In a recent post, where I analyzed the relative underfunding of male breast cancer research, the issue of whether incidence and funding should be related came up. So I decided to take a look and discovered, perhaps not surprisingly, that breast cancer research is funded at a higher level than might be expected purely on the basis of its incidence.
For this analysis I decided to focus on research funding from the National Cancer Institute (NCI) as reported on their website. Please note that they do not report funding for all cancer subtypes, so I was restricted in this analysis by what they do report, and included only cancers that had data for all years in the analysis (2007-2011).Then I obtained estimated annual incidence for both sexes in the USA from the American Cancer Society’s annual report for the same cancers.
When you graph incidence vs NCI $ you get a somewhat linear relationship, with some noise in it. Floating quite far above the trend line is breast cancer. If it fell on the line, it would receive about $400M in NCI funding a year, as opposed to the $600 it gets: a $200M bump.
This is notable, particularly when you consider that the other major federal cancer dollars, flowing from the Department of Defense, also contribute above this amount to research on breast cancer. So, why does our federal government spend disproportionally on breast cancer research (amongst other cancers)?
Here is the visualization (click to launch interactive version):
While I don’t know, specifically, of course, I can give you my best assessment. It is that governments don’t act in a vacuum, even when funding research. I believe that effective advocacy has been enormously influential in determining how cancer research dollars are spent.
Grant allocations at NCI are driven by both programmatic considerations and the quality of grants that are submitted. Program offices at NCI craft a portfolio of research grants that is intended to reflect breadth as well as NCI’s priorities. NCI’s priorities are in turn set by the Director, influenced by external advisors and ultimately subject to control by the government. The NCI Director reports to the National Institutes of Health (NIH) Director, and upward to the Secretary of Health and Human Services. Appropriations that fund NIH are controlled by Congress. At the same time, NCI can only fund what is submitted, and can only fund within that pool projects that are found meritorious by peer review. The funding level of any one cancer is therefore both influenced by the decisions made by the officers at NCI acting on behalf of the tax payer through their elected representatives, and by the volume and quality of submissions that come through the door.
Of course the two are connected. The volume of grant submissions on a disease is closely linked to the size of the community of researchers focused on that problem, naturally. The size of the community is in turn correlated with the amount of funding available, both NCI and all the other sources. As I will consider in more depth in a future post about the nature of grant funding for biomedical research, funders and researchers live in symbiosis. A simple example: a grant on breast cancer research allows a scientist to train students and fellows in this area of research, and they are more likely to work on that cancer as they establish their own research program later on. Arguably the size of the community is likely also connected to grant quality – larger communities are likely to have more high level researchers, and to be more competitive.
What about NCI’s priorities? Of course NCI is run, and advised by cancer experts and so as an institution has a professional opinion on priorities. However, NCI is also influenced by the communities it serves – including cancer patients and their advocates. NCI has been including patient advocates in their processes for some time now, a clear indication that they are listening. I would go further and say that the funding levels at NCI are strongly influenced by such advocates. Not only do foundations and patient organizations exert direct (lobbying) and indirect (cultural change) pressure on the government, but by providing additional research funds, they enrich the research environment for that area leading to more scientists being active in working on that disease. It is very common for laboratories to have both federal and foundation grants.
I am more familiar with my own research area – brain tumors. Funding for nervous system tumors is also well above the line in the analysis above, and I know that this is in part due to the fact that the foundations who work in this area effectively collaborated several years ago on influencing the funding levels for this disease group. The most common primary brain tumor, glioblastoma, has a much worse prognosis than breast cancer and so there is not the same strong survivor community around this disease, which is why the effort had to be made largely by the foundations, often led and supported by people who have a personal connection to the disease. They have been remarkably effective at bringing brain tumors significant attention. Of course they have been investing considerable resources raised from the public in research also.
In my recent analysis of funding for breast cancer, the largest amount of funding by far came from the federal government via NIH and the Department of Defense. Third in line was Susan G. Komen for the Cure. In data available in the SciVal Funding database grants totaling about $195M over the years 2009-2012 from Komen supported research. This approaches 10% of what the NCI declares in research as described here (you can explore the contributions from different funding agencies in the visualizations in this post).
Could Komen do more? Yes. Always, yes. Can the split between screening, prevention and research for cures be challenged? Yes. Should we do more research on metastatic breast cancer, which gets only about 11% of the breast cancer dollars? Yes. Same for male breast cancer, which represents 1% of the disease, but gets 0.05% of the research funding. But while we advocate for a change in how the pie is distributed, we may wish to acknowledge that the reason the breast cancer pie is relatively big, is in part due to effective advocacy.
In summary I believe that the strong, effective and long-term advocacy by the many participants in the breast cancer community have been extraordinarily effective at raising awareness in the broadest sense, and thereby raised funding levels. I am not a big fan of pink (hey, I’m a guy…), and certainly not the excesses of pinkification documented so effectively by many of my fellow bloggers and most recently Peggy Orenstein in the NYT. However, I for one readily acknowledge that Komen and the many other foundations in this space have changed the environment, and in many ways for the better, certainly when considering research funding.
[Note on methodology of funding analyses presented here: The difference between the annual level of funding reported by the NCI on their website (round $600M; above) and what can be found in the SciVal Funding database when you search for “breast cancer” (around $1bn) needs some consideration. First off, there is a methodological difference. I realize that a word search in a database is prone to overestimating funding, because any grant that uses the words “breast cancer” in its title or abstract will be captured. This means that grants that focus on several cancers, including breast, could be included. NCI may be more stringent, and uses a field in its database called “cancer type” to identify grants. This allows that a grant can only appear once in the data, and under one category. I assume that the NCI reports any 1$ only once, also in one category (prior reports where budget % and $ correlated 1:1 suggest this is the case). So that is another reason for the discrepancy: for example, a grant on triple negative breast cancer and ovarian cancer might be listed only under one cancer by NCI. Ultimately, it is hard to know exactly how much support goes to a given cancer type in the complex funding landscape. In each analysis we are comparing like with like.]