Friday, July 19, 2013

Unraveling Risk-based Monitoring Terminology

“Judging by the sheer number of conference presentations, forum discussions, and webinars, not to mention all the new vendor offerings built to facilitate the process, risk-based monitoring might be one of the most frequently discussed topics in clinical research today.”

That sentence opened our December blog post on risk-based monitoring, and it’s as true today as it was then. (In fact, I’ll say it’s even truer today, if only to spark debate about truth being an absolute.)

The more a topic is discussed, the more words people use to discuss it.  Of course, a few subject-specific, well-defined terms can facilitate conversation, but over time, meanings blur.  Industry-established definitions give way to corporate-centric, implementation-specific usage.  We’ve all seen this recipe for miscommunication before.  The German and the Frenchman know that they speak different languages and can take measures to assure accurate translation.  Two people who speak the same language and use the same words may not even realize they might not be saying the same thing.

This blog post looks at the terminology used in what are arguably the three most important and oft-quoted papers on Risk-based Monitoring in the industry:

·        FDA draft guidance Oversight of Clinical Investigations – A Risk-based Approach

·        EMA Reflection Paper on Risk-based Quality Management in Clinical Trials

·        TransCelerate BioPharma Position Paper on Risk-based Monitoring Methodology

I carefully reviewed these three documents looking for consistencies on which we can rely, and inconsistencies of which we need to be cautious.  For each set of terms below, I first present the conclusions I’ve drawn about usage consensus, or lack thereof, and then suggest some tips to avoid confusion in discussions you might have with colleagues and clients.  The sane among you will probably be interested in only my conclusions and suggestions; you linguists, skeptics, and masochists may be interested in the supporting details I added to the end of the post.

Sorting out terminology and definitions is a tricky business.   If you have different interpretations than I present here, or can add to this analysis, please post a comment and we can start a discussion.

Remote versus Off-Site Monitoring

Sources agree that these terms are synonymous and used to describe traditional on-site monitoring activities that, thanks to technology, are now able to be conducted elsewhere.

You can use these terms interchangeably, but continue reading to avoid confusion with the term “central”.

Remote/Off-Site versus Central Monitoring
Now that we’ve established that remote monitoring and off-site monitoring can be used interchangeably, let’s compare the pair to central monitoring.

Sources agree that central monitoring includes remote/off-site monitoring activities, but includes other functions, as well, predominately statistical analysis. 

Sometimes the literature uses “remote/off-site” and “central” interchangeably, but that’s risky.  Due to its statistical component, “central monitoring” has the widest range of interpretations of all the terminology in use, and some of the interpretations are very different from the comparatively straight-forward “remote/off-site monitoring” definition.  Read on.

(BTW, FDA uses “centralized”; TransCelerate uses “central”.  Tomato, tomahto.)

Central Monitoring and Statistical Analysis

“Central monitoring” implies that data across study sites are collected centrally, and available for statistical analysis.  The type and extent of  statistical analysis that is performed to adapt a monitoring plan can vary significantly.   Sometimes analysis is limited to Key Risk Indicators or critical data points, and sometimes more rigorous statistical testing is performed on every data item, as any outlying element can be indicative of data quality.

When using the term “central monitoring” outside a narrow culture, make sure to clarify whether you mean a slight superset of remote/off-site monitoring, statistical analysis of Key Risk Indicators, or something more.  Different computing platforms offer different features; if it’s appropriate, mentioning the IT platform that is being used could clarify the type of central monitoring being discussed.

Adapted versus Adaptive Monitoring
If you’ve stuck with this analysis thus far, you deserve some lighter commentary…

Usage is too loose to be of much danger.

Use interchangeably with giddy abandon.

Centralized versus Centralised Monitoring

It depends on which side of the pond you are.

Stick to verbal discourse and there is no difference.  (Tomato, tomayto.)

Closing Thought (before I lose you all)

My college French professor was the first to point out to me that irregular verbs, the verbs that refuse to follow regular conjugation patterns, are the most common verbs in every language – the verbs to be, to go, to do, to have, to see, etc. They’re used so often that they bend and blur and evolve on their own. I think the jumble of definitions that arises with an oft-discussed topic like risk-based monitoring is somehow analogous to the unstructured forms of irregular verbs. The more they get thrown around, the more they go their own way.
By Laurie Meehan

We at Polaris hope you find this information helpful. You can visit us at our website, or you can contact us at
Supporting Detail (as promised/threatened)

Remote versus Off-site Monitoring:
The FDA guidance does not formally define either of these two terms, but says, “source data verification and other activities traditionally performed by on-site monitoring can now often be accomplished remotely, as both trial data and source data typically become part of the central submission.”  The TransCelerate paper equates the two terms, and formally defines both as “monitoring activities as defined either within process documents or in the MP that occur away from the study site location (e.g. at a Monitor’s home or in a sponsor representative’s office.)”  The EMA paper uses neither term.

Remote/Off-site versus Central Monitoring:
The definition of central monitoring in the FDA guidance consists of 8 bullet items, only some of which are the technology-enabled remote/off-site version of traditional on-site monitoring.  The TransCelerate paper says, “Off-site Monitoring Activities are performed by Monitors and can be distinguished from Central Monitoring which could also be performed by Monitors or other roles within clinical operations or by other functions (e.g. Statistics, Data Management, Safety).”

Central Monitoring and Statistical Analysis:
The TransCelerate paper focuses on analyzing Key Risk Indicators to determine and adapt a monitoring strategy.  The appendices are ripe with examples that show how thresholds can be used to dictate actions and trigger on-site visits.  The paper doesn’t preclude analyzing all study data (both critical and non-critical) for outliers, but doesn’t discuss much about the role that this type of statistical analysis might play or provide details on its use.  The definition of “central monitoring” in the FDA guidance is more general.  “Conduct aggregate statistical analyses of study data to identify sites that are outliers relative to others” might refer to Key Risk Indicators.  It might also refer to the “analysis-of-every-data-element” approach that some platform vendors have implemented to detect outlying, hence higher risk, sites.  The EMA paper speaks only of the potential to develop central monitoring systems that use statistical methodology, and leaves it to future papers to delve into more detail.

Adapted versus Adaptive Monitoring
Of the three sources, the EMA reflection paper uses the term “adapted” the most, but never formally defines it.  “Adapted,” as it relates to monitoring, can refer to the practice of tailoring monitoring procedures for each protocol.  (“Safety monitoring procedure adapted to each trial”)  It can also refer to mid-study changes to the monitoring procedures that were set out at the beginning of the study.  (“[Priorities] should be carefully set out so that risk analysis is carried out and control measures are designed in a way that is continuously adapted to them.”)  The TransCelerate paper uses the term “adaptive,” almost as generically, when it defines risk-based monitoring to be an adaptive approach.

Centralized versus Centralised Monitoring:
No factual foundation whatsoever; I went out on a limb.