Call us: + 31 (0)20 - 4350 580

Home > News > Blog > The growing challenge of handling multistream d...

The growing challenge of handling multistream data capture in clinical trials - ACDM 2019

- José Matamoros Luna MSc, Clinical Data Manager
The growing challenge of handling multistream data capture in clinical trials - ACDM 2019

This year’s annual conference of the Association for Clinical Data Management (ACDM) took place in Amsterdam. During a two-day event, the emphasis was on the growing challenges of multistream data capture in clinical trials, with additional focus on risk based monitoring (RBM) and regulatory inspection findings.

In line of last year’s conference topic about Electronic Health Records (EHR), Martin Douglas (University of Münster) called attention to the huge amount of data collected in EHRs and in Electronic Data Capture (EDC) systems; electronic Case Report Forms (eCRF). He stressed that there is an overlap between documentation in EHRs and eCRFs, specifically for clinical research, which makes it difficult to integrate data in compatible medical data models. Douglas opted for transparency on medical eCRFs used in clinical research as well as routine documentation forms used in EHRs. He is involved in creating an open access platform, called The Portal of Medical Data Models. This platform aims to create standards in data elements and code lists, which will benefit standardizing clinical research performed in an academic setting.

Good Clinical Practice (GCP) inspections are conducted in all lines of work in clinical research, yet the findings are not publicly discussed at conferences. Instead, Andy Fisher (MHRA Inspector) exchanged several GCP findings, which are relevant for clinical data managers (CDMs). One relevant finding is based on data quality issues between data from source and data entered in EHR/eCRFs. If no correct quality check or validation process is in place, it can lead to high amount of variability between data, resulting in discrepancies of clinical trial endpoints. Additionally, quality controls are not described properly in data management plans, leading to complication on which key data should be of high quality in order to take key decisions. An example of this are dose escalation trials, which are commonly performed in early phase (oncology) trials. Data cleaning or adequate quality control process, such as targeted monitoring visits, proper SDV or 100% data cleaning on safety data, are important in decision-making. Results from dose escalation (oncology) trials are crucial to make dose escalation decisions and determines whether a drug is safe to check for efficacy in consequent phase II trials. Another finding related to data entered in eCRFs was regarding the data audit trail. From an inspection point of view, having the data audit trail presented in a dynamic format compared to a stand-alone fixed file (e.g. PDF-format) adds more value for assessing compliance. Applying GCP requirements correctly to clinical trials is vital in order to draw correct conclusions.

Handling multiple data from several streams (e.g. eCRF data, advanced medical technology wearables, vendor data) is becoming more challenging for CDMs according to Kirsty Millar (Phastar). Safety, statistics, EHR, biomarkers and PK/PD analysis data has to be reviewed, meaning that it is the responsibility of the CDM to set-up multiple agreements with each data stream. From project start, it is important to ensure that CDM responsibilities and communication paths are in place. Risk management is key to identify risks (e.g. data quality, timelines, compatibility and integration of data), such as data security, quality and reconciliation of data with all parties. It is significant to ensure that early engagement with vendor is in place to prepare comprehensive data transfer and reconciliation plan with adequate timelines so errors are avoided. Plan a data strategy to reduce duplication, and propose to not collect data in the eCRF, which is available in another stream. In the future more streams of data will be introduced (e.g. big data, EHR, eSource, artificial intelligence), making it pivotal for CDMs to be early involved in trials to assess risks. 

Source: Sciencesoft.com

Joke de Wever (BDLS) emphasized that data visualization can play an important role, especially when handling data from different sources, to support decision-making and increasing compliance in a timely manner. To get more added value from data, visualization techniques are used to assess patterns, outliers, trends as well as target bottlenecks. By creating dashboards specifications (queries, safety or enrollment data), which can be used by CDMs, CRAs and Medical Monitors, the risk management process can be improved to support decision making. Benjamin de Vriendt (BDLS) continued to explain that using a standard Study Data Tabulation Model (SDTM) structure to present data (at country, site, visit and patient level), allows pooling of data from multiple EDC systems. If standard structures are present, it would allow across-study reporting. Multi-level risk management would be available to assess issues on a study, compound and therapeutic area level.  

New indications and high variable indications within oncology trials lead to several issues within the data management process (e.g. protocol requirements, design of data collection/validations, complex data review, increasing protocol complexity).           Improving the efficiency and effectiveness of data management is crucial according to Maria Craze and Viviana Rodriguez from MSD. Both opted several approaches to tackle the challenges mentioned above. Firstly, it is important to develop robust but flexible validation tools. By creating quality metrics, setting improved timelines of new components requirements (submission – testing – delivering) and increasing flexibility in validation tools. Another approach would be to improve acquisition and management of complex high volume external data types. By re-designing external data processes, such as setting up-to-date requirements, validation tools, processes and responsibilities for new data types. Lastly, Viviana mentioned that improving knowledge and expertise on SDTM, CDISC and oncology is important for CDMs to work efficiently in oncology trials. CDMs should have continuous training in E2E, SDTM, CDISC and data validation tools to ensure the high quality of data in oncology trials.

The increase in data volume coming from multiple data sources will have a significant impact in the work of CDMs, resulting in an increasing involvement for them within clinical trials. Usage of different methods of risk based monitoring approaches (data visualization tools) as well as following GCP compliance and constant training are important factors to deliver quality data in clinical trials.

As one of the few oncology CROs in the world, SMS-oncology is an expert in developing oncology specific eCRFs and databases. We work according to CDISC standards by default to ensure our clients always have access to SDTM/ADaM datasets required for authorization purposes - as quality of data is the number one priority.