Identification and Stratification

Identifying Need, Matching Care

“Enhancing Depression Identification and Stratification with a Claims-Based Analytical Framework,” The Journal of Behavioral Health Services & Research, 2025.

Why severity matters

Depression care covers a wide range of interventions, from digital self-care to intensive clinical treatment. The stepped care model aims to match that support to the person’s level of need.

The potential benefit runs across the system. Members can receive help earlier, insurers can direct care more effectively, and healthcare providers can balance demand across interventions of different intensity. But that depends on answering two questions: who needs care, and how much?

The diagnosis does not tell the whole story

The information needed to answer those questions is often already present, but scattered. Screening scores, diagnosis codes, medication records, and treatment histories each describe part of a person’s condition. Relying on one indicator can leave people unidentified or give an incomplete picture of their severity.

The Identification and Stratification framework—ID&Strat—was developed around a straightforward premise: those partial views should be considered together.

We examined whether combining available claims and clinical indicators could improve population-level identification and provide a more complete assessment of depression severity than individual measures alone.

What combining the evidence changed

The published study identified more than 258,000 additional members compared with diagnosis codes alone. Combining indicators also assigned higher severity to many cases initially classified as mild or moderate by a single measure.

The differences mattered beyond classification. Average total healthcare expenses for the severe group were approximately 2.5 times those of the minimal-severity group, showing how depression severity relates to the broader medical burden.

Who was identified?

720,882 total
462,676Recorded depression diagnosis
+258,206No recorded diagnosis

Who was reclassified?

Share assigned a higher severity than indicated by a single measure

Initially mild46%
Initially moderate19%
0%50%100%
Popkov et al. (2025) · 2022 Highmark Health data · doi:10.1007/s11414-025-09972-0

A foundation for stepped care

The contribution is a documented framework for connecting fragmented evidence and making differences in need visible across a population. That creates a basis for studying how lower-intensity support can reach people with milder symptoms while more intensive resources are directed toward those with greater needs.

Identifying more people is only part of the purpose. The larger question is whether a clearer understanding of need can help the healthcare system provide appropriate care earlier.

Published paper: “Enhancing Depression Identification and Stratification with a Claims-Based Analytical Framework,” The Journal of Behavioral Health Services & Research, 2025. Read the full article.

Conference presentation: “The Benefits of a Complex Identification and Stratification Model for Depression: Insights from a Large Claims Analysis Study,” ISPOR 2024; abstract published in Value in Health. The model was subsequently optimized for the published study.