While I don't cover big pharma on its own, I am keenly interested in it and follow it with some passion, having in my career worked for several pharma firms as a consultant.

I have spoken to several pharma folks about data since arriving at Forrester. Some of what I learned was surprising. Some of what I learned reinforced my views coming to this job.

Important leaders are saying that good data management gives companies a key and differentiated competitive advantage. We are hearing this in almost all of our conversations with pharma leaders. It is 100% top-of-mind for big pharma CIOs. http://searchcio.techtarget.com/video/JJ-Pharma-CIO-Healthcare-data-management-will-revolutionize-her-business

But what are the trends, and what are the best practices?

We are hearing from all the pharma stakeholders four stories that are driving the questions that are being asked of the data:

  1. Pharma needs to get away from its focus on molecules and pivot to a holistic view of disease. As per a senior IT manager at a major pharma in a meeting with me last week: "We have to deliver whole solutions, and not just pills." 
  2. Pharma needs to understand prescribing behavior in the formulary and in the physician's office better in order to influence it and thus drive sales. As per a senior marketing manager from a meeting recently: "In the old world, we just sprayed and prayed," meaning that the marketing campaigns aimed at the physician did not discriminate as to who that physician was.
  3. Genomic-based drugs are driving changes though the amounts and types of data that the industry must manage.
  4. The recent changes in American patent laws from first to invent to first to file is driving a need to more coherently manage the data associated with creation of intellectual property.

What do these systemic industry changes mean to the questions asked of data?

To address the core problem moving from molecule to disease (#1), we see fragmented efforts to map taxonomies to each other. There is an urgent need to do this, and below are a few examples of taxonomies that must be mapped to one another:

  • Indications (the class of things that drug approvals allow doctors to prescribe medication on when those uses are "on label" and thus can be marketed to)
  • ICD 9 (old diagnostic codes)
  • ICD 10 (new diagnostic codes)
  • ATC (used for the classification of drugs in the pharmaceutical business) 

Because there is no effort underway to standardize any of these maps that I know of, there are no reference guides readily available to do this mapping. Therefore, the best practice is to staff up these efforts with domain experts and to prioritize regions of the taxonomies that are important to the party that is underwriting these efforts. I do see significant opportunity here for the NIH to step in and put together a working group to do this mapping and do it as a standardization effort. Of course this standardization would have to be maintained as most of the data standards listed above are not settled, but rather drift with innovative changes.

To address the need to market in a more targeted way to the physician populations (#2), we see big pharma purchasing data that maps prescriptions to doctors and also doing significant analysis employing methods of social graphing to measure influence circles. We see this work taking place within only the most sophisticated environments, and we see these environments struggling to find appropriate business models to draw from in order to do this work. Probably the place where this work is mature is in the intelligence community, and thus it would be a best practice, not one we have seen, to have the pharma folks who are pioneering this work create potential best practice sharing sessions with the intelligence community. Alternatively, recruiting talent from these communities is something we see in firms in the information business like LexisNexis and Thomson Financial.

To address the problems of harnessing the gene sequence (#3), we see the most advanced work coming not from pharma, but from the clinics. Mount Sinai is a great example of a center of excellence here; in particular, look at the Icahn Institute (http://icahn.mssm.edu/departments-and-institutes/genomics). Thus the best practice of pharma, and we are hearing that some leaders are moving in this direction, is to team up with academic institutions in order to move this work along. Many different types of IP and financial configurations are possible to both underwrite work and protect intellectual property.

To address the patent compression issue (#4), we see efforts of natural language processing coming to the fore. In particular, we see a great effort to create better-structured data from somewhat unstructured data entered into electronic lab notebooks. We are mostly being made aware of this by the vendor community, and thus are not sure to what extent this is a profound issue of data management in big pharma. The data management teams we have spoken to in big pharma are not focused on this problem; i.e., look for this to possibly be a solution in search of a problem.