Podcast: Play in new window | Download
This week Zach and Kelly talk about the surprising lack of scientific evidence supporting the lifestyle changes suggested by doctors to people with gastroesophageal reflux disease (AKA acid reflux disease). Elizabeth Iorn’s blog post on Science Exchange starts a discussion on the importance of reproducibility in scientific research, and factors explaining the recent rise in retractions of scientific articles.
Links
Zach at ROFLCon (solo and Super Art Fight)
Kelly is awarded an American Association of University Women Dissertation Fellowship!
Zach and I will be the Chartists in Residence over at I Love Charts on May 1!
Gastroesophageal Reflux Disease (GERD)
Gastroesaphageal Reflux Disease
Scientific article: Are lifestyle measures effective in patients with gastroesophageal reflux disease? An evidence-based approach
WebMD: Top 10 Heartburn Food Triggers and Lifestyle Modifications for GERD
Scientific Misconduct
New York Times article on scientific misconduct
Dr. Iorn’s response to NYT article
Weekly Weinersmith Episode 16: Science Exchange and Dr. Elizabeth Iorns
Article on the importance of sharing computer source code: Shining Light into Black Boxes
Yes, it’s a problem (coincidentally, the first author’s name is “Gerd”):
http://citrixweb.mpib-berlin.mpg.de/montez/upload/PaperLibrary/GG_etAl_Helping_doctors-1.pdf
http://www.mpib-berlin.mpg.de/en/news/current-research-results/better-doctors-better-patients-better-decisions-envisioning-healthcare-in-2020
Related: http://www.plosmedicine.org/article/info:doi/10.1371/journal.pmed.0020124
Your podcast about GERD research and retractions was fascinating. I’m sorry you suffer from this Zach! Incidentally, the same day I listened to your podcast, there was an informational video playing at my local gym (Monterey, CA) about the “foods to avoid” etc. for GERD. Coincidence?
Your podcast also brought up for me some burning questions about cholesterol recommendations. So many people that I know in their mid thirties and beyond (friends, family, myself!) have been told by their doctor that their cholesterol is “borderline high.” Most of these people I know are really healthy, eat well, are not overweight, and exercise regularly. Yet, we’ve all gotten the requisite pamphlet about “foods to avoid” and “increasing exercise” to bring down our cholesterol. I wonder, what is going on? Is there really an epidemic of borderline-high cholesterol in healthy adults? Or are the criteria for cholesterol “normal” ranges changing? What’s the best up-to-date evidence of the link between blood cholesterol levels in routine tests and heart disease, and the links between the “recommended behavior changes” and cholesterol or heart disease. Lastly, there is some controversy surrounding the prescription (over-prescription) of cholesterol lowering drugs by doctors, influenced by pharmaceutical companies. Conspiracy theory? I don’t think so, it’s just plain good for business for these companies and many may be well intentioned. But can we separate myth from reality? The pharmaceutical companies certainly don’t have any inspiration to do this, or fund it. Thoughts?
Regarding reproducibility, my experience (in theoretical astrophysics) has been that if you’re starting a new project which builds on previous work, step 1 is always to reproduce the previous work as a special case of your new work. For instance, if you’re adding a parameter w to your model of some phenomenon because previous work had always implicitly assumed w=0, step 1 is to see whether you get the previous results in your model by setting w=0, even though you can handle other values.
I realize this might not be as easy to accomplish in a field where your experiments are expensive, but still, if your goal in say psychology is to answer “What effect does Y have on well-known effect X”, you should have a control where you verify that X is real. Often in psychology X is not real because it was only discovered at p=0.05, and the authors just got lucky (in that they are now famous and have tenured positions) / unlucky (in that the effect they discovered is not real).