All content copyright, Creative Commons Attribution-Noncommercial-Share Alike 3.0
Unported, with the attribution: "The Lead CRA Blog", and a link to the post.

Thursday, June 13, 2013

Risk Assessment in Monitoring

Clinical trial monitoring necessitates timely and accurate risk assessment. Risk assessment has become particularly relevant given that industry pioneers are increasingly and aggressively considering a risk-based monitoring paradigm. This paradigm is endorsed by consortiums like TransCelerate and supported by guidance issued from regulatory agencies such as FDA and EMA. So, how should a CRA assess risk accurately? What should be the balanced approach for site monitoring given that it is a cost intensive exercise with patient safety and quality being at the core of it?

Framing and classifying risk correctly

Acknowledgement of risk in the right context is important prior to its assessment.   Known or fixed risk factors at a program or protocol level should be considered first. At a program level, these risk factors could pertain to organizational dynamics, operational complexity (nature and extent of outsourcing), investigational product class, logistics and supply chain. Protocol level risk factors (such as TA, number of invasive procedures etc.) and site-specific risk factors (such as site experience, historical site performance etc.) should be accounted for, before developing the monitoring plan. Dynamic or unknown risk factors should be looked at using a relevant set of risk indicators to help evaluate the unknown risk elements during the progress of a study.

Using the correct risk indicators

It is important to identify risk indicators that serve as the right trigger to recalibrate monitoring activities. These indicators should be direct or surrogate pointers to   issues and risk that arise during the course of a study. A reasonable set of indicators for tracking site and study performance could be centered around quality, timeliness and efficiency. Timeliness of data entry, query resolution timelines, number of protocol deviations and adverse events; screen failure, early termination and site enrollment rates, are all good candidates for ‘signal detection’. Analytics-driven dashboards that provide trends and benchmarked assessment of these risk indicators for all sites in a study and that help identify outliers across the study performance spectrum are excellent tools to leverage for better risk assessment. Site performance across these indicators should ultimately convey a message that builds trust of regulatory agencies, in the quality of data being captured and reported during a clinical trial.

Leveraging technology to improve risk assessment

Successful adoption of technology, data integration from disparate sources and leveraging analytics optimally is critical for improving monitoring efficiency.  Cohesive technology-based monitoring solutions can assist greatly in early and ongoing risk assessment, by leveraging the real power of EDC systems to focus on critical processes and critical data. These targeted monitoring solutions should link protocol objectives to endpoints to ascertain risk and enable adjustment of monitoring activities in real-time. Leveraging technology-based solutions early on, through a quality by design-based approach during protocol and monitoring plan development, is ideal to mitigate risks at the later stages of a trial. Adoption of EDC systems supports remote and continuous transcriptional verification of clinical data; rather than as a retrospective exercise.  This keeps the quality and costs in control while keeping focus on aspects that “matter more”.

Verifying or reviewing the relevant information

The practice of doing 100% SDV is now being seriously questioned from a need and ROI perspective. Data analytics from the Medidata Clinical Cloud™ - based on thousands of clinical trials across the globe indicates that nearly 97% of clinical data entered for the first time is available for downstream use in its final form. Initiatives such as TransCelerate support that SDV has minimal effect on data quality – only 7.8% of total queries are SDV-generated and a smaller subset (2.4%) of queries are generated for critical data. Monitoring activities should cover both transcriptional verification (Source Document Verification or SDV) and SDR (Source Document Review), the later revolving around investigator involvement, protocol and guideline compliance. SDR is more important from a quality and regulatory approval perspective and therefore technology-enabled offerings which holistically address these aspects of the clinical development spectrum are likely to be more valuable to site monitors and CRAs.

Managing and mitigating Risk

Ongoing assessment of risk based on pre-defined risk indicators, critical processes and data should be the foundation of any monitoring plan. When needed, monitoring activities should be scaled up with the expectation of returning to the standard level in future. This return to baseline should occur post a root cause analysis of issues, thereby mitigating future risk. The approach of recalibrating verification and review activities should be closely aligned with the technological capabilities before the start of a study. It is important for monitors to reinforce expectations around data entry and query resolution timelines and ensure adherence to predefined SOPs and workflows stated in integrated risk management plans. This holistic approach involving data-driven process change and technological enablement, which starts early on with protocol design and site engagement will empower monitors better to perform their duties effectively.

Author information:

Associate Product Manager, Medidata 

Geeks Talk Clinical
Geeks Talk Clinical Medidata Blog

Special Thanks to today's guest contributor, Medidata.  Will you add to the conversation?  I invite you to author your own Lead CRA post. I'm accepting new guest contributors now, send me an email if interested.

Further reading at the ClinOps Toolkit blog:

You may also like...from The Lead CRA archives:


NadiaBoBadia said...

Thanks again to Shantanu and the Medidata "Geeks Talk Clinical" blog team for this summary of risk-based monitoring. The role of monitors has to change. The hypothesis is that data quality can be achieved without 100% SDV; it will be interesting to see this in practice.

I for one, still have a lot of questions about centralized monitoring, how to select which items to target for partial SDV (looking for different data patterns and risks), a discussion of other diverse monitoring methods, how will sponsors track their central monitoring efforts and reconcile that everything was completed (how will they leverage technology to do this?), etc.

I plan to address these questions and more in future blog posts here and at the Clinical Operations toolkit blog - stay tuned!

Kelly Gagnier said...

Great Summary!
I am a Senior Site Manager and I am involved in a task force for Risk Based / Adaptive Monitoring. I was fortunate to be involved in the Transcelerate Initiative for the Restrospective review of SDV vrs queries generated for critical data. It was quite fascinating. My company has a few studies with the Adaptive Monitoring Approach and so much can be reviewed centrally. It is important to have the large picture and ensure the site is following the protocol with no major deviations and that patient safety is #1. CRAs have evolved from 100% SDV workers to more in depth Site Managers.

NadiaBoBadia said...

Hi Kelly, I am so glad you stopped by to check out the blog. Your background on the task force is very interesting to me and I would love to network further.

I welcome tangible metrics and I was glad to see the retrospective review of SDV vs. queries generated. An additional metric that I plan to explore on the ClinOpsToolkit, would be to also account for the amount of entry that happens prior to and during SDV when the monitor picks up on & flags safety or other data that has not been entered. Many items aren’t even queried in the system (just flagged for entry in the source document review instead). A significant amount of under-stated data revision (or entry) occurs because of attention, conversation, and action taken during the monitoring visit. I don't see any of this accounted for in the metrics Transcelerate provided.

For this blog, I am still looking for a guest contributor to write about the CRA perspective of risk-based monitoring, perhaps you would know someone who would be interested? Please email me.

I've written a response from the Trial Manager perspective for Risk-Based Monitoring at the ClinOpsToolkit blog. I'd love for you to check it out.

Finally, I'm not sure that I agree that all CRAs necessarily have evolved from 100% SDV to Site Managers, but I am delighted to hear of the progress your company has made with adaptive monitoring. I will be doing my part to assist that CRAs who read this blog hone the correct Site Manager skillset.

Post a Comment

Note: Only a member of this blog may post a comment.