Biotype-assigned Augmentation Approach in Resistant (late-life) Depression (BAARD)


Background: Depression in older adults often proves challenging to treat, with up to 50% not responding to initial antidepressant therapy and fewer than 20% achieving remission. Utilizing a robust multi-site clinical trial network, we have conducted two landmark randomized controlled trials (RCTs) that demonstrated the efficacy of augmenting antidepressant treatment with aripiprazole (ARI) or bupropion (BUP). These interventions achieved a 29% remission rate in treatment-resistant late-life depression (TRLLD). 

Objective: Building on the findings of the OPTIMUM and OPT-NEURO studies, we aim to advance the BAARD study whos primary objective is to improve treatment selection for late-life depression (LLD) by leveraging precision biomedical information. 

Data Integration and Analysis: The BAARD initiative seeks to develop a clinical decision support tool to enhance treatment selection for TRLLD. This effort integrates diverse expertise, including geroscience, psychopharmacology, cognition, molecular subtyping, neuroimaging, computational psychiatry, and qualitative research methodologies. 

Significance: To achieve these goals, we will utilize comprehensive demographic, clinical, cognitive, genetic, proteomic, and neuroimaging data from approximately 700 participants. The development and testing of the BAARD tool aim to significantly improve remission rates and transform care delivery for this vulnerable population.

​​Collaborating Institutions: The BAARD project is an international research initiative that leverages multi-modal data - encompassing clinical assessments, neuroimaging, biomarkers, and genomics - through the collaboration of leading institutions including Washington University, Columbia University, UCLA, University of Pittsburgh, and CAMH.

Link to NIH Project details:  1UG3MH137353-01


Directory Structure on SCC 

Path:  /external/rprshnas01/netdata_kcni/dflab/data/BAARD

Description: 

This directory contains data and documentation for the BAARD (Biotype-assigned Augmentation Approach in Resistant Depression) study. It is structured to organize raw, processed, and reference datasets across multiple data modalities, including clinical assessments, blood biomarkers, genotype data, neuroimaging, and neurocognitive assessments. Files are organized into subdirectories by data type to support analysis and collaboration. 

clinical/

  • Description: Contains demographic, clinical, and symptom-related data collected from study participants.

  • Subfolders:

    • raw/: Original clinical data files as collected.

    • processed/: Cleaned and standardized clinical datasets ready for analysis.

    • readme.txt: Detailed documentation on clinical variable definitions and data preprocessing steps.

mri/

  • Description: Includes structural, functional, and diffusion MRI data.

  • Subfolders:

    • raw/: Unprocessed neuroimaging data in original formats.

    • processed/: Preprocessed MRI data (e.g., motion-corrected) suitable for analysis.

    • readme.txt: Information on scan protocols, file formats, and processing pipelines used.

neurocognitive/

  • Description: Contains results from standardized neuropsychological assessments.

  • Subfolders:

    • raw/: Original scores and raw cognitive test data.

    • processed/: Cleaned and harmonized datasets for analysis.

    • readme.txt: Includes cognitive task descriptions and scoring methodology.

blood_biomarkers/

  • Description: Blood-based biological measures including inflammatory markers and other proteomic data.

  • Subfolders:

    • raw/: Original laboratory data.

    • processed/: Quality-controlled and normalized biomarker data.

    • readme.txt: Details on sample processing, assay methods, and units of measurement. 

genotype/

  • Description: Genomic variant data.

  • Subfolders:

    • raw/: Original genotype data (in VCF file format).

    • processed/: Filtered genotype data with covariate information.

    • readme.txt: Details on preprocessing and analysis. 

Dashboard Code Repository

An interactive dashboard used to support decision making in the development and evaluation of the BAARD tool.

🔗 https://github.com/BAARD-CAMH/dashboard

Data Analysis Code Repository

All scripts and code utilized in the processing, harmonization, and analysis of multi-modal datasets for the BAARD project.

🔗 https://github.com/BAARD-CAMH/data-analysis

Data Modelling Code Repository

This repository contains scripts for implementing predictive modelling approaches for the BAARD project.

🔗 https://github.com/BAARD-CAMH/modelling


  • No labels