The EDC, CTMS, CDM Confusion

In the August 14, 2006 issue of the eCliniqua newsletter, Mark Uhling willingly accepts a verbal beating by Jack Lawler of Cephalon. Mark's transgression seems to be the assumption that EDC can take the place of CTMS. Jack argues that there is no commerical EDC system out there that can provide the type of services that a typical CTMS system provides. I agree.

Note: I can't provide you with a link to eCliniqua since it is an email newletter. To subscribe, just go to the Bio-IT World web site and find the eCliniqua link.

My thinking is that a lot of people are confused by the features delivered by any of these systems (EDC, CDM, CTM, CDR, etc.) much less understand how they need to work together. A key problem is that we are in transition mode between the traditional mode of clinical trial conduct and a desire to rework processes and leverage newer technologies.

In the traditional world we write a protocol, set up paper Case Report Forms (CRF), collect the data on the forms, enter the data into a Clinical Data Management (CDM) system, clean it there and then throw it over the wall to the statisticians.

In the new world we create a protocol (perhaps using an automated protocol publishing system), build electronic CRFs in an EDC system while simultaneously building the EDC database structure, collect and near-simultaneously clean the data electronically, pipe the data from the EDC system to the CDM and then throw it over the wall to the statisticians.

So where does the CTMS part come in? Well, EDC vendors are happy to tell you that much of the information you need to report on trial status is automatically collected or derived from their software. This is where Jack Lawler only partially agrees. Why? Because the EDC systems don't allow you to go across trials and especially not to those still being done with paper. Also, none of those systems let you pay investigators based on milestones, or anyting else for that matter. So, the need for a CTMS component is clear.

Now, in my view, the new world is still not optimized. What is needed is a paradigm shift in the way technology is used. But what should this look like? Well, to paraphrase Shakespeare, "first, let's kill all the CDMs." With apologies to both Oracle and PhaseForward, CDMs were designed to deal with paper systems and heavy data entry volumes. Moving EDC data to a CDM system just makes no sense any more. Before you start ranting, however, let me be clear. Even the EDC systems will need to evolve further before the CDM lights can be turned off. But, we are almost there.

So, instead of the CDM, we need a final resting place for all the data. By all, I mean every bit of data from all trials and all systems used to collect the data. Not just EDC but also lab systems, genomics databases, AE systems, etc. This new resting place is called a Clinical Data Repository (CDR). CDR is a topic that I have discussed in other posts (Here and Here).

This CDR is a magical place since it avoids the need to throw data over the wall to the statisticians. In fact, people from all walks of life (roles) go to the CDR and use it as required. All the data, for example, can reside in the CDR which may, in fact, be an Oracle database. Then, statisticians and programmers can go there and create any type of output that may be needed. Yes, this includes SAS datasets and trasport files but also export files for creating data warehouses or marts for further analysis.

OK, so how does CTMS come into the picture? Even here, it may not be the traditional CTMS that makes the most sense. We need to think of services rather than applications to gain the flexibility needed in a data centric world. So, the CTMS functions can be delivered through an application layer that hovers above all of the data and repositories.

A Portal or Dashboard, for example, would tap into the underlying systems and repositories to aggregate the data needed for trial management. Yes, it will also need to have its own functions such as investigator payments.

From an architectural perspective, the optimal environment for trial conduct will be one that collects as much data in as passive a manner as possible, exchanges those data seamlessly between applications and services and requires only minimal data entry from anyone.

And no, we're not there yet.

Comments

[steven davies]

You are right that these sales letter delivered by any of these systems EDC, CDM, CTM, CDR and etc. need some optimization.