Americans are living much longer than they did a century ago. A baby born in 1900 was expected to live only to 47, but by the end of the 20th century life expectancy had increased by 30 years, to 77. Among the many reasons for this increase in longevity and overall improvement in Americans’ health — improved diet, better sanitation, improved working and living conditions — is biomedical research.
“Many people have argued that research-driven biomedical innovations such as new drugs, new devices, and new treatments have, in fact, been responsible for increases in longevity,” says Professor Frank Lichtenberg, who specializes in the study of how research and development drives productivity. The United States devotes a lot of money to developing biomedical innovations: combined, the public and private sectors spend about 140 billion dollars on medical research every year. The Federal Government, primarily through the National Institutes of Health (NIH), does very basic upstream research, while private industry does more applied research to translate it into treatments and commercialize it, such as developing new drugs.
However intuitive it is to link medical research to increased longevity, gathering evidence is not easy. Lichtenberg wanted to make a broad assessment of the impact of biomedical research on mortality — what is the social return on investment? It’s not a straightforward matter of comparing total biomedical research with overall declines in mortality. Instead, Lichtenberg used a common method in social science research called difference-in-differences analysis to look at various types of cancer. (Unlike many diseases, reliable cancer data are available thanks to the widespread existence of cancer registries in the United States that track cancer patients throughout their lives.)
“We know that in the last 20 years or so, cancer mortality rates have declined. There is progress against cancer,” he says. “But it has dropped much more for some types of cancers than for others.” Lichtenberg’s hypothesis was that biomedical research is a cause of the drop in cancer mortality, and that cancers with the greatest declines have benefitted from more research studies about that cancer.
Lichtenberg used the National Library of Medicine’s PubMed database, which tracks details about published medical research, including each paper’s publication date and whether it received financial support from the NIH or the biomedical industry. Of the approximately 1.5 million papers published about cancer since 1975, about one-third cited research funding from public or industry sources. (Some of the other two-thirds that didn’t cite funding are review articles or editorials that did not introduce new research findings.)
Combining the PubMed data from all cancer-related research-funded publications with the data from cancer registries, he compared different types of cancers, the number of research papers for each type of cancer, and reductions in age-adjusted mortality, controlling for the number of people diagnosed with cancer in a given year.
Because it takes years, even decades, for biomedical researchers and developers to turn initial findings into effective treatments, Lichtenberg looked specifically at subsequent drops in the age-adjusted mortality rate — the probability of any one person dying from cancer per year. He found a significant inverse correlation: the more rapidly research about a particular cancer accumulates, the greater the subsequent decline in mortality from that type of cancer. Prostate, stomach, and colon cancers have all had large declines in mortality, roughly corresponding to increases in research publications about those cancers. For other cancers, there were much smaller decreases in mortality (and pancreatic cancer showed a slight increase), corresponding with fewer research papers about those cancers.
So if research doesn’t impact treatment in the immediate future, how long does it take? About 15 years, Lichtenberg finds: today’s cancer mortality rates depend on the number of research publications there were about that type of cancer up to about 15 years ago.
“The 15-year lag means that any substantial reductions in medical research today, such as those that came out of the recent budget sequestration, put future longevity gains at risk,” Lichtenberg warns. “The median time between research and publication about that research is about six years, so today’s NIH budget cuts would result in fewer publications about six years from now. Fifteen years after the decline in published research, we would begin to see smaller reductions of mortality. We would see the consequences of the cuts 21 years from now,” he says. “That’s what the very real long-term consequences of reducing biomedical research are.”
Frank Lichtenberg is the Courtney C. Brown Professor of Business in the Finance and Economics Division and in the Healthcare and Pharmaceutical Management Program at Columbia Business School.
Read the Research
"The Impact of Biomedical Knowledge Accumulation on Mortality: A Bibliometric Analysis of Cancer Data"