend monitoring, and FUH compliance for 840 patients requires data-driven workflows, not chart-level manual review.

The telehealth infrastructure dimension is equally important. Practices that haven't built analytical visibility into their telehealth encounter quality — comparing outcomes, completion rates, and billing accuracy between telehealth and in-person behavioral health encounters — are operating a major portion of their clinical program without quality feedback.

The post-COVID behavioral health landscape is not returning to 2019 norms. Planning for a clinical program that serves the 2019 volume is planning for a practice that will fail to meet its current patient population's needs. The data is available. The question is whether it is being used.