Integrating clinician and patient case conceptualization with momentary assessment data to construct idiographic networks: Moving toward personalized treatment for eating disorders

Abstract

Eating disorders are serious psychiatric illnesses with treatments ineffective for about 50% of individuals due to high heterogeneity of symptom presentation even within the same diagnoses, a lack of personalized treatments to address this heterogeneity, and the fact that clinicians are left to rely upon their own judgment to decide how to personalize treatment. Idiographic (personalized) networks can be estimated from ecological momentary assessment data, and have been used to investigate central symptoms, which are theorized to be fruitful treatment targets. However, both efficacy of treatment target selection and implementation with 'real world' clinicians could be maximized if clinician input is integrated into such networks. An emerging line of research is therefore proposing to integrate case conceptualizations and statistical routines, tying together the benefits from clinical expertise as well as patient experience and idiographic networks. The current pilot compares personalized treatment implications from different approaches to constructing idiographic networks. For two patients with a diagnosis of anorexia nervosa, we compared idiographic networks 1) based on the case conceptualization from clinician and patient, 2) estimated from patient EMA data (the current default in the literature), and 3) based on a combination of case conceptualization and patient EMA data networks, drawing on informative priors in Bayesian inference. Centrality-based treatment recommendations differed to varying extent between these approaches for patients. We discuss implications from these findings, as well as how these models may inform clinical practice by pairing evidence-based treatments with identified treatment targets.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 34 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.