Dissertation Title:  "From Idea to Product - Translating Knowledge between the Lab and the Clinic"

Translating Inventions into Products: Inventors’ Educational Background And the Speed of Technology Licensing From Academic Medical Centers

The great leaps that have been made in basic life sciences in recent decades have brought to light the need for translating these findings into practical applications. NIH’s support for translational research emphasizes the need for interdisciplinary and multidisciplinary team collaboration as well as training of interdisciplinary researchers, specifically MD/PhDs. In this paper I use invention and patent data from two large Academic Medical Centers with over $700 million in combined research funding to understand whether the educational background of people in the inventing team, along with invention characteristics, influences the hazard of licensing of the inventions. I use licensing as a proxy for translating an invention into a product. I find that, contrary to expectations, having cross-domain expertise on the team – i.e. a combination of MDs and PhDs or MD/PhDs, significantly lowers the hazard of licensing inventions compared to inventions by teams that are made up solely of clinicians or bench researchers. Furthermore, the impact of cross-domain teams on licensing of cross-domain teams varies by technology type.

Buyer Behavior in Technology Markets: Technology Proximity between Firm Portfolio and In-Licensed Patents (with Iain Cockburn)

Markets for technology promise to increase productivity by better allocating innovative capacity across firms. Research on the demand side of these markets, however, has been limited. In this paper, we use a new dataset of patents available for licensing from a large, innovative academic medical center (AMC) to understand the structure of these markets. Our data includes data on all firms that showed interest in these patents by signing a confidentiality agreement and later decided whether to license or not license the focal technology. Strikingly, we find that of the 285 patents we observe, about 30% of patents available for licensing are never even looked at, and of those that are looked at about 25% are not eventually licensed. Firms with a higher number of own patents and older firms are more likely to take a license. A licensed patent is looked at on average 3.24 times, compared to 2.23 times for patents that have been considered but never licensed. Because market safety issues are ameliorated in this market, we hypothesize that the lack of demand is due partly to the necessity for complementary technologies in the licensing firm. We measure technology proximity, as captured by different measures of the overlap of International Patent Classes between the AMC patent and the firm’s own technology. We find that technological proximity is indeed a determinant of the decision to in-license once a patent has been looked at. Firms are more likely to license technologies that are close and complementary to their own. While this is true at the broadest proximity measure level, we also note that at the more granular level, conditional on high-level proximity, greater similarity between the licensee’s patents and the AMC patent makes execution of a license agreement less likely. This implies that “close” fit is good but “very close” fit is detrimental for in-licensing. Additionally, we offer improved measures for technology proximity between patent portfolios.

Does Diversity and Focus in Routine Work Influence Creative Output: Evidence from Cardiac Surgery (with Rob Huckman)

Routine, non R&D, work has rarely been seen as the wellspring of innovative ideas. In this paper we propose that routine work can result in innovation and its design can influence the quantity and quality of innovation. We use the setting of cardiac surgery to show that a surgeon’s clinical focus has an influence on the quantity and quality of his academic article publications, which are our proxy for innovative output. We use a panel dataset of 162 surgeons who perform procedures in academic medical centers in New York State from 1994 to 2004. We find that early-career cardiac surgeons who perform a more-diverse set of cardiac surgical procedures have a higher number of cite-weighted articles. We also find that focal experience doing valve procedures is associated with a higher number of cite-weighted valve articles while related experience in non-CABG procedures is associated with a higher number of cite-weighted CABG articles. We believe our results to be generalizable to other settings with routine work that involve non-identical inputs and require problem solving.