Friday, January 17, 2014

Cancer vs malaria: burden, treatment spending, R&D spending, R&D results

Summary: I compare the burden, treatment spending, and R&D spending for cancer and malaria worldwide. Cancer causes somewhat more than twice the DALY burden of malaria, but has almost 14 times the global R&D budget per DALY, and almost 60 times the global treatment budget per DALY. Funding for malaria, which is controlled by donors, has a much higher share dedicated to R&D than cancer spending. That R&D also seems to produce more results, indicating diminishing returns at work in medical R&D.

GiveWell recently posted a rough sketch cost-effectiveness analysis of cancer research. I was a bit surprised by the choice of cancer, rather than one of the diseases of the global poor (its focus in recommendations so far), which reminded me to drag out some old notes comparing the global funding pictures for cancer and malaria.

Cancer vs malaria, in disease burden, treatment spending, and R&D spending

Examining the Global Burden of Disease website's visualization tools, in 2010 cancer was estimated to account 7.6% of global DALYs, with malaria accounting for 3.33%.  However, treatment and R&D spending levels were much more disparate.

Combining a report on global cancer burden and spending with a report on malaria R&D and one on malaria treatment and prevention I get the values in the following table:


CancerMalariaRatio
Percentage of 2010 global DALYs7.6%3.33%2.28
2009 R&D spend$19,000,000,000$612,000,00031.05
2009 treatment budget$217,000,000,000$1,600,000,000135.63

The GDP costs of cancer are also many times larger, because malaria is largely confined to poor countries, while cancer strikes both rich and poor countries

Treatment vs research: do rich countries have too much treatment relative to R&D?
Interestingly, R&D expenditures are larger relative to treatment for malaria (38%) than cancer (9%). This makes sense in light of expensive cancer care for individuals in rich countries, via both private markets and public health systems. Individuals have a private incentive to pay for their own treatment, but only much weaker motivations to donate to the R&D. Voters, too, may be more interested in deploying treatment immediately, than funding global public goods in research with delayed benefits.

In contrast, donors pay for most malaria treatment and R&D, and are more free to pick the mix of research and treatment that they think will most help potential victims of malaria. This is particularly true in light of the large role of the Gates Foundation, an integrated entity with an eye to cost-effectiveness in its public health endeavours. The fact that the Gates Foundation's malaria efforts are so heavily focused on research should give pause to those donating to distribute malaria nets, asking about the difference in focus.

It seems very plausible to me that on moderate timescales rich country citizens could benefit greatly if their health systems reallocated effort from treatment to research in similar fashion (and randomized and recorded a larger portion of treatment, so that typical patients contributed more to research).

Malaria R&D has been delivering better results than cancer, suggesting diminishing returns
In recent years malaria research has had a number of striking successes. The Gates Foundation funded a synthetic biology process to produce the antimalarial drug artemisinin more cheaply. A number of vaccines are in development and some, especially RTS,S, have promising results in trials.

In contrast, the much larger budgets of cancer research have only modestly at best affected survival times for cancer, with a few bright spots such as Gleevec and the HPV vaccine. This suggests that there were more "low-hanging fruit" to pluck in the malaria field so that a little funding could go a long way. Diminishing returns may also contribute to the lower R&D relative to treatment for cancer. [Eliezer Yudkowsky draws attention to the possibility that the Gates Foundation is much better than government science agencies, which are separate from aid agencies, in identifying or prioritizing projects with good impacts on malaria.]

4 comments:

Doug S. said...

The other problem with cancer survival rates is that people with cancer tend to be old. The five-year survival rate for "being an 80 year old male" is only 67%...

Carl said...

Doug, the Global Burden of Disease figures account for that. It's disability-adjusted life-years, not deaths.

Igor Terzic said...

While DALY is a valuable metric to decide how to allocate the resources for treatment, it seems a bit suspect when used to decide how to allocate resources for research.

Notwithstanding the difficulty of curing cancer vs malaria those diseases have a crucial difference: if otherwise healthy everyone will eventually get cancer but not everyone will get malaria.

The question we should be asking when dealing with how to allocate research is not how much DALY is currently caused by cancer/malaria, but how many healthy years do we gain if we solve cancer/malaria. The gain from malaria is bounded by cancer while the gain from cancer is not bound by malaria.

Also, AFAIK DALY is calibrated to our current life expectancy(LE) in developed countries which suffer from cancer but not from malaria. If we solve malaria, reference LE does not go up. If we solve cancer, LE does go up. A bulk of difference in DALY between malaria and cancer is that malaria strikes younger victims. As LE goes to infinity, DALY ratio between malaria/cancer goes to 1(at least)

Robert Wiblin said...

Here's vaccine spend to burden of disease ratios for a wider range of diseases: https://docs.google.com/spreadsheet/ccc?key=0AsOdBVMpAjJHdE5yR2tFbmFtZTlIbUI0UzdHRWtQclE&usp=drive_web#gid=0