Ask Emily: How do I separate fact from fiction on cancer treatments?

Dear Emily,

My wife has a pretty common problem: cancer. While we’re very fortunate that her brand of the condition (not-yet-metastatic melanoma) hasn’t led to any life-threatening complications just yet, we’re trying to be vigilant on any life changes or treatment protocols that might help her stave off development of her cancer.

We’re both open to alternative forms of treatment beyond what’s we often hear of (surgery, radiation, chemotherapy). But it’s distressing to find so little good information on alternative therapies out there in the world and online. There has to be some way to filter out what hard data exists from the scams and the hearsay that’s out there. With your background separating medical fact from fiction, could you provide any advice on how we could best perform our own research? Are there any tools or resources you’d recommend to people looking to try and find alternative treatments, but who aren’t willing to completely compromise on science and data?


First, I’m very sorry to hear about your wife’s illness and very much hope that complications stay away.

The question you ask is a good one and relevant to virtually anyone who faces any medical issues. Even digging into research on standard therapies can be daunting, and the evidence can be surprisingly uninformative. Moreover, the internet is the obvious place to look for information–but there is no shortage of misinformation there. But this doesn’t mean you cannot make it work for you.

The basic approach I recommend is to prioritize your research according to the studies that will provide the best-quality evidence. Usually, this means looking for evidence from trials that were randomized and controlled. Such studies enroll a group of people and randomly assign some of them to receive a treatment, then compare their results to a randomly assigned control group. These types of trials are most likely to lead to firm conclusions about the cause and effect of different factors.

Failing that, go next to reviews that summarize many observational studies. (Observational studies just compare people who get a certain treatment to those who do not, without randomizing who gets the treatment.) These are not as good as randomized studies, but they can still be informative. Failing that, go to single observational studies. And failing that–well, I’ll try to address that at the end.

But first, here is a quick rundown of specific sites I’d recommend.

Step 1: Cochrane Reviews

If you are trying to figure out the efficacy of some treatment, the best-case scenario is that there are a bunch of large, randomized, controlled trials. In the medical space, these are sometimes summarized in review articles published by Cochrane, an independent nonprofit that organizes medical research. Cochrane reviews are typically well-done and easily searchable.

You can start at this website and search for keywords of interest. These reviews are typically focused on standard treatments, but you can sometimes find reviews of alternative treatments. For example, here is an article on the role of mistletoe in cancer treatments.

Step 2: Pubmed

Assuming you strike out on Cochrane, you may need to go next to Pubmed. This website is a service of the National Institutes of Health (NIH) and catalogs essentially all published papers in medical journals in a searchable format. It can be overwhelming, but it can also be hugely useful.

The website’s search function can be easily filtered by method. Let’s say you’re interested in the impact of vitamin D supplements on melanoma–a commonly discussed alternative treatment. I’d start with a simple search (say, “melanoma and vitamin D”) and limit the results to randomized controlled trials. The result gives you just one completed trial (see here). What is especially nice about Pubmed is that they show you not just the title but also the abstract of the paper. In medicine, you can almost always learn the conclusions of the paper, and the basic methods, from just the abstract.

If you fail to find something here, I’d begin to look beyond randomized trials and see what you can learn from observational studies. Figuring out which of these are good can be a bit of an art form, but you want to look for studies with a large sample size (lots of people) and which use regression to adjust for other differences across people (like education, age, or health status).

The risks and rewards of online searches

Lastly, you may be wondering what to do if you cannot find anything from these sources. Is it a good idea to dive into the pool of articles on the internet?

My view is no. If there is evidence in favor of some treatment–even very minimal evidence, like one descriptive study of 25 people–it will appear in medical literature. If you cannot find evidence about your question in this literature, it means there is no evidence. Unfortunately, the experiences of some strangers on the internet are not a safe substitute for hard data.

With all that in mind, good luck in your searches–and the best of luck to you and your wife.

Emily Oster is an associate professor of economics at Brown University and the author of “Expecting Better: Why the Conventional Pregnancy Wisdom Is Wrong — and What You Really Need to Know.

Got an everyday problem that could use an economist’s point of view? Send Emily your questions at

home our picks popular latest obsessions search