Innovation at the intersection of creativity and automation
Chapter 3 of 3. Need to catch up? Read the previous post in this series about scientific workflows.
In the same way that the guillotine concentrates the senses, the need to improve productivity in the pharmaceutical industry has become a life-or-death imperative. Improving productivity does not mean working harder and faster while doing the same job as before. Improving productivity lies in innovation—in the technology and processes that clever minds bring into existence. Moreover, what this innovation must accomplish is vividly clear. We must reduce the time and cost of drug development; increase the probability of successful experiments; and bring better drugs to the marketplace.
I believe that this current crisis demands we pursue innovations that will give scientists time to think. The insight and intuition we need for big leaps in productivity will come in creative thought, not in the details of executing manual and sometimes mundane day-to-day tasks.
To find the time for creative thought, scientists and project managers must come to grips with several important issues. First, we need to free scientists’ time by assigning tasks to people at the appropriate skill level. Second, we need to understand those tasks so that proper work instructions can be communicated. Third, we need to define and harness the data flow to allow for automation where it is achievable.
Imagine a system in which data are available in the needed format without requiring a scientist to spend 50% of his or her overall project time on dataset creation. Imagine graphical display tools that present smart display options by anticipating results. Put this system in the hands of trained scientists and suddenly they are no longer hunched over their computers crunching numbers. Suddenly, they have time to think, and most importantly, they have time to think about solutions to the questions plaguing drug development projects.
This system is not a pipe dream; in fact, the organizing principle—the idea of workflows—is already available. Workflows are a new idea in pharma R&D. The work of scientists has been viewed as more of an art than a process, so it can be difficult to imagine how one would go about identifying and defining a scientist’s tasks so that at least some of the common and repetitive elements are amenable to automation. An understanding of those repetitive elements and definition of the accompanying metadata provide the underpinnings for automation. That’s how systems can be made to seemingly think with you. And the richer the metadata, the smarter the system (for more about metadata, read Understanding Metadata from the National Information Standards Organization).
Of course, considerable effort will be required to study what needs to be done and how to do it. And we need to develop complete and clear requirements for each task in the execution of a workflow (requirements were discussed in my previous blog entry). But the time and effort expended to develop a workflow will ultimately save even more time and effort. Plus, each run through the workflow allows us to learn even more about the tasks and provides the input for a continuous quality improvement process that can improve the speed, efficiency and effectiveness of the workflow.
So what will we do with all that extra time to think? Here are some suggestions. First, spend as much time as possible with scientists from other specialties. Second, talk to those other scientists about the problems they are facing in their work. Third, ask whether you could do something different that would make someone else’s job easier or your results more informative. This is especially important when your data is going to be used by someone else downstream. Fourth, work to develop a common understanding of the disease process or drug effects of interest and systematically discuss what is known versus unknown. Fifth, consider whether data collected for other purposes or reported in the literature might be relevant and useful for addressing a particular challenge. Finally, use some of the time left over to have some fun.
The imperative to improve the productivity of the pharma R&D effort cannot be ignored. The only choice is in how to respond. For the unlucky, this will mean working harder and faster while doing the same job as before. For the lucky, and hopefully most of us will fall into this category, it will mean taking a long hard look at what we do and figuring out how to do it better and faster with smarter tools and automated workflows.
National Information Standards Organization. Understanding Metadata. Bethesda, Maryland: NISO Press, 2004. https://www.niso.org/publications/press/UnderstandingMetadata.pdf. Accessed May 25, 2011.