Automating Serendipity

In the 1990’s, a licensing leader at a major Pharma was reading through dozens of journals each month, looking for insights on possible investments.  He noticed a new study on an old drug revealed it acted on a slightly different pathway than previously assumed.  Similar drugs were found to have adverse effects (AE’s), so the owner never filed for market approval in the USA.   His Discovery organization shared recent new insights on the nature of the disease, leading him to recognize that the other pathway triggered those AE’s.  Realizing the potential, his firm secured the rights to register and market the drug in the United States for five years – and at a great price.  It resulted in the launch of Glucophage, a drug that netted Bristol-Myers Squibb well over eight billion dollars in sales in the US in five years.   Added benefit:  the cost of that product was a small fraction of the supposed $2B it takes to develop NME’s in 2011. 
        … Nice results. 

This serendipity of ideas and understanding came by hard work and chance.  But what other capabilities positioned the firm for this success?  And could that success be reproduced with new and purposefully acquired abilities?  I studied this and similar cases, looking to help clients improve their return on R&D expertise and assets.  The key factors that were “atypical” from standard practice in the 1990’s were: 

  • The scientist in licensing was aware of the broad range of unmet therapeutic needs of society – not just in terms of medical conditions, but also by how society would value a solution 
  • He was applying not just new information on the pathway but was applying a new approach for examination.  Looking at the problem through this “lens” (differential pathways for efficacy and AE’s) was again, unusual for the 1990’s 
  • There was an unusually short lag between sharing new insights from Discovery and areas like Licensing.  Again, atypical in most Pharma’s of the 1990’s where silos were de rigueur 
  • The above three “capabilities” resulted in better insights and understanding of value.  Understanding before competitors, the market or the owner of the rights caught up.  This allowed BMS to act quickly and with confidence to great advantage and profit

Architect’s toolbox:   
1) Combine disparate sources of knowledge to help synthesize new insights
2) Apply new ways of looking, “lenses”,  at these insights 
3) Speed the flow of information and “lenses” across your experts and functions, and
4) Educate and encourage knowledge workers to continuously examine how new ideas,      
     through these capabilities, can create value and competitive advantage

* Some in R&D roles take exception to the last point.  They feel that such distractions hamstring true scientific thinking and efforts. 

Many in 1995 had this mental model (on the right) for what created success in Pharma R&D .  Components:  1) Discipline in “aim”, or goals, plus 2) “Preparedness” – which for some meant investing in research “all around the aimed target”.   Aim was a bit constraining and created some blinders in execution, but the biggest flaw for this model was 3) “Luck”, or the wistful wishes and looking at others enviously.  With luck, one could get results with less aim and preparedness investments.  Too simple?  Yes – but this is to make a point.  Even today, folks describe luck as a major component of drug R&D success.   Francis Collins of the NIH has recently stated:  

“Both the need for and the risks of this strategy are clear in mental health. There have been only two major drug discoveries in the field in the past century; lithium for the treatment of bipolar disorder in 1949 and Thorazine for the treatment of psychosis in 1950.  Both discoveries were utter strokes of luck… “ 

I submit that the old adage “you make your own luck” was, and remains operative.  Taking the principles from above, one can craft a better model.  1) Substitute velocity and breadth of sharing across the organization for preparedness.  Set and follow goals, but more importantly:  2) Set the stage for creating new ideas by converging disparate sources of information (and share them!)  Related to this, educate your scientists to continuously examine where value and advantages lie (converging science, markets, and a bit of finance.)  Finally, 3) Ensure your organization is always evaluating new ways of looking at problems and opportunities.  Test these new “lenses” across all the high velocity, disparate knowledge you are sharing.  This nimbleness of mind also creates nimbleness of organization – something I coach regularly to R&D clients.   
More details on how firms can build in capabilities using this new model:    

Converge and deliver disparate sources of knowledge: 
In the Glucophage example, the converged disciplines were individually selected.  It was not popular in the 1990’s to apply that combination of strategic market needs, plus an appreciation of risk/costs, and the underlying science of the disease pathway (such as we understood it .)   Today, all aspects of Pharma/ BioPharma R&D can benefit from bringing in new and nontraditional ideas.  For instance, the value of your candidate might plummet below justified funding, if a competitor leapfrogs your development program.  Conversely, systems today can indict a class of compounds you are considering. They can sense and call out a potential AE area before your specific candidate advances to the clinic.   

New information technologies can suddenly break through “unsolvable” problems or streamline a ponderous process.  But most of these chiefly make it easier and faster to bring new combinations of information to experts that then can apply more creative focus and time using that expertise.  An example can be seen here.  We found a majority of surveyed researchers “walked away from” what they thought were “ultimately solvable scientific problems”.  They cited excessive costs/ time to integrate, distill or organize in-hand information.   Technology can help but it is NOT a panacea.  Instead, it can enable and potentially create new competitive advantages once the firm has developed a strategy or guiding framework.  They then use this to select and prioritize capabilities to apply at specific areas of the R&D pipeline.  We create a Capability Impact Framework to help firms think this though and then guide initiative(s) which build out  those new talents.  

Finding and applying new ways of looking, new “lenses”:
Scientists are always looking for new ways to interpret and assess their information or problems.  However, those “new ways” are often extensions of standard statistics and visualizations that are part of basic science training.  New “lenses” around how drug molecules interact with receptors, disease models, etc. are rapidly being identified and piloted.   Some use advances in analytical probes (for example, fMRI, data gathering, information processing, while others involve other disciplines (e.g., behavioral science guiding compliance, nanotechnology, in-body sensors …)   There are many examples of novel “lenses” that have guided firms to improved performance and results – I have myself been involved with many.  

One powerful example came from rethinking of what was formerly thought of as a necessary R&D cost: diagnostics.  Turning tradition on its head, we helped firms to think of how R&D tools that helped probe diseases might be of value to patients and doctors.  Value that was comparable and almost inseparable from the therapy itself.  Ten years later, many firms have “diversified” into diagnostic offerings, utilizing biomarkers and selling new tools that also improve R&D and to open up new avenues of research.  Some firms stated flatly that they will not develop a NME without offering a corresponding diagnostic.  Caling this diversification is a misnomer since such offerings are integral to the drug franchise. That thought exercise, or “lens” of looking at elements of cost, data and expertise, as salable offerings has transformed the industry.  What will be the next new perspectives to help improve returns on R&D and to benefit society?  

Summary:  Note that firms uncover new “lenses” continually (but are only slowly adopting.)   These include leveraging in-silico models, proteomics, semantics, biomarkers and more in 2011.  The key here is to develop a process to identify, vet and assess new lenses that all R&D staff might explore in an organized fashion (although others are experimenting with using the equivalent of internal crowdsourcing.)  Caution:  Once a novel tool which creates advantage is identified, it can be a challenge to recognize and protect it as valuable intellectual property.    

Increase the velocity of information moving across disciplines and groups: 

In 2000, dozens of key info nodes in R&D were isolated, requiring manual conversion & curation

BMS benefited from having the corporate HQ and the primary research center at the same location.  There, professionals from many disciplines often met and shared lunch tables or intramural sports.   While this seems quaint now, remember that in decades preceding 2000, a core problem for Pharma’s were organizational “silos” – even within R&D.  Again, without the Licensing leader hearing quickly about the new bio-pathway insight, he might not have taken the lead and created the windfall.  Connecting R&D needs and information via lunch needed to be “automated” somehow to drive the convergence and “lensing” efficiently.  Firms have since enacted many programs to connect or integrate groups, disciplines and information systems.   Advances in information technologies have also helped this (see this link on progress connecting data across Pharma R&D.)  

 This remains a challenge – while translational medicine is making advances, other knowledge domains still are not taking part effectively.  Recently, firms have also been experimenting with social networking tools, trying to create “virtual lunch tables” that address the human component, without geographical bounds. 

Summary:  Rapid flows of information across traditional boundaries are essential for increasing chances for “serendipity”.  Technology is but one aspect of this.  One first needs to examine what capabilities are in hand, and what can be gained by nontraditional blending of information in a usable form.  In essence find out what increased capabilities are possible and within reasonable reach.  Sometimes a slight shift in focus is worth more than millions in underused hardware and software.  

Educate all R&D professionals of the basics of what creates value (a subset of #1 above): 

The Glucophage insight could have been made by the owner of marketing rights, or the scientists that studied the disease pathways.  But they lacked this key need for serendipity.  These insights and the value to the firm and patients only came together because  the Licensing leader also had an understanding of what was needed in the marketplace and how economically important the insight was.   Otherwise, it might have remained as a journal article or intellectual exercise, vs. benefiting millions through a renewed class of drugs.  

In 2002, while guiding a top-three firm on redesigning their overall nonclinical R&D process, I was challenged by the team of scientists leading the initiative.  They asked “How can we understand the economic implications of our decisions – years ahead of launch – without diverting attention from our science?  Do we all have to get an MBA?”  While there is no easy answer to this, it was a revelation to hear such questions from scientific peers for the first time in decades.  A sea-change was afoot. 

Firms have since worked on many ways to educate R&D professionals on what creates value.  Some argue that in smaller firms like startup biotech’s whose scientists are already in continual dialog with other disciplines because they are all co-located.  Their scientists by need “ wear several hats” and get their MBA with on-the-job  training.  Elsewhere, others have set up “mini-MBA” programs for scientists that on the order of weeks instead of years.  The key remains that scientists can and need to learn value principles, be motivated  by, and to respect the power of the market (and how society will respond to their efforts.)  As noted earlier, social networking platforms (within the firm) can also cross-pollinate some of that education.  Continual dialog in these platforms can provide context and inject disruptive ideas to advantage, all approaching the realtime effectiveness of “having lunch”. 

Summary:  Firms have to ensure all of their scientists are aware of the business they are in and to some extent, how it works.  As firms look for better returns on R&D before staying in that game, they must educate all staff on what drives business decisions – and that R&D decisions are themselves business decisions with consequences.  If firms do not link what their R&D to what society values, they cannot sustain that investment, and perhaps cannot survive.  There are many programs that can achieve this, but the key first steps are 1) Create a shared understanding of purpose,  2) Eliminate science’s “demonization” of business, and 3) Include creating value as a part of their creative challenge. 

To recap, Pharma’s of the 1990’s pointed to a “luck” factor in others R&D returns and success.  But analysis suggests the luckiest firms were working to create their own luck.  How they did that reveals certain principles that some have adopted.  A few of these worked to “automate” those capabilities to ensure greater “chances of serendipity”.  This can benefit any stage of the R&D value chain, and not just in lead candidates.  It can also create efficiencies, competitive advantages, and  just solve really hard problems.   The value of a “lucky” combination of these principles and capabilities in the 1990’s created $8B in value for BMS.  What firms today can ignore the need to create that kind of “luck”?   



One Response to Automating Serendipity

  1. Watson has a paying client. One of the more exciting directions for healthcare and translational medicine takes a big step towards your MD’s doorstep. But also towards the above capabilities I describe as needed in the R&D Architect’s toolbox. Especially “Combine disparate sources of knowledge to help synthesize new insights” and “apply new ways of looking, “lenses”, at these insights” See

Leave a Reply

Your email address will not be published. Required fields are marked *