Economic Burden Associated with Negative Symptoms Identified Through Natural Language Processing Among Patients with...

The objective of this study was to identify documented negative symptoms of schizophrenia using natural language processing (NLP) and to characterize treatment patterns, healthcare resource utilization (HCRU), and costs among this patient population.

This US retrospective cohort study used electronic health records (EHR) linked to administrative claims data from January 2016 through February 2023. Adult patients (≥ 18 years) with at least 2 schizophrenia diagnosis codes were included. Negative symptoms were identified by NLP. Patient characteristics were assessed in the 12 months preceding the index date (first-documented schizophrenia diagnosis). Treatment patterns, HCRU, and costs were measured over the 12 months after the index date.

A total of 79,326 patients were enrolled in the EHR cohort and 14,992 (18.9%) had documented negative symptoms. Avolition was identified most often (44%) followed by blunted affect (42%). In the EHR cohort, 11,293 patients (14.2%) had linked claims. Patients with documented negative symptoms had more HCRU including days hospitalized, outpatient visits, and all-cause healthcare claims than patients without documented negative symptoms (all P < .001). They also had higher healthcare costs including inpatient, all-cause healthcare, schizophrenia-related costs (P < .001), and outpatient costs (P = .029). Only 34.6% of patients with documented negative symptoms received psychosocial interventions.

These observational data add to the limited published literature on negative symptoms in patients with schizophrenia in the United States. The association of negative symptoms with low utilization of psychosocial interventions, increased costs, and high healthcare utilization emphasize the need for better management and treatment options for patients with schizophrenia.