Essays on Biomedical Innovation

Abstract

The choice of what scientists decide to study is fundamental to determining what innovations are produced. Yet, rather than being solely determined by a scientist's idiosyncratic preferences, this choice is shaped in important ways by the incentives, organization environment, and supporting institutions that surround a researcher. This dissertation consists of three essays that use biomedicine as a setting to explore how institutions and organizational environment can shape the rate and direction of innovation.

In the first essay, I examine how knowledge certification by professional medical society clinical practice guidelines shapes the use of extant knowledge. Employing a difference-in-differences approach, I find that after inclusion in a guideline affected subfields of knowledge grow in size and scientific impact compared to controls carefully matched on observables. Rather than the aperture of subsequent innovation narrowing, subfields shift towards exploration as they become more translational, more intellectually distant, more disruptive, and build on more diverse and less established prior research.

In the second essay, I investigate how exposure to frontier research in a training program can alter the career trajectories of potential innovators. I study the careers and innovative output of physicians who applied to the Associate Training Programs of the National Institutes of Health (NIH) during the turbulent period surrounding the Vietnam War. I find program participants entered research-focused positions at higher rates and garnered more publications, citations and grant funding than synthetic controls. In particular, the direction of their research efforts was durably imprinted with a distinct "translational" style of biomedical research that was characteristic of the NIH at the time.

In the third essay, I study how a specific institution, grant peer review, shapes scientific risk taking. While risk is an inherent aspect of innovation, those projects with high degrees of risk may be more likely to lead to breakthrough innovation yet may face challenges in winning the support necessary to be carried out. I analyze R01-equivalent grants from the NIH, and after carefully controlling for investigator, grant and institution characteristics, find that grants with high levels of risk taking are renewed at lower rates.