Collaboration and Consensus for Safety Data

By Ernie Bush

February 1, 2010 | The Bush Doctrine | In my previous column (see, “Data Hide and Seek vs. Safety Assessment,” Bio•IT World, Nov 2010), I dis­cussed an upcoming conference devoted to the issues around managing the enormous volumes of pre-clinical safety data within the pharma industry. Subsequently senior leaders of toxicology and IT from six large phar­ma companies convened to investigate whether a collaborative project could effectively tackle the problems as described in that column.

It was very satisfying to note that my fall optimism appears justified and that the industry perspective on these issues has evolved to a point where the defini­tion of a collaborative project could be reached expeditiously and with a high de­gree of consensus and conformity. Below are a few high points and open questions from that conference.

Points of Agreement
In previous attempts (stretching back many years), we have not been able to convince a group of pharma companies to address pre-clinical safety data manage­ment by working collaboratively. In what must surely be a sign of changed industry perspective, a collaborative approach was quickly and enthusiastically endorsed as the model for this project. I suspect this reflects both a growing comfort with the concept of pre-competitive collaborative projects and the simple fiscal realities facing large pharma and therefore limit­ing funding for IT projects.

The pharma representatives were also agreed that toxicolo­gists want access to ALL the study data. Whenever we have dis­cussed safety data management previously, one sticking point has been the ‘granularity’ of the data access. In general there are two prominent perspectives: some investigators only want ac­cess to results, e.g. significant changes of group means between treated and control animals; while others need access to all the data down to individual values for each measured parameter on each individual animal. This latter perspective is most often held by card-carrying toxicologists. The scale of these two al­ternatives can be estimated by looking at how many pages in the final study report these sets of data represent. Whereas the significant results may take between 5 and 10 pages in a report, the individual animal listings can occupy 500-1000 pages—two orders of magnitude more volume and complexity. Therefore, even though the meeting participants knew they were endors­ing a much more difficult objective, they were nonetheless unequivocal in their position that what was required for utility in their departments was the ability to retrieve and review data down to the individual animal level.

By design, the meeting agenda allocated a lot of discussion time to defining the high value ways companies would use a system to access their current and archived safety data (i.e. use cases in software-development-speak). When confronted with sophisticated options such as, How do we share data across or between companies? or, How do we data mine preclinical safety data generate new knowledge? time and time again the simple questions were given the highest priority. For example, nearly all of the pharma representatives most wanted to find answers to questions such as, Have we seen this before? and, How does this compare to historical controls? 

Areas for Further Development
The participants did not reach a consen­sus on every point, however. Although they eventually agreed that it made logi­cal sense that CROs should participate in a project to manage preclinical safety data, (largely because CROs are increas­ingly the source of this data), there was a general air of uneasiness with this ar­rangement. So while there is a growing sense of the value of collaborations with­in the industry, this has not yet reached the level of comfort for collaborations between the pharma industry and their suppliers.

Surprisingly to me, the meeting par­ticipants struggled to make a strong economic case for institut­ing a system to gain access to their troves of preclinical safety data. I think everyone believes that having improved access to their safety data stores will improve decision making but no one could enunciate a way to quantify this to any significant de­gree. Can one put a dollar value on having effective and efficient access to their preclinical safety history? What business case would you propose? I look forward to hearing from you.

Ernie Bush is VP and scientific director of Cambridge Health Associates. He can be reached at:  ebush@chacorporate.com.