Two abstracts presented by Holmusk at the Schizophrenia International Research Society (SIRS) 2022 meeting help to provide new insights into schizophrenia that could ultimately support more individualized care pathways.
Both studies drew large, deidentified datasets from NeuroBlu, Holmusk’s large and rapidly growing real-world dataset created specifically for behavioral health using de-identified electronic health records (EHR) data. NeuroBlu houses data from over 560,000 patients, providing the world’s richest source of both unstructured and structured data for a range of mental and behavioral health conditions.
One of the studies explored utilization of health care services among people who are receiving mental health care and who have also been diagnosed with a substance use disorder. The second study leveraged data analytics and natural language processing (NLP) to examine how different symptoms of schizophrenia impact two factors: scores on the Clinical Global Impression Scale (CGI-S) and number of days spent in a psychiatric hospital.
Both abstracts were led by Rashmi Patel, MD, PhD, Holmusk’s Vice President for Medical and Scientific Affairs.
How do substance use disorders impact health care utilization?
The first study included NeuroBlu data from over 20,000 patients. These patients were grouped into one of three broad types of mental health disorder: schizophrenia and related disorders, mood disorders, or other mental disorders. They also had one of four comorbid substance use disorders: alcohol, opioid, cannabinoid, or cocaine.
Of patients with substance use disorder, 8.4% had a diagnosis of schizophrenia or a related disorder, and nearly half (47.2%) had a diagnosis of a mood disorder. Individuals with substance use disorders and a comorbid mood or psychotic disorder were found to have a two-to-five-fold higher risk of visiting the emergency department with a year in comparison to people without a comorbid mental disorder. This could be indicate a higher risk of overdose or mental health crisis among patients with both a substance use disorder and a mood or psychotic disorder.
“Analysis of this real-world data highlights the impact of comorbid substance use and mood disorders or schizophrenia on clinical outcomes,” said Patel. “The combination of a substance use disorder and mood or psychotic disorder is associated with worse clinical outcomes. It will be critical to establish holistic services that ensure patients receive the care they need both for their mental health disorders and their substance use disorders.”
How do positive and negative symptoms impact measures and outcomes?
The second study gathered data from 4,400 patients using NeuroBlu and analyzed those data using proprietary natural language processing models (NLP). EHRs contains free text with critical information about symptoms that present in patients with mental disorders, and NLP can extract this information and convert it to structured data to enable comparisons between different symptom groups.
Schizophrenia is a serious mental illness characterized by positive symptoms (e.g. delusions or hallucinations) and negative symptoms (e.g. poor motivation, blunted or flattened affect, or social withdrawal) that may contribute towards substantial illness burden and poor functioning.
The study included data on 14 positive and 15 negative symptoms and found that patients with a greater number of positive or negative symptoms recorded in the EHR during their initial visit were more likely to have more severe illness at the outset as recorded by the CGI-S, and were likely to spend a longer time in a psychiatric hospital.
“We thank the Schizophrenia International Research Society for the opportunity to present these studies,” Patel said. “Using NeuroBlu, we generated real-world evidence that could be useful in informing care for patients with schizophrenia and developing better individualized treatment plans. EHR data and NLP have the potential to improve our understanding of modifiable factors associated with worse outcomes in serious mental illnesses that could help to develop more effective treatments and improve access to care.”