Bad Science Watch: acupuncture and lung cancer pain

Having needles shoved in your skin can seem like a major medical intervention, but science says it's a sham. Is a new study about lung cancer pain and acupuncture going to change that picture? In short . . . no. It's terrible science. Photo from Ref. 4.

Having needles shoved in your skin can seem like a major medical intervention, but science says it’s a sham. Is a new study about lung cancer pain and acupuncture going to change that picture? In short . . . no. It’s terrible science. Photo from Ref. 4.

I saw this headline in my Google News feed this morning: “Acupuncture Reduces Pain in Lung Cancer Patients – New Findings” [1]. The article was posted on a credulous site that promotes acupuncture, a practice that has never been proven to yield any benefit over the placebo effect.

So I thumbed through the news article. It claimed:

The focus of the study was the evaluation of acupuncture and its effect on lung cancer related symptoms such as nausea, anxiety, pain, depression and a sense of not

feeling well. A total of 33 lung cancer patients received 45 minute acupuncture treatments at a rate of 1 – 2 times per week for at least 4 treatments. The results demonstrated that acupuncture effectively reduced pain levels in over 60% of patients. Additionally, 30% of patients noted improvements with a sense of well-being with at least 4 acupuncture treatments. This number jumps to a 70% improvement in well-being when patients received 6 or more acupuncture treatments.  [1]

The original paper making the claim of such findings is references in [2]. A read through the abstract of the paper reveals that the news site summarized it appropriately. Reading the abstract also reveals a set of major scientific flaws:

  1. There is no placebo control group in the study.
  2. There is no group that receives the best known and possible medical pain intervention for lung cancer patients.

The placebo effect is a powerful, built-in effect wherein belief about outcome affects outcome. If a patient receives a sham, non-medical treatment, but the treatment provider believes it’s effective and the patience believes it’s effective, then the patient can get better even though no medical intervention actually occurs. It must be controlled for before declaring a treatment successful. Degree of belief determines outcome, independent of the activity of the treatment.

While for acupuncture it is difficult to provide a sham control procedure, in fact sham acupuncture (where the needles press against, but never puncture, the skin) exists and has been used in properly controlled studies of acupuncture. Those independent studies have shown no effect of acupuncture over placebo. At best, a patient is paying for the privilege of being deceived.

This study [2] is bad science. Without a control group, the most accurate and scientifically honest interpretation is that the patients experienced the placebo effect. Studies of placebo have shown that “more sham intervention equals faster positive outcome,” because the belief is that more treatment means more healing. The results – more acupuncture led to more reports of pain relief – are consistent with the null hypothesis. That is, the data are consistent with acupuncture behaving as a placebo, where more placebo speeds the placebo effect.

You can find lots of good information about the amazing placebo effect, and understand why it’s critical to control for it, in ref [3]. What is most scientifically DISHONEST about this paper is that the study fails to do the most important thing: compare against the BEST POSSIBLE TREATMENT. It’s not enough to test against nothing (e.g. placebo) and then claim your treatment is equal to or better than doing nothing – it’s medically and scientifically honest to test against the best pain control interventions. 



[2] Kasymjanova, G., M. Grossman, T. Tran, R. T. Jagoe, V. Cohen, C. Pepe, D. Small, and J. Agulnik. “The potential role for acupuncture in treating symptoms in patients with lung cancer: an observational longitudinal study.” Current Oncology 20, no. 3 (2013): 152.

[3] David Moermon. “Cultural Variations in the Placebo Effect: Ulcers, Anxiety, and Blood Pressure”. Medical Anthropology Quarterly. Vol 14, Issue 1. 2008


A scientist in pseudoscientist’s clothing

A news aggregator for stories that support anti-science like "GMO food labeling" seems to have picked up my story about separating science from policy. Oops. My headline worked.

A news aggregator for stories that support anti-science like “GMO food labeling” seems to have picked up my story about separating science from policy. Oops. My headline worked.

This morning, I was sent a direct notice on Twitter indicating that something called “FoodIDTheft” had aggregated one of my tweets into some automated news feed. I was intrigued – since “FoodIDTheft” sounded like an inflammatory name intended to convey that this site helps people keep their food’s identity from being stolen.

I checked out the site. It’s generated by, and it’s an aggretator site for bad food science and sexy food headlines with little substance. It’s clearly intended for people who want to believe that simple nonsense gimmicks or “emerging science” will save their food, rather than understanding good science and separating science from policy based on science.

Here is the Twitter account’s self-description: “  is a website that alerts consumers about threats to the food we buy for ourselves and our families. Sponsored by Citizens for Health.” Nice use of propaganda tactics, “Citizens for Health” – whomever you are.

I was pleased to see my recent blog post about separating the “labeling food” issue from the scientific issue of “biology and genetic modification,” so that we can arrive at a correct policy to deal with corporate food interests. My story was not in support of labeling food as “GMO” – labeling food as “GMO” is, at best, naive . . . at worst, it demonstrates utter and total scientific ignorance . . . and in both cases, it fails to achieve the actual policy goal, which is to curb corporate power and the proliferation of proprietary food.

So . . . thanks story aggregator! I know you probably interpreted my blog post title to mean I intended to support “labeling food as GMO” – you were wrong. That headline made me look like a pseudoscientist, but instead if you read the actual content of the post you’ll see that I am a scientist whose clever window-dressing fooled you into thinking you were getting confirming content to support your unscientific world view.


What’s that whining?

Is that whining I hear?  ID proponent Casey Luskin argues ID is scientific - using a total misunderstanding of the scientific method.

Is that whining I hear? ID proponent Casey Luskin argues ID is scientific – using a total misunderstanding of the scientific method.

The Discovery Institute posted a long and painful defense of Intelligent Design as a “scientific idea” today [1]. It was painful because it seemed based on a pre-college level of understanding of the scientific method. Their primary defense was founded on shoe-horning intelligent design into a scientific method-looking structure, but evidence of their misunderstanding and misrepresentation of science is evident right in their application of the steps of the method.

The gross stages of the scientific method are: to make an observation of a phenomenon in the natural world; construct a hypothesis – a testable idea – to explain the observation; conduct a test of the predictions of the hypothesis; and then draw conclusion about the test (was the hypothesis disproven? If not, conduct another test of its predictive powers).

Casey Luskin writes the post, and here is one example of his shoe-horning bad science into something that sounds sciency:

Observation: Intelligent agents solve complex problems by acting with an end goal in mind, producing high levels of [complex specified information (CSI)]. In our experience, systems with large amounts of specified complexity — such as codes and languages — invariably originate from an intelligent source. Likewise, in our experience, intelligence is the only known cause of irreducibly complex machines.

Hypothesis (Prediction): Natural structures will be found that contain many parts arranged in intricate patterns (including irreducible complexity) that perform a specific function — indicating high levels of CSI.

Experiment: Experimental investigations of DNA indicate that it is full of a CSI-rich, language-based code. Biologists have performed mutational sensitivity tests on proteins and determined that their amino acid sequences are highly specified.22 Additionally, genetic knockout experiments and other studies have shown that some molecular machines, like the flagellum, are irreducibly complex.23

Conclusion: The high levels of CSI — including irreducible complexity — in biochemical systems are best explained by the action of an intelligent agent.

There are so many glaring errors in this – some so big, you might miss them at first – that it’s painful. PAINFUL. Here are some obvious ones.

  • Complex specific information (CSI) is a concept that was articulated by William Dembski in some of the few published and peer-reviewed papers produced by any proponent of Intelligent Design. It’s been roundly refuted by mathematicians and philosophers [2], its foundations shown to be mathematically inconsistent and incorrect, and its assumptions weak and poor. Not only that, Demsbki’s filter for detecting design has been shown to find design in pure randomness. Oops. An unreliable method is a terrible foundation for a solid hypothesis, as any beginning student of the scientific method can tell you. This is akin to forming a hypothesis based on the assertion that the “earth is flat.”
  • The hypothesis naively and narrowly assumes that the form of intelligence we know about – human intelligence – is the only kind of intelligence that can produce information in a system; or, more to the point, that when we see complexity that could have resulted from human intelligence, we are to assume that some intelligence caused it. It’s unscientific hubris of the worst kind to assert that complexity, based on engineering choices humans might make, represents the only kind of intelligence. There is no generic definition of intelligence (certainly not from complex specified information, which is discredited). Since there is no evidence-based foundation for this assertion, this assertion is weak, and thus forms another weak basis for the hypothesis.
  • They invoke “experience” to justify assertions. Wrong. Evidence and data is what you use to justify the a proposed explanation of some observation. “Experience” suggests a personal set of small-statistics, biased by your set of expectations; data and evidence, however, are collected via a rigorous protocol and subjected to quality assessment. “Experience” is no substitute for evidence.

The worst error here is the most obvious, and hardest to see. Let me simplify Luskin’s example, removing all the fluff to boil down his example to its core.

Observation: Intelligent agents result in complexity.

Hypothesis (Prediction): I will observe complexity in nature, and that can only be explained by intelligence.

Experiment: Other people have noticed that there are some hard-to-explain (“complex”) structures in nature.

Conclusion: Complexity is best explained by the action of an intelligent agent.

See it? Luskin’s example is a tautology – assuming the premise to draw the conclusion. The truth of the premise cannot be refuted, by construction. Nowhere in here does he begin with or even ADMIT the most established explanation of complexity in nature – Natural Selection. Instead, he weaves a logical loop that never ends – a Moebius of Misunderstanding.

This is not science. This is bullshit.

A proper hypothesis-test-conclusion would be the following:

  • I observe X – something that seems complex to me – whose explanation is not immediately obvious to me.
  • I hypothesize that there is a path by which this structure arose in nature, and apply the known physical laws to attempt to explain it.
  • I construct an experiment to determine if X could have arisen naturally, from forces in the natural world (chemistry, physics, biology). This experiment could, for instance, be to try to recreate the structure in the lab through Natural Selection forces or, more simply, to look in the natural world for related structures that might have evolved into X through the application of Natural Selection.
  • Either the hypothesis passes the test or not. I publish in a peer-reviewed journal either way, subjecting my method and results to a bloodbath of criticism. If successful, it gets published and the idea enters the public domain, to be confirmed or refuted by independent experimentation.

At no point do I conclude: “Too hard – must have been designed by some intelligence.” That’s called “giving up.”

Intelligent Design Creationists try to represent themselves as something other than a straw man, and in doing so demonstrated that they are both a straw man of themselves and, worse, they are begging the question.

Don’t even get me started about the irony of them whining about being treated like a straw man, when they haven’t demonstrated that they understand the basic principles of the Theory of Natural Selection during the entire time they’ve been seeking to recast it as a straw man and then knock it down.


[2] For examples of refutations of Dembski’s discredited ideas, with references, see

Getting the “science” right in “science policy”: labeling food as “GMO”

I’ve been engaged in a small conversation on a social network regarding the labeling of food as “genetically modified.” I’ve written specific thoughts about the scientific problems with that labeling recently in this blog [1]. I thought it might be useful to explain here how I got engaged in this most recent conversation, which is not yet concluded (I think), and what I’ve said for my part in the conversation.

It started when I noticed that somebody on my “following” list had re-posted comments from another person (whom I am not following). The comments were in reference to the editors of Scientific American coming out in writing as being against GMO food labeling [2], for all the reasons that scientific organizations, like the AAAS, have also come out against them: they are scientifically inaccurate, and thus are the use of bad science to inform science-based policy. The editors state their position clearly and crisply, “Instead of providing people with useful information, mandatory GMO labels would only intensify the misconception that so-called Frankenfoods endanger people’s health . . . ”

The comments on this article that made some vague implications about consumer choice and the “naivety of scientists,” with heavy overtones of having conflated several issues – some scientific and some not – into one straw man issue. The original poster complained that the Scientific American editors ” . . . can’t possibly let people have the facts . . . ” and referred to the naivety of scientists regarding the substance of the issue, which was not about the science but rather about control of the world’s food chain (with specific reference to Monsanto).

Both the original issue – GMO food labeling – and the comments that started me talking about this,  grossly oversimplify the science (biology and genetic modification), setting up a straw man to be knocked down (“GMOs”), and then risk achieving the wrong goal. The right goal has nothing to do with genetic modification itself, but rather with the practices of some of the larger companies that employ this and many other practices in a way that creates a proprietary and protected food market under their control.

I posted a response to the person who shared the comments, saying,

I am curious: why do you believe that labeling food as “GMO” (genetically modified organism) will better inform consumers? All food – organic and otherwise – is genetically modified. It’s one thing to oppose food made by a food maker because of their economic or political practices; but to insist on non-scientific labeling of food is not the solution. I think, what you probably prefer, is clearer labeling about who made the food. That’s scientifically accurate, and lets consumers decide if they will send their money to, for instance, Monsanto. Labeling a food as “GMO” implies something bad or different about it; that’s scientifically inaccurate, since all food is genetically modified, only with varying degrees of control (the invention of agriculture was the invention of human genetic modification of plants through breeding and control – even organic farming selects winning traits and breeds out losing ones, and that’s the definition of “GMO”).

Scientists oppose this not out of naivete, but out of the desire for scientific accuracy in policy making. It’s a complete mis-representation of food science and biology to imply that GMO = bad. It’s a different thing to insist companies clearly label themselves as the manufacturer so consumers can make clear choices to spend currency on one company or another.

I want Monsanto to answer for their practices, and pay the price for their shady business in the marketplace. But when science is thrust into the public sphere, I also want it accurately represented in policy. GMO labeling, as currently pursued in states, in inaccurate and represents a gross misunderstanding of biology.

Another person engaged in the conversation, which was great, but seemed to have missed the point of what I was saying. They again conflated “food labeling” with “labeling GMO food” – not the same things. I support labeling food . . . just not using the straw man of “GMO,” but rather for the substantive cause of mapping the food supply, its origins, its producers, and its handlers. “GMO” is useless, in that sense.

Here was my reply to the comment I received:

Please don’t misunderstand what I am saying:

  • I support labeling food to better indicate where and by whom it was produced; I would argue that the pinnacle of this labeling process would be full food-chain tracking (e.g. to help in unraveling food-born disease outbreaks).
  • I support corporate food policy reform, especially regarding the patenting of genetic modifications; I support the enforcement of “publish all scientific trials” about food and drugs (including null results, so that we can actually assess the rate at which foods/drugs have the claimed benefits).
  • However, since all food is genetically modified, if one is going to label some foods as “GMO” then scientifically one must label all food as “GMO.” Since science is being injected into this particular policy discussion, the science must be correct. As a citizen scientist, I will not stand idle while people misappropriate the field and its terminology and, in doing so, achieve the wrong political and policy end. Nobody has studied the environmental or health effects of every plant humans have modified since inventing agriculture; nonetheless, every time we bred for certain traits we modified, in a gross and largely uncontrolled way, the genomes of animals and plants – with unknown repercussions that are potentially far worse and far less-well-understood than the precision genetic modifications we are capable of executing now (along with double-blind studies, etc.).

I suspect that, at the real core of this debate is a desire to curb the business practices of large food corporations – food and gene patents, control of food commodity markets, and crushing family farms under legions of lawyers. I believe that my first two bullets above would go a long way to addressing that issue, without misrepresenting genetic modification as the straw man to be knocked down.

When you say “commonly consider” in reference to genetic modification and its definition, I would argue that using “common knowledge” as a defense of labeling is akin to advocating mob rule in science, and thus in science-based policy. Majority opinion does not rule in science; if that were the case, we’d have to teach astrology along side astronomy (since 50% of U.S. residents believe in astrology, that the stars decide our fate). I reject the notion of applying common public misconceptions of biology to then defend the labeling of food as “GMO.”

Rather than foster public misconception of science, I rather support the idea of better educating the public on matters of science – including biology, genetics, food science, and related issues – while at the same time curbing corporate overreach into the pantry and decisions made about food. These are two separate issues, in reality. GMO food labeling conflates them in a bad way that misrepresents science, and doesn’t actually address the real problem that, I think, people actually want to solve.

I care deeply about accurate use of science when informing science-based policy. To label one food as “GMO” is to completely get the core of biology incorrect. Besides, what people really want is to hold companies accountable for their behavior in the marketplace. Bastardizing science in the name of this quest will only do two things: breed fundamental misconceptions about science that will inhibit our society’s ability to function in a world that demands more creativity and innovation through science; and fail to accomplish the real goal, which is curbing or putting more checks on corporate power in a global economy where control of the food supply is, in fact, at stake.

[1] “Why you believe what you believe” (August 11, 2013)

[2] “Labels for GMO foods are a bad idea.” (Scientific American, August, 2013)