Do Firms Underinvest in Long-Term Research? Evidence from Cancer Clinic Trials

IP Speaker Series presenter Williams examines the existence of private investment distortion and related potential effects on patient life outcomes.

In the last five years alone, eight new drugs were developed and approved to treat metastatic lung cancer, all targeted at patients with the most advanced stage. On the other hand, the FDA has never approved a drug to prevent lung cancer, the leading cause of cancer deaths in the US. In fact, only six drugs have ever been approved to prevent any type of cancer at all. While scientific challenges or market demand could account for this, what if the reason can be chalked up to private firms’ investment choices—namely their interest in drugs that most quickly can be brought to market?

This was the hypothesis posed by Heidi Williams, Class of 1957 Career Development Assistant Professor in MIT’s Department of Economics, who recently came to campus to discuss her paper, “Do Firms Underinvest in Long-Term Research? Evidence from Cancer Clinic Trials,” as part of BU Law’s Intellectual Property Speaker Series.

The paper—co-authored by Eric Budish, associate professor of economics at the University of Chicago Booth School of Business, and Benjamin N. Roin, assistant professor of technological innovation, entrepreneurship, and strategic management at MIT—examines whether private research investments are distorted away from long-term projects in favor of drugs that can be commercialized the quickest. By examining data from cancer clinic trials over the last three decades, Williams, Budish, and Roin present evidence that suggests there is such a distortion in investment, and that patient life outcomes may be at stake as because of it.

Using the context of cancer clinical trials

Williams noted that cancer clinical trials offer an interesting and relatively effective method for considering this question for several reasons. For one, 100% of reported cancer cases are documented in the US, and treatments are organized around the effected organ and stage of the cancer, offering a natural categorization of observed and potential research and development activity. Secondly, estimated survival rates linked to diagnoses and stage of cancer offer a benchmark for determining the probable length of the clinical trial period.

Williams explained: approval for a cancer drug, in almost all cases, is dependent on its proven efficacy in terms of survival rate. In other words, clinical trials must show, with statistical significance, that the drug can extend life expectancy for a targeted kind of cancer. Therefore, results are only as quickly realized as the rates at which patients die. Firms that invest in drugs to treat late-stage cancers, where life expectancy is relatively low—say, a year—have a much shorter clinical trial period than drugs that aim to treat early-stage cancers, where patients may live for 10 years or more, or to prevent cancer all together.

The hypothesis

With this reality in mind, Williams explained the authors' hypothesis: the shorter the the clinical trial period based on patient survival rates—or the shorter the lag time between invention and commercialization—the more incentive a firm has to invest in that product because they will see a return on their investment much sooner.

The obvious financial benefits of a shorter time between invention and profitability offers one explanation for investment distortion toward late-stage cancers. But Williams acknowledged that the current structure of the patent system might also play a role.

Since patents are awarded to innovators in fixed periods of market exclusivity (e.g., 20 years in the US), and firms file their patents at the discovery (invention) stage rather than the time of the first sale, firms investing in longer-term projects are losing a significant portion of their market exclusivity and protection of their invention while the drugs are in lengthy clinical trials.

The data

Williams then moved into analyzing the paper’s data, which correlates to the proposed theory that firms lack incentive to invest in long-term projects. Firstly, the authors found that patients with localized cancers—who have a 70% chance of surviving five years—had about half as many opportunities to enroll in clinical trials as patients with metastatic cancers—who have about a 10% chance of surviving five years.

Proposing that patients with longer life expectancies could just be less inclined to participate in clinical trials, not wanting to undergo the potential pain or stress, Williams pointed to analysis of data from hematologic cancer trials. Hematologic cancers are unique in that accelerated approval for drugs is possible: designated surrogate endpoints—measurements that are proven strong indicators of future improved life expectancy—not actual survival rates, can be used in the approval process.

The authors found that there was not, in fact, the same distortion in R&D investment for drugs treating diagnoses with lengthier survival times for hematologic cancers. They saw much more private investment in drugs for patients with longer survival rates when surrogate endpoints were used in clinical trials, and there was no significant drop-off in patient participation as their life expectancy increased. Williams suggested that this conclusion offers evidence that should firms be allowed to use surrogate endpoints for other types of cancers—and thus have shorter clinical trial periods, ultimately resulting in a shorter commercialization lag time—we likely would see more investment in drugs treating cancers with longer survival rates.

The authors also observed publicly financed trials vs. privately financed to compare investment distortion based on clinical trial length. Williams noted that though both public and private investment dropped off with longer commercialization lags, publicly funded R&D did so at a statistically significantly lower rate.

Investment distortion's impact on survival rates

But has this distortion in investment actually impacted patient survival rates? Could the medicine have improved enough and could patients have lived longer were more investment made in longer-term projects? The authors were curious whether firms’ abilities to leverage surrogate endpoints might have actual social value (beyond quicker profitability for the firms). For this exercise, Williams pointed to what she referred to as “back-of-the-envelope calculations,” based again on data collected from hematologic trials.

The authors compared gains in survival rates for hematologic versus non-hematologic cancers between 1973 – 2003—essentially, where science has helped patients live longer over the last three decades. While survival improvements were similar across both categories in cases where surrogate endpoints did not shorten commercialization lags (i.e., for cancers with low survival rates in 1973), improvements in survival rates for hematologic cancer trials using surrogate endpoints significantly outpaced those for non-hematologic cancers not using surrogate endpoints. In fact, as commercialization lag increased, so did the difference between improvements in survival rates for hematologic cancers compared to non-hematologic. The paper estimates that for US cancer patients with non-hematologic cancers diagnosed in 2003, lack of investment caused by longer commercialization lags accounted for approximately 890,000 lost life-years.

Williams briefly touched on some of the paper’s policy take-aways, including allowing firms to use surrogate endpoints in clinical trials, increasing public funding for R&D projects with longer commercialization lags, and changing patent design to start the patent clock at commercialization rather than invention.

The IP Speaker Series

BU Law’s IP Speaker Series, organized by Professor Mike Meurer, offers students and faculty the opportunity to interact with leading thinkers from around the world in an interactive workshop setting. Williams’ presentation gave members of the IP and health law programs in the BU Law community the chance to engage with this exciting subject, and the resulting active discussion and Q&A were testaments to the great interest students and faculty took in Williams and her co-authors’ work.


heidi williams

Heidi Williams, Class of 1957 Career Development Assistant Professor in MIT’s Department of Economics, recently came to campus to discuss her paper, “Do Firms Underinvest in Long-Term Research? Evidence from Cancer Clinic Trials,” as part of BU Law’s IP Speaker Series.

Related links


Sasha George (’16) on her SiP with the DOJ's Environment and Natural Resources Division >> Full story

Shari Redstone ('78) and Beth Roberts ('83) named to Women in Entertainment Power 100 >> Full story

BU Law sat down with four alumni to discuss their best moments working in public interest law >> Full story

All news

Reported by Boston University School of Law

Last edited March 6, 2015

BU | School of Law | Prospective & Admitted Students | Our Faculty | Alumni | BU Law Central    Social Media Icons

facebook-icon twiter-icon linkedin-icon youtube-icon google plus icon rss-icon blog-icon instagram