Software for Remote Clinical Trial Monitoring

The software receives electronic case report forms (eCRFs), operational data on the sites’ compliance with protocol SOPs, and source documents, such as informed consent forms (ICFs), lab test results, and X-rays, from integrated systems. The solution can ingest source patient health data from text, image, audio, and video files, along with the readings from wearable devices used for remote patient monitoring in clinical trials. Clinical data may also include standardized coding, such as ICD-10 for diagnoses, CPT for procedures, as well as structured CCDA summaries. Additionally, sites can upload scanned copies of paper source documents, such as reports from external laboratories.

Once ingested, the data is sent to a preprocessing module that automatically removes patient-identifying details, performs optical character recognition (OCR) for the content of scanned documents (including handwritten notes and signatures), and transcribes audio recordings, among other tasks. After that, the data goes into the landing zone of the data warehouse, which serves as a repository for raw trial data.

The batch data processing and analytics module extracts CRFs and preprocessed source documents from the raw clinical data repository (the landing zone) and performs automatic SDV/SDR. This includes verifying CRF data against the source documents (identifying transcription errors), checking CRF data quality (detecting blank fields, missing signatures and dates, inconsistencies, etc.), and confirming compliance with protocol and SOP requirements.

In addition to the batch processing and analysis module used for scheduled analysis, incoming CRFs and source documents are processed by the stream data processing module in real time. This enables CRAs to constantly monitor key risk indicators (KRIs) associated with site performance, trial performance, and patient safety. If potential high-risk threats are identified (e.g., previously unknown adverse effects), the module alerts sites and CRAs through a notification and alerting service.

After the batch and real-time source data processing, highly structured clinical data, which is ready for complex analytics, is stored in the presentation layer of the data warehouse.

The machine learning (ML) and artificial intelligence (AI) engine analyzes trends in CRF data to identify anomalies and outliers that may indicate issues with compliance and data quality at a particular site. For instance, lower adverse event rates at a specific site compared to others may suggest potential underreporting.

The ML/AI engine also processes incoming site data in real time to identify pharmacovigilance signals and compliance issues that require urgent attention.

Lastly, the engine uses data about identified errors, clinical data defects, and protocol deviations to assign a risk level to specific sites or document types. For example, it can assign a high-risk status to sites with frequent data gaps or with recurring discrepancies in investigational drug accountability records.

From the structured clinical data repository (the presentation layer of the data warehouse), the analyzed source data is directed to the business intelligence (BI) and reporting module, which visualizes and presents it to CRAs in the form of interactive dashboards. These dashboards offer convenient tools for data exploration (e.g., filtering, searching, and drill-down capabilities), report generation, and routine remote data monitoring.

All the CRFs and source documents with errors, defects, and anomalies that were flagged after automatic checks are available in a dedicated workspace for SDV/SDR. Here, CRAs can manually review the source documents that are critically important (e.g., serious adverse event reports) or have a high-risk status.

The BI module also sends real-time KRI updates and the results of ML/AI-driven risk analysis to the risk indicator review dashboard. Here, CRAs can dig into the source data associated with each potential risk, establish new KRIs, and adjust QTLs based on their findings.

Reports on the results of remote monitoring and audits with all the documented protocol deviations are transferred to the integrated clinical trial management system (CTMS), where they become available to site staff.

A project management module is connected with an issue management tool in the user interface, enabling CRAs to create tasks based on their action plans for mitigating risks and protocol deviations. These tasks become available to their team members, project managers, and data managers for assignment, execution, and monitoring.

Via an EDC query management module, CRAs can send queries to investigators to correct errors discovered during monitoring and then track the resolution of these queries. A video conferencing tool lets CRAs conduct remote monitoring visits to oversee SOP compliance and execute corrective actions, such as site staff training.