The Blau Center Data Science team is a multidisciplinary group of researchers, data scientists, engineers, and psychiatrists who combine efforts to better characterize the clinical trajectories of patients with mental illnesses, as well as to improve the existing electronic medical record (EMR) architecture currently available for psychiatric patients
EMRs are an invaluable source of information on millions of patients’ clinical states. However, for psychosis and severe mental illnesses, the characterization of standard of care, in terms of treatment protocols and outcomes, is complicated by stringent definitions and diagnoses that encompass heterogeneous presentations. The Blau Center’s team aims to improve the existing EMR architecture through systems capable of rendering it more accessible and navigable for clinicians while producing ready-to-use streams of data for secondary purposes.
Due to the nature of clinical psychiatric evaluations, which mainly rely on the direct observation of behaviors and symptoms, most information related to how psychosis presents in primary care, standard treatment protocols, and outcomes after treatment, is reported in raw-text clinical notes written by providers.
Despite the ubiquitous applications of natural language processing (NLP) methods on open-domain language, NLP tools able to extract the phenomenology, identify treatment patterns, and optimize the care flow of patients from raw text are still lagging behind. The Blau Center data science team is contributing to this effort through the implementation of machine learning and neural language models to extract, analyze, and interpret salient data of psychosis patients within EMR clinical notes. Such techniques can be used to process large datasets of thousands of patients and reveal important insights into real-world clinical practice within our health care system.
Data Science Activities
The Blau Center data science activity aims to serve as the Department of Psychiatry’s data science engine for advancing the research and treatment of mental illnesses in the Mount Sinai Health System. The data science initiatives at the Center include:
- Standardization of psychiatric intake forms to render the process more efficient, effective, and useful in terms of the data collected on Mount Sinai psychiatric outpatients.
- Implementation and maintenance of the Blau Center Clinical Database, which streamlines the data collection process at the Blau Center into a navigable and searchable tool where forms can be accessed and analyzed. These standardized data are then linked to EMRs to create a detailed characterization of patients’ interaction with the Health System.
- Generation of real-time statistics and clinical trends to help stakeholders in the Department of Psychiatry (e.g., department administrators, unit chiefs, clinic managers, clinical researchers) portray current standard of care for patients with mental illnesses.
- Clinical notes mining for investigation of treatment efficacy and safety.