Team members from Holmusk recently attended the DGPPN (German Association for Psychiatry, Psychotherapy and Psychosomatics) meeting to share insights derived from real-world data drawn from Holmusk’s industry-leading NeuroBlu Database.

Impacts of COVID-19 on illness severity and care delivery

Researchers from Holmusk conducted a study to understand how the COVID-19 pandemic impacted illness severity and healthcare utilization.

The first part of the study analyzed changes in CGI-IS illness severity scores in patients with behavioral health conditions who had at least two recorded clinic visits prior to the pandemic and at least two visits during the pandemic. This analysis, which included data from over 7,500 patients from the NeuroBlu Database, found that about a quarter of patients experienced worsening CGI-S scores, 38.1 percent saw no change, and 36.6 percent showed improvement.  

The second part of the study examined types of health care visits among over 14,000 patients, again comparing a pre-pandemic period with a pandemic period. This analysis showed that most visit types declined (administrative contact, emergency, inpatient, and outpatient), while the number of tele-health visits increased.

“These findings suggest that the COVID-19 pandemic may have impacted the way in which mental healthcare is delivered,” said Emily Palmer, PhD, a research scientist for Holmusk and lead author of this study. “One limitation of the study is that all of its data is from the U.S. Because the European Psychiatric Association has found significant differences in service delivery during the pandemic among different countries, further analysis of additional real-world data sources is needed to better understand regional differences in these trends.”

Associations among symptoms, severity, and hospitalization

Another Holmusk study examined associations among symptoms, illness severity, and hospitalization. The study used data from 4,440 patients with schizophrenia drawn from Holmusk’s industry-leading behavioral health database, which includes real-world clinical data from over 1 million patients.

The study also leveraged unique features of the NeuroBlu Database. Because the database includes Clinical Global Impression-Severity (CGI-S) scores for the majority of its patients, researchers were able to examine outcomes. They were also able to design the study to use the database’s rich symptom data, which is extracted from free text using natural language processing models and transformed in to research-ready variables.

Schizophrenia is a heterogenous condition which can present through many different symptoms. This study, which used data on 14 positive and 15 negative symptoms, found that people with more symptoms were more likely to have higher CGI-S scores, or worse illness severity. Patients with higher symptom burden were also more likely to have longer stays in the hospital than those with lower symptom burden.

“Using our natural language processing model, we were able to extract information about symptom burden from EHR data,” said Miguel Rentería, Principal Data Scientist and one of the study’s authors. “This is a valuable method that could be used in the future to identify patients who could benefit from improved outcomes if their specific symptoms were treated.”