# Uncovering the Internal Structure of the German Version of the CORE‐OM: A Network Analysis

## TL;DR

This study uses network analysis to explore the structure of the German version of the CORE-OM mental health assessment tool and identifies four key areas of focus.

## Contribution

The study introduces a novel approach using network analysis and exploratory graph analysis to refine the structural validity of the German CORE-OM.

## Key findings

- The network analysis revealed high correlation stability with an average item predictability of R² = 0.42.
- Exploratory graph analysis identified four distinct communities in the German CORE-OM.
- Confirmatory factor analysis confirmed the EGA-derived model as the most parsimonious fit.

## Abstract

The Clinical Outcomes in Routine Evaluation – Outcome Measures (CORE‐OM) is a pantheoretical diagnostic instrument that has been widely used in mental health research. Nevertheless, the exploration of the factor structure of the CORE‐OM yields diverse results.

This study aimed to explore the internal structure of the German CORE‐OM using network analysis and compare several competing factorial structures of the CORE‐OM with traditional confirmatory factor analysis (CFA) to gain a more comprehensive understanding of its structural validity.

A total sample comprised 4496 (63% female) participants from an outpatient population. In a first step, we used network analysis (n
1 = 2248) to assess relationships between the items, followed by explorative graph analysis (EGA) to analyse community structure. Finally, we specified five competing models, including the one derived from the EGA, and used CFA in a second sample (n
2 = 2248) to identify the best‐fitting structure of the instrument.

The estimated cross‐sectional network demonstrated high correlation stability. The average item predictability was R
2 = 0.42. The EGA identified four distinct communities in the German CORE‐OM (General Problems; Interpersonal Problems; Positive Resources; Self Harm Risk). Confirmatory factor analysis showed that the EGA‐derived models had the most parsimonious fit.

These findings suggest a refined structure for the CORE‐OM, highlighting key item relationships and offering potential improvements for scoring and clinical use.