ThoughtSphere Platform Release Notes v1.16

We are thrilled to announce the latest enhancements to the ThoughtSphere Platform in version 1.16. This release introduces more advanced ML automation, expanded query management capabilities, a new Listing Review Experience, and Python support in the Clinical Mapping Workbench. These enhancements boost performance, accuracy, and workflow efficiency.

Automated Trend & Outlier Detection Models

• Missing Visits vs Discontinued Subjects Algorithm: Configurable ML model that analyzes correlations to proactively identify subjects at risk of discontinuation based on missing visit patterns.
• Adverse Events vs Missed Visits Analysis: New statistical ML model for analyzing correlations between adverse event counts, and missing visits per subject to identify safety trends and possible areas of AE under-reporting.
• Lab Outlier Detection Dashboard: A new Lab Outlier Detection dashboard to detect outliers using the IsolationForest algorithm.
• Automated Actions from Statistical Monitoring: All statistical Monitoring models (including those above) can now be configured to auto-create Actions when outliers and anomalies are detected, to further automate statistical monitoring of studies.

Clinical Data Management Upgrades

• New Enhanced Listing Review Experience: To increase communication and capture of insights during review, a tabular format is now reviewable for Listing Reviews. In addition, users can now track Queries, Actions, and Comments at the row-record level. Metrics for pending Actions, Queries, and Changed Records (since last review) are listed to guide the user to review areas of interest.
• Enhanced Query Management & EDC Integration: End-to-end query lifecycle support with automated EDC query closure, bidirectional sync status tracking, improved visibility into related queries during discrepancy review, and expanded reporting on query writebacks to EDC systems.
• Improved Data Check Configuration Experience: New drag-and-drop rule setup, support for boolean expressions, and in-line error highlighting make it faster and easier to build, validate, and troubleshoot Data Checks and Reconciliations.
• View SDTM Data in Discrepancy Workflows: A dedicated tab for viewing SDTM Data during Discrepancy Reviews has been added so transformed data can always be viewed when available.
• Advanced Smart Check Capabilities: The Smart Check model now predicts checks regardless of matching contextual data, allowing wider coverage of predicting data checks.
• Expanded Clean Patient Tracker Configuration & Dashboards: Users can now assign subjects to the CPT using rules based on status, visit, and site, with improved dashboards offering detailed cleanliness insights for both Cut-off and Subset views.
• Persisting Filters in Data Check Inventories: Study and CtQ Filters persist across My Check Reviews, All Check Reviews, Unassigned Check Reviews (likewise for Reconciliations).

Data Ingestion & Outbound Configuration

• Extended Outbound Exports for ADaM: DTYPE columns are now included in outbound files across all supported formats (CSV, XPT, SAS).
• Incremental Metadata Ingestion in ODM: Clinical metadata ingestion now supports incremental ingestion to optimize data synchronization workflows (CDISC 1.2 & 1.3, Rave, Inform ODM Metadata).
• Updated Python Library for XPT v8: The outbound configuration now uses an updated Python library to support export of XPT files in version 8 format.
• ADaM Dataset Ingestion Support: Implemented ingestion pipeline support and configuration for ADaM datasets. ADaM data can be loaded as-is while skipping the initial mapping steps.

Klini (Clinical Chatbot) Upgrades

• Reasoning & Memory Upgrades: Klini, the LLM-powered clinical and operational analytics chatbot, now features expanded memory handling, more powerful study-level querying, cross-study data aggregation, and enhanced reporting for missing data.
• User Driven UI Enhancements: Klini can now visualize queried data as Line, Bar, Pie, and Column charts, with options to select axes in the charts. Klini now offers plain English explanations of how data was queried/analyzed based on the user prompt. The study context is called out for the user, and drill downs are available to relevant Clinical Analytics in the platform.
• FAQ Support: Klini now allows definition and maintenance of FAQs for improved user guidance, at the Role level or Global level.

BI Designer & Dashboards

• New Visualizations & Chart Enhancements: Added support for Sunburst charts, multi-series Pie charts, and a color gradient-enabled World Map; Dynamic Series can now be split into individual graphs for easier analysis and export.
• Improved Program Oversight & Monitoring: The Program Oversight Dashboard now includes an enhanced Monitoring Plan by CtQ with a new Sunburst chart format and updated % Study Enrollment derivation.
• BI Designer Usability Upgrades: Standard Data Catalogs used in visualizations are now displayed in the BI Designer, and chart exports reflect updated naming for separated series.
• Expanded Study-Level Insights: A new Study Calendar has been added to show form-event schedules, and the Rejected Records Report now includes more detailed headers and record-level data.

Clinical Mapping Workbench

• Python Language Support in Clinical Mapping Workbench: Python can now be used to write advanced expressions when mapping data through the Clinical Mapping Workbench. The Out-of-the-box library of mapping functions, along with test execution capability, are available in Python to support the mapping process.
• ADaM DTYPE Mapping & Execution Support: The platform now supports ADaM DTYPE mapping and derivation, including import/export of DTYPE metadata and record creation via standard functions. DTYPE mappings can also be tested using the Test Execution workflow, with DTYPE-derived records clearly highlighted in the preview.
• Mapping Workbench Enhancements: The Clinical Mapping Workbench now features multi-column lookups via the enhanced $search function, a new $formName function for advanced multiple column lookups and form-level data logic. Ready-to-use wrapper functions for date retrieval and timestamp conversion were added. Additionally, R library functions now include tooltips to guide proper usage within the workbench.
• Standardization of Item Names Across Platform: Display of Item Name vs Item OID throughout the platform has been standardized for consistency and easy viewing (including in OOB dashboards).
• Improved ADaM Define Experience: ADaM Define configuration has been aligned with SDTM Define configuration for a more consistent and user-friendly experience. ARM capabilities have been added to support outputs from Pinnacle 21.

Want to Learn More? To all our existing customers, let’s talk about the next steps to updating your platform version so you can capitalize on all the new features and capabilities! Please reach out directly to your customer support lead or contact us at support@thoughtsphere.com. To non-existing customers that want to learn more about our latest release and see a product demo, reach out to our sales team at sales@thoughtsphere.com. We look forward to connecting with you!