R.O.I. on Cancer Spending: Better Than We Think?

In SuperFreakonomics, we write about how chemotherapy is ineffective for treating many forms of cancer, and that it is almost always very expensive. (We also write about the potential conflict of interest when clinical oncologists profit from the administering of these drugs, but that’s another topic for another day.)

Furthermore, it is commonly thought that the nearly-40-year “war on cancer” has largely been a failure, since the age-adjusted mortality rate for cancer is essentially unchanged over that time.

But that’s a deceptive metric. Consider this:

Believe it or not, this flat mortality rate actually hides some good news. Over the same period, age-adjusted mortality from cardiovascular disease has plummeted, from nearly 600 people per 100,000 to well below 300. What does this mean? Many people who in previous generations would have died from heart disease are now living long enough to die from cancer instead.

So how are we to think about the tremendous cost associated with fighting cancer these days?

In a fascinating and important paper new working paper called “An Economic Evaluation of the War on Cancer” (abstract here; pdf here), Eric C. Sun and five co-authors try to measure the degree to which spending on cancer R&D has proved efficient and worthwhile:

For decades, the U.S. public and private sectors have committed substantial resources towards cancer research, but the societal payoff has not been well-understood. We quantify the value of recent gains in cancer survival, and analyze the distribution of value among various stakeholders. Between 1988 and 2000, life expectancy for cancer patients increased by roughly four years, and the average willingness-to-pay for these survival gains was roughly $322,000. Improvements in cancer survival during this period created 23 million additional life-years and roughly $1.9 trillion of additional social value, implying that the average life-year was worth approximately $82,000 to its recipient.

Perhaps even more interesting:

Health care providers and pharmaceutical companies appropriated 5-19% of this total, with the rest accruing to patients. The share of value flowing to patients has been rising over time. These calculations suggest that from the patient’s point of view, the rate of return to R&D investments against cancer has been substantial.

This is good news, of course. It is also a reminder that if you hear a debate about health-care costs and it doesn’t heavily single out cancer costs, then the debate is radically incomplete.


The statistics also don't seem to take into account earlier detection. What if we are detecting cancers 4 years earlier? Not saying that is all of the 4 year difference, but it is a large chunk of it.


I'm curious: you guys live with statistics and you're willing to bring up a paper that tries to estimate massive flows of money over time and yet I've seen no mention of studies that discuss the number of people who die because they don't have health insurance. A very recent study by a group associated with Harvard found that nearly 45,000 Americans die each and every year because they don't have health insurance. Given the volume of posts about health, why not discuss that? And given the ferocity of feelings about the issue, why not discuss how we're so willing to spend vast amounts to save lives in one area, such as preventing a few hundred or few thousand lives lost or harmed by terror, and yet so many are unwilling to spend money to save 45,000 lives a year.


Cancer is for the most part a chronic disease which can be slowed considerably by treatment. Cancer medicines however are exceedingly expensive. It is a societal question to determine if the cost is justified. I take issue with your comment "clinical oncologists profit from the administering of these drugs" without mentioning the pharmaceutical industry which sets the astronomical prices and spends obscene amounts of money to market lobby for their use and coverage. This is the root of the cost issue, not the physician.


My wife is living proof that chemotherapy works on some people. There are many different types of cancer with many different response rates to chemotherapy. My wife was lucky enough (that is if you consider being diagnosed with breast caner at that age of 40 as "lucky") to be diagnosed with a particular type of cancer that while very agressive, it responds well to a drug called Herceptin. Between the February biopsy and October surgery, Herceptin had reduced her tumor from something we could feel to what the pathologist called "a complete pathalogical response." There was no trace of the tumor at all. She is here with us today because of chemotherapy, something my grandmother and aunt were not so lucky to have.


"Women with breast cancer treated with chemotherapy and after that with Tamoxifen had about 20% improvement in disease-free survival compared to women treated with tamoxifen only." (www.isrameds.com)


There are two things a pharma/biotech company will take into account before deciding to dedicate billions of dollars to a research area: scientific doability and market attractiveness.
Gains in the understanding of how tumors can be attacked at a molecular level have been considerable, which is why more companies want to put more money at risk to develop new products in this area. However, the vast majority of these projects will fail due to the remaining high hurdles of scientific doability.
Without the market attractiveness i.e. high prices, there would not be such increased focus into product development. New products are what we need and will continue to need in this area for a very long time.

Diana DeVito

Are you including insurance under "health care providers"? If so, how can insurance be covering less than 20% of the costs? If that's the case, they're evading their responsibilities.


Thanks for this post! I've always thought that the proper question is not how many people die of cancer - since we *all* have to die of *something* - but rather, how long people are able to live with cancer and how comfortable those years are. For the same reason, I've always been confused and a bit frustrated by disease awareness campaigns that focus on how a particular disease is the "#1 killer" or "#3 killer" of a certain population. Even if great strides were made in eradicating the "#1 killer" disease, something else would be the #1 killer. The focus should be on improving the prognosis for a disease, not merely decreasing its prevalence.

Mark (Chicago)

Craig (comment #1) identified the possible confounding effect of lead-time bias on the results of this study (i.e., earlier diagnosis appears to confer longer survival). From a brief read of the paper, it appears that the authors were aware of the potential effect of lead-time (pp 34-37) and investigated its effect on their data.

Their conclusion, however (on pg 36), is suspect: "In sum, lead time bias likely has quantitatively modest impacts on our results; if anything, it causes us to understate the value of survival gains, and the share of survival accounted for by improved treatment." Am I reading this correctly? Lead-time bias will cause us to OVERSTATE the value of survival gains by increasing the apparent length of survival.


In regarding Jonathan's post above and other media quotes I've seen, I've often wonder where these "45,000 pts a year die because of no insurance". If you have a serious illness like cancer and are truly indigent, most people qualify for Medicaid where most treatments are covered with no co-pays. I have seen middle to lower middle class people with serious illnesses having to spend down most of their money first before qualifying and unfortunately, that is the hard choice that people have to make for survival. Maybe we need better social workers and counselors to direct people to the help they need.


After being diagnosed with leukemia in 1989 and surviving two other cancers since then, I am convinced that whatever we are spending it is worth it! Because if I was diagnosed in 1969, I am probably just a tombstone somewhere now.


I would like to read about the root causes of cancer. While I believe aggressive treatment has improved people's lives - this is anecdotal as I have several cancer - survivor friends who are all happy to still be around - the thing that interests me here is how the spending is accounted for on the social and data ledgers.

What I mean is: As I understand it, spending on cancer R&D and treatment falls into the gross domestic product figures...which basically is a number that, if it goes up, things ought to be better because there is more economic activity.

Shouldn't spending like this be accounted for in a different way? Same with spending on, say, jail building? They both exist because of problems we haven't really fixed...they are, in the big picture, "whack - a- mole" solutions, expensive band - aids.

This can't be easy or it would be getting done, is my guess. I have read about "happiness" index..but do not know if it has gone so far as to re-organize the correlation of economic activity to happiness or a "better" world



New research on marijuana and cancer suggests that we've been thwarting the War on Cancer by fighting the War on Pot.

I draw your attention to the following peer-reviewed scientific research published in Cancer Prevention Research in August 2009:

A population-based case-control study of marijuana use and head and neck squamous cell carcinoma.


The authors of this study have concluded that "moderate marijuana use is associated with reduced risk of HNSCC."

The reduction is significant -- about 30-50%. And they mean "reduced" compared to people who don't smoke at all.

That's the same kind of result that the pulmonary disease researchers led by Dr. Donald Tashkin at UCLA Medical School found in their population-based study on marijuana and lung cancer.

Marijuana users, it turns out, exhibit the same or even slightly lower probability for developing lung cancers as people who don't smoke anything at all.

How could that possibly be true, one might ask in vehement protest, given the undisputed fact that the smoke from marijuana contains the same carcinogens found in all burning vegetable matter?

Research over the last 35 years has revealed that the active ingredients of marijuana function as natural anti-carcinogens. The carcinogens in the smoke appear to be canceled out by the anti-carcinogenic behavior of the components that get you high.

This is the direct opposite of the behavior researchers have noticed in nicotine. Nicotine appears to function as a cancer accelerant. The THC in marijuana appears to function as a cancer retardant.

A review of anti-carcinogenic properties of THC and its chemical cousins in marijuana was published in Nature Reviews - Cancer in October 2003:



American cancer researchers actually knew back when the War on Pot began that THC was able to kill lung and breast cancer cells in vitro.

The government didn't allow this research to be widely publicized because it might "send the wrong message about drugs to children."

Researchers have since then added another eight types of cancer cells to the list of cancers that can be killed or shrunk by THC and the other active ingredients found in marijuana.

Given the bulk of this research, it's entirely possible that .the War on Pot has been thwarting the War on Cancer for the last 35 years.

But the mainstream media cannot cover this story, alas. It might send the wrong message about drugs to children.



Data, data, data...So the fact that the moneys invested in cardiac research have resulted in a 50% decrease in mortality while the moneys invested in cancer research have provided no such relief is used as to argue that cancer research really does have a good ROI? You statisticians are a funny bunch - use the same data to argue any and every side, while, as #1 points out, there are plenty of alternative explanations that make a lot more sense than those proposed.

But I guess that's what happens when people trained in math and economics, and not medicine, interpret clinical results (and it's the same reason why the news media will blare the conclusions of lousy papers published in a journal like JAMA, when any scientist looking at the data will tell you the experimental design was flawed, the methodology was highly biased, and the paper is meaningless (as is often the case with pseudo-clinical studies in which the patients are not randomly selected and physicians are not blinded to who is receiving treatment vs. placebo, or for that matter when the "researcher-physician" is being paid by the drug company funding the study to get the results they want).


Alan Salzberg

I second Craig's comment concerning the bias of earlier detection: are we just detecting cancer earlier or are we actually prolonging life?

Also, general life expectancy has increased 2 years between 1990 and 2000, and more than a year for 60 year-olds, whether or not they have cancer (see, for example, http://www.infoplease.com/ipa/A0005140.html ). This should also discount the calculated increase for cancer patients.

Finally, willingness-to-pay data, in addition to being biased upward for wealthier people (wealthier people will say they are willing to pay more because they value each dollar less), is not the appropriate measure. We should instead look at the money spent as a limited resource and see if we can save more lives through some other measure--more comprehensive healthcare, prevention, etc.



Assuming for the sake of argument we could "cure cancer" tomorrow so that in 2011 nobody would die of cancer. Calculate the impact that would have on Medicare and Social Security. I think you'll see that the ROI on curing cancer would by tremendously negative.


Did you notice that the study received funding from Genentech, a company that markets some of the most expensive cancer drugs around?


"Health care providers and pharmaceutical companies appropriated 5-19% of this total, with the rest accruing to patients."

I don't have access to the full paper, so I can only wonder what this means.

If someone spends $32K per year (or whatever it may be) for chemotherapy and hospital treatment and doctor visits, isn't all of it going either to health care providers or pharmaceutical companies? What does it mean to say that health care providers and pharmaceutical companies "appropriated 5-19% of this total", with the remainder going to the patients?

My guess is that the statement has some special meaning understood (possibly) by economists but not by peasant folk.

Can someone translate this from economics-speak into English?


First, Herceptin isn't chemotherapy. It's a new immunological drug.

Second, apples and oranges comparisons abound. People living long enough to contract cancer has nothing to do with the ineffectiveness of chemotherapy.

If people were given an honest appraisal of their chances and were apprised of it's expense, they wouldn't take most chemotherapies. It's just a money maker for some doctors and hospitals.

Fact...most people die within four months of contracting a life threatening cancer. Sorry.


If anyone wants to know the real reason for the low ROI in cancer research, and most other areas of research, just visit your local academic research center. Find the chairperson of the oncology department and ask him/her how much grant money they have received in the course of their career. Then ask how many drugs or detection assays they've developed that are currently available to patients. You'll find that the reason he/she is chairman is because they are well established and have literally brought hundreds of millions of dollars into the university over 20 or 30 years. As far as drugs, they'll tell you about all the drugs they are "evaluating" (i.e. candidate drugs the doctors are being handsomely paid to test on terminal patients) but that no, they've never developed a drug or assay that has made it to market.

So why do these people hold chairs? For the same reason that the same senators and congressmen stay in office for 30 years; not because they pass great laws and look out for their constituents (in our case cancer patients and their families), but because they are the best at working the system and are the best politicians. Only in academics (or military appropriations, etc.) can someone receive hundreds of millions of dollars while not having produced anything useful, and still be praised and promoted as opposed to being put out of business.

As far as the drug companies go, they have a saying that they'd rather invest in the jockey than the horse. In other words, because drug development is such a long-term proposition, the biotech industry is a lot more like three card montey than a manufacturing industry. New ideas aren't funded by big pharma, they get funded by academic grants and then small business grants (SBIR/STTR), and then these small companies get bought by slightly larger companies until they reach a mass such that big biopharma might buy them out or partner with them.

So the primary question investors and big pharma ask when evaluating a potential technology isn't so much whether or not that technology will lead to a useful drug, as that's improbable and, even if it does come to fruition, unlikely to happen until after the people who make the decisions have retired. So instead the question is how successful the management team for this particular endeavor have been at raising capital and making profits for their investors in the past. The decision isn't about the drug and it's potential, which is too far off in the distance for the business types to care about, but how well can we play the shell game with these people and make a profit by obtaining those next rounds of funding and attracting the next set of investors. So, as in academics, people that are good at bs'ing investors and stock holders get rich, while less politically/industrially connected people with great ideas are completely ignored.