Mar 28 2013

Do 97% of UK Doctors Prescribe Placebos?

A recently published survey at PLOS One of UK primary care doctors reports that 97% have prescribed an “impure placebo” at least once in their career. Most news reporting of this survey leave off the “impure” bit.

Let’s take a closer look at what this means.

The survey asked about “pure” placebos, which are inactive sugar pills or a similar inactive treatment, and “impure” placebos, which are effective medicines given in a way that might not have a clinical effect. The survey found that only 12% have ever prescribed a pure placebo, and only 1% do it on a regular basis. However, 97% have prescribed impure placebo ever, and 77% on a regular basis.

The survey had a 48% response rate, which is not bad for a survey, but this is why surveys are not considered strictly scientific. The low response rate introduces the potential for systematic bias – perhaps people who choose to respond do so because of a ┬ácertain attitude or belief that biases their responses.

That aside, let’s look at what the authors consider “impure placebo.”

  • Positive suggestions
  • Nutritional supplements for conditions unlikely to benefit from this therapy (such as vitamin C for cancer)
  • Probiotics for diarrhea
  • Peppermint pills for pharyngitis
  • Antibiotics for suspected viral infections
  • Sub-clinical doses of otherwise effective therapies
  • Off-label uses of potentially effective therapies
  • Complementary and Alternative medicine (CAM) whose effectiveness is not evidence-based
  • Conventional medicine whose effectiveness is not evidence-based
  • Diagnostic practices based on the patient’s request or to calm the patient such as
    • Non-essential physical examinations
    • Non-essential technical examinations of the patient (blood tests, X-rays)

In other words – if a physician ever does something to enhance the therapeutic relationship with the patient, that’s an impure placebo. I would have to consider myself an “impure placebo” prescriber, then. Sometimes I will do a brief neurological exam on a follow up patient not because it is definitely needed (it’s unlikely to change my management) , but because patient’s expect it will be part of their office visit and it makes them feel as if they are getting the attention their condition requires.

What about subclinical doses? This is a gray area – patients respond to medication differently (different metabolism and sensitivity), and a subclinical dose for one patient may have an effect on another. It’s common when using symptomatic treatment to start at a low dose and slowly increase the dose until the patient gets unwanted side effects or until they get a good response. What if they get a good response at a low dose, one that is “subclinical” on average, but may be having an effect on the patient? Is that an “impure placebo?”

Nutritional supplements are another gray area. For example, I check B12 levels on most of my patients. Low B12 can affect nervous system function. If the level is low, I prescribe supplements to get them into the normal range. I am optimizing any parameters that might affect neurological function, even if I don’t think the low B12 is causing their current symptoms. If the patient gets a positive placebo effect from taking the B12 also, that’s great. Is that an impure placebo?

Off-label use is also a tricky area. Many drugs have good evidence and a solid rationale for uses that are not FDA indicated. Drug companies usually don’t bother getting every possible indication for their drugs, because that is expensive. They count on off-label use – which can be entirely evidence-based and effective. Do the authors, however, consider all off-label use impure placebos?

In short – there are many problems with interpreting this survey, and the 97% figure that is being widely quoted is highly problematic and misleading. This is a useful area of medical practice to explore, but it is very complex and would require much more nuance than provided in this survey.

Note – Scott Gavura also wrote about this article at Science-Based Medicine.

5 responses so far

5 thoughts on “Do 97% of UK Doctors Prescribe Placebos?”

  1. SARA says:

    If you low dose a patient and get a response, how do you know it’s not a placebo effect?

    When the placebo wears off and the real medical problem recurs, do you up the dosage or change it or wonder if your diagnosis is wrong?

  2. With individual patients it’s very hard to know if they are really responding to any treatment (therapeutic dose range or not), or just getting a placebo effect, or the illness would have improved on its own anyway. There’s no blinding or control and you can’t do statistics on individual patients. You are applying group or population data probablistically to individuals.

    I can’t really answer the second question (at least not in a comment) – there are way too many variables that need to be considered. Suffice to say, lack of response to treatment does provoke a rethinking of the approach.

  3. locutusbrg says:

    I have often found that PLOS one publishes marginal research.

    Kind of a loaded dice here. Who really practices utilizing none of these techniques. The only thing that surprises me is the 3% who responded negatively. I suspect that they just did not take the time to read and understand that the “impure placebo’s” should be included.

    If you have ever prescribed or recommended any old medication you are probably recommending “Conventional medicine whose effectiveness is not evidence-based”.
    I can not honestly say that I have never given an antibiotic just to CMA over a infection that I believed was probably viral. No placebo intent just concerned about liability and uncertainty.

    I have often given equivocal or evasive answers about scam treatment because the patient is new to me and I am trying to develop a relationship before blunting pseudoscience. Does that count?

    If you really think about the variety of “impure placebo’s” they are just to broad enough to be inclusive to all professionals. Plus just vague enough to appear exclusionary to lay people who do not understand that treating humans means acknowledging the inherent problems of the human brain.

  4. Diane says:

    My lab actually studies antibiotic interactions, and some of my coworkers have shown that sometimes a combination of antibiotics, each at a subclinical dose, can together be effective. So I am a bit surprised that the authors listed prescribing subclinical doses, in and of itself, as a placebo.

    By the way, PLOS One has only minimal peer review. They are quite explicit about it. They review papers only for technical accuracy, not for the quality of the research. Consequently, you must take PLOS One papers with a grain of salt.

  5. I suspect that this is common in the US too: prescribing an antibiotic for a virus is probably the biggest example.

    I am told that the physicians that do this, do it when their patients demand it. I wonder how accurate the 97% is, and I wonder to what extent the UK physicians are doing this when a patient demands it.

Leave a Reply