Scaling Research Solutions for Society’s Real Problems
To transform US research labs to better serve society, we need to bring in a new type of scientist who specializes in scientific operations.
The COVID-19 pandemic demonstrated that the US scientific community has the ability to surge to a higher level of productivity very quickly. Although the NIH–Moderna collaboration that jump-started the development of a safe and effective COVID-19 vaccine is the most visible example of this capacity, other institutions and collaborations accomplished similar leaps under the pressure of the pandemic. I had the great fortune of witnessing close-up a diagnostics effort organized by the Broad Institute of MIT and Harvard, that rapidly converted our genomics platform for COVID-19 testing, ultimately reaching a processing capacity of nearly 150,000 samples a day and testing over 20 million samples to date.
If such impressive accomplishments are possible, shouldn’t our scientific institutions be capable of doing much more under ordinary conditions? With society’s outstanding need for breakthroughs in science and technology for problems as far-flung as climate change, health disparities, and pandemic preparedness, doesn’t science have an obligation to do more?
The scientific system’s successes over the past 18 months have also spotlighted its failings: the fruits of US science and technology have benefited some groups disproportionately relative to others, with race, geography, and socioeconomic status being key determinants. To ensure that our scientific research benefits everyone in our country, we must get more innovation out of our labs and into our communities. This change will require not only producing solutions, but then rapidly identifying, scaling, and distributing them to the entire population.
One way to accomplish this kind of rapid scaling is to rethink the organizational structures of our research centers, specifically by staffing them with a new kind of scientist: a professional manager or operations specialist who can amplify the impact of research beyond disciplinary boundaries, thus expanding the scope of the organization. To transform the US research enterprise to better serve society, we should consider recruiting and incorporating these new types of scientists in research teams and more widely adopting their unique approach.
At the Broad Institute, our COVID-19 diagnostics effort built on existing organizational divisions that had already invested in industrial-scale laboratory facilities, software engineering teams building production systems for handling health data, and a diversity of industry partnerships. These innovative operations and teams—unusual for a research laboratory—are overseen by full-time professional scientific staff. With titles ranging from scientific adviser to alliance manager to director of scientific partnerships, at the Broad these organizational leaders—who often have extensive scientific training—work to magnify the impact of the traditional principal investigator (PI)-driven laboratory.
The initial challenge for COVID-19 testing was to reconfigure a licensed clinical laboratory that had been designed for DNA sequencing to handle viral diagnostics. Early during the pandemic, our equipment was adapted to detect SARS-CoV-2 at the rate of a few hundred tests per day. By taking advantage of a highly modular approach to automation—driven by the operational expertise of the organizational leaders—within 6 months capacity was scaled to more than 100,000 COVID tests per day. And as the pandemic has continued, our professional scientific staff has played crucial roles in connecting with industry partners for logistic support to expand testing across New England. We have also launched new capabilities such as pooled testing and viral whole-genome sequencing for tracking the emergence and spread of SARS-CoV-2 variants.
For simplicity, we can call these organizational professionals scientific product managers (SPMs). This title has echoes of the role played by product managers in the software industry, where software teams are often co-led by a technical lead (i.e., a senior software engineer) who works alongside a product manager. The product manager is an organizational generalist, often a former engineer who also earned an MBA, whose responsibility is to deeply understand the product ecosystem, customer needs, and technical capabilities. They may also be involved in communicating with key leadership in finance, business development, sales, and operations to attend to the process of launching successful products. The product manager works closely with the technical team to set priorities, create product roadmaps, and establish timelines for implementation. As such, they are considered key organizational leaders in the tech industry and are highly sought after as CEO candidates.
In the context of a research organization such as the Broad, our PIs are analogous to technical leads, while SPMs are more like product managers. Of course, the “product” in this case isn’t a singular entity such as Gmail, Spotify, or MATLAB, but rather represents a scientific research agenda. And in the broader context of a collaborative ecosystem of partners in tech, pharma, venture capital, and government, SPMs can play a vital role in creating a hybrid culture that blends elements from these disparate communities.
Employing SPMs makes many initiatives at the Broad resemble a fusion of industrial operations and traditional research laboratories. A good example is Machine Learning for Health (ML4H), where I lead strategy and operations. ML4H is a hybrid between the type of software engineering team you’d find at Google, Apple, or Amazon, and the type of clinico-genomic research group found at academic medical centers. For this project our SPMs work on prospecting and vetting potential partnerships; identifying new software features that would benefit the research community; working closely with clinicians and our technical team to plan and execute new feature roadmaps; liaising with counterparts in industry to facilitate large, multiyear, interdisciplinary collaborations. With SPMs on board, ML4H can invest in the production of tools, resources, and policies in addition to scientific publications—all while incubating and launching cross-institutional research.
Unlike traditional PIs who typically think deeply and intensely about a relatively small number of focused problems, our SPMs are flooded with information—much of it organizational and operational. They then use their extensive research training to shape decisions large and small, wearing the diverse hats of scientific leader, thought partner, tactician, and diplomat.
We have found that the best SPMs are generalists rather than specialists. General scientific knowledge, critical thinking ability, and strong communication skills are major assets that our SPMs use to overcome bottlenecks in the research process. Their roles take them from the macro to the micro, from strategy to execution, from vision to minutiae. As generalists, they can collaborate on research while also being embedded in finance, development, and communication offices to understand critical dimensions of science funding, budgets, and planning—all things that scientists are typically shielded from—but which often determine the success of an endeavor.
When we hire SPMs we look for highly trained scientists who have left academia for careers in finance, consulting, and advertising analytics as well as those who have done stints in industrial laboratories or in the wider academic ecosystem of science publishing or policy. Many of the skills that these scientists have gained during their diverse experiences are particularly valuable in scaling up academic science. By bringing these people back into university-based or university-adjacent research environments in operational roles, we can organically expand the scope of impact that our scientific organizations are capable of.
Scaling the model of SPMs to more labs would not be difficult—there are many highly trained PhDs and other skilled professionals available, and there are also good working models for training and integration. At Broad, the organization invests considerable resources into every worker. While some SPMs are hired into highly structured roles, others are given considerable latitude in shaping their own role and might be given the better portion of their first year to explore how the institution works. As participant observers, the trainees network within the Broad, speaking one-on-one with dozens of PIs, graduate students, postdoctoral researchers, staff scientists, administrative assistants, and other key colleagues, while writing white papers on internal strategy. They may speak with lawyers from tech transfer offices, local entrepreneurs, and venture capitalists to understand the potential for science translation, identifying key bottlenecks for their local ecosystem. SPMs rely on mentors to find their path and settle into a stable role. This model, or others, could be adapted by institutions looking to incorporate SPMs into their work.
Although the Broad had the benefit of being launched in the wake of the Human Genome Project and was created with substantial philanthropic support, other scientific ecosystems could take an alternative path to developing such nontraditional expertise. Universities, for example, could pool resources across multiple institutions and commit to collaborating with industry to identify and develop instrumentation, software platforms, or other standardized tools and processes that would otherwise not be incentivized. This approach would require that universities embrace not only the SPM model but also work across the research and development cycles at the highest levels. Finally, universities must make a significantly deeper investment into local scientific communication, education, and community building to ensure that this new institutional capacity is directed at problems that matter to society.
I believe that fully integrating a cohort of full-time professional staff into laboratories and research institutions could transform the current scientific landscape. Augmenting existing research institutions with additional operational capacity should allow traditional PI-driven laboratories to continue to be the productive intellectual engine they have been since the Second World War, but with greater impact beyond the institution’s walls. And shifting the culture of our laboratories could encourage new models of science to take root, so that scientists are judged by their societal impact rather than mainly by publications. As this model becomes more the norm, we can expect to see faster innovation and a more rapid translational pipeline from new scientific insights to products, services, and policies that benefit society.
Although SPMs can transform the work of individual laboratories, we should aim higher, and use SPMs to incubate the development of collaborative ecosystems in US cities in an organic, grassroots way. Over time SPMs could profoundly change not only the capacities of our research facilities but also the way these labs conceptualize their mission. As SPMs connect and network within their home institution and into the local scientific ecosystem to create collaborations, they could also work with finance and development offices to create new philanthropic strategies and propose corporate partnerships. Such transformative scientific ventures could go well beyond the traditional understanding of scientific research, allowing each city and each scientific ecosystem to find its own voice and take advantage of whatever resources and institutional capacities are available.
The impact of SPMs should not be exclusively technical in nature. In medicine and public health, for example, we might imagine full-time staff devoted to science communication and public outreach, working closely with community leaders at religious organizations, food banks, and homeless shelters to ensure that local needs are consistently being represented to university researchers and other thought leaders.
Conversely, these lines of communication can ensure that advances originating in universities are being translated into local action. And by establishing strong relationships well in advance of public crises, we can ensure that our research institutions can anticipate and more rapidly work to organically and equitably distribute the fruits of our research enterprise when called upon to do so.
The success of US science has been built on decades of highly evolved decision making and a rich ecosystem of research institutions. Rising to the challenge of the twenty-first century will require that we continue this introspective tradition. As we confront unprecedented challenges as a society, we need to take a broader, more integrative view of research, and also incorporate serving the needs of our country as a fundamental metric by which we evaluate scientific success. To accomplish this expanded approach, we need to find ways to scale our work to reach more people more quickly.
Today, when we think about the words “scale” or “scalable,” we think of the tech giants or unicorn startups that capture larger and larger numbers of users for digital platforms at a dizzying pace. We experience the weight of Facebook’s awesome power over the news media, elections, and the global ramifications of its algorithmic tweaks and product decisions. We feel the visceral impact of Amazon’s relentless march into one vertical after another: from books to groceries to cinema to the infrastructure of the internet itself. What we haven’t tried is scaling the impact of our research environments by creating cross-institutional science and technology ecosystems so that a similar immediate impact could be felt for the country’s most pressing problems. SPMs offer a way to seed the development of such environments by connecting our cutting-edge research with our local communities and applying our nation’s vast investment in science and technology more directly to making life better for more Americans.