How can we answer important questions?

A recent commenter asked the following question:

If you are a clinical researcher, how do you test a treatment for recurring body-wide pain and extreme fatigue that has lasted more than a year? How do you use science to test whether X treatment works? What measurement(s) do you make pre- and post-treatment to produce a clinical study that supports or indicates the underlying reality?

In this hypothetical, let’s say the reality is that the treatment produces significant reductions in the body-wide pain and fatigue in 98% the afflicted. How do you produce a legitimate, science-based study that supports this?

For someone who is suffering, the details of designing a clinical study may not be the first thing on their mind.  But for researchers, it must be.  Clinicians must then be able to assimilate relevant data to use in treating their patients.

The first task in looking at this question is, “what is the question?”  In this case, we need a disease or syndrome with a useful operational definition.  Since the commenter hasn’t given us one, we’ll choose “fibromyalgia”.  This syndrome is difficult but not impossible to study.  It’s difficult because it is syndromic; we can define a list of signs and symptoms and create an operational definition, but we cannot pathologically define the illness.

In this case, we can measure exactly what the commenter asks.  We can choose an intervention, say, a sleep aid called Miraculum, and design a randomized controlled trial.  Patients can be recruited, randomized to placebo or Miraculum, and outcomes of interest can be measured.  There are many tools to measure these outcomes including visual analog pain scales and quality of life measures.  The numbers can be crunched and interpreted, and voila, we have an answer (and usually more questions as well).

I hope this helps our curious commenter.

Previous Post

14 Comments

  1. Jack

     /  August 31, 2010

    This might be a stupid question not requiring a lengthy answer, but at least give some links when you bother to answer it with a fucking blog post. For someone in the position of asking that question, I don’t think your answer will do (if anything).

  2. I’m not sure I understand what you’re asking, but the link to the original is present above.

  3. Christina Pikas

     /  August 31, 2010

    I think you probably need a lot more detail in this area: “outcomes of interest can be measured. There are many tools to measure these outcomes including visual analog pain scales and quality of life measures.”

    How would one quantify “recurring body-wide pain and extreme fatigue”? If it’s something you can measure, then you can do so before and after an intervention, right?

    I actually agree that your answer is less than completely helpful. You might also talk about “a useful operational definition”

    But this comes from someone who is good at working with people to ask the right questions, and who doesn’t know any answers!

  4. Thanks—time to clarify things.

    An “operational definition” is a definition (of an illness in this case) that can be usefully applied to subjects/patients. One cannot measure something until you define what you are measuring. In this case, the commenter was not asking a good research question and we need to help him ask a valid research question.

    If he is simply asking about a single patient, we must strive to figure out the etiology of the symptoms.

  5. Mu

     /  August 31, 2010

    Didn’t they write a whole series about that phenomenon? I think the answer was 42, so I didn’t read all the way to the end to find out if they ever got the question.

  6. spit

     /  August 31, 2010

    You should add to this “hypothetical” that Miraculum is a tightly controlled substance, which makes researchers have to jump extra hoops, makes doctors paranoid about prescribing it even in the most obvious circumstances of extreme cataplexy, and means that only one company is currently allowed to produce it and is free to jack up the price to a truly obscene level that makes every insurance company massively question its every use. They’re not going to buy a “quality of life” survey, no matter what its merits.

    I realize you’re trying to stick to the we-can-do-science here, but this particular “hypothetical” is kind of politically loaded.

  7. David

     /  September 1, 2010

    If “miraculum” is oxybate, ten why be coy? The science behind it is fully described in posts on the FDA’s website:
    http://www.fda.gov/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/ArthritisDrugsAdvisoryCommittee/ucm203434.htm

    You can read the best case made for the product, by the sponsor company, and you can read the FDA’s analysis. The drug produces improvement in pain (marginal) and does not clearly produce improvement in sleep. In exchange, there are risks of substantial adverse events, overdose, dosing errors, abuse, diversion, and criminal misuse. The drug is used as a “date rape” cocktail.

    A panel of experts, shown that information, voted overwhelmingly that the product should not be approved. They made the right decision.

    no conflict of interest.

  8. Folks, I pulled Miraculum ex culo. It is not in re any particular real drug.

  9. Darn. I wanted me some Miraculum.

    Off topic complaint – there’s an auto-refresh thing going on here that’s truly annoying. It closes up comments and pops one back to index page.

    If you are doing this, shame! Shame on you! If it’s a scientopia thing, could you let them know it’s highly disrespectul to the reader to turn the page or close the book automatically?

  10. Vicki

     /  September 1, 2010

    I’m guessing that you would measure fatigue largely by asking the patients questions, both things like asking them to quantify how tired they are on something like the 0-10 pain scale, and about what activities they can do/chose to do. Someone with chronic fatigue is going to notice whether they had the energy to get to the grocery store, or were able to visit a friend or family member. That it can’t be measured mechanistically with something like a fever thermometer or blood pressure cuff doesn’t mean that it can’t be measured. “How are you feeling today?” is a legitimate medical question.

  11. plutosdad

     /  September 1, 2010

    I am interested in how you measure pain. You mention a few in the last paragraph, but someday a post on how we measure pain would be interesting. It seems like something that’s especially hard to do especially if the measurement relies on self-reporting. So I assume researchers try to get away from that.

    For instance I tend to under report pain. I think because of how I was raised. This makes getting treatment difficult for me, since I’m not writhing around or whining and the doctor doesn’t take my pain that seriously. I wish there really was an objective pain measurement, attach some electrodes to someone and say “it says your pain is at 7.5” that would be great! 🙂

    So I’m just wondering the same about researchers, since it seems they would also need to overcome differences in how people self report.

    thanks

    • Epinephrine

       /  September 2, 2010

      @plutosdad

      Yes, different people will report on things like pain differently. There are ways to account for this. One is the random assignment of people to groups – the theory being that with large enough groups, you’ll get a similar number of underreporters in each group, and if you don’t, their effect on the group will be small.

      Other ways to limit this effect is to use scales that aren’t purely subjective; instead of rating your pain on a scale, use a pain scale that corresponds to different degrees of limitation, either activity limitations or other types. A pain scale with descriptions can help a lot, an example is here. Since this scale provides some ways of gauging the pain other than a numerical scale it can be more useful.

      Another way is to let subjects serve as their own controls. One can still do random assignment, but perhaps have each subject receive either the drug for a period and then a placebo, or the placebo for a period, then the drug. This crossover type of experiment ensures that you get a placebo and a drug-associated pain rating for each subject, and by ensuring that half the group has the drug first and half the group has the placebo first you can account for order effects. Each subject can serve as his/her own control, by comparing drug/placebo scores.

  12. spit

     /  September 1, 2010

    palmd — fair enough, and apologies then. I guess I read too much in, but Miraculum _does_ sound an awful lot like Xyrem, and it’s a rather topical topic at the moment for Fibro.

    And David — it’s hard for me to buy that it doesn’t improve sleep when a major reason it’s approved for narcolepsy with cataplexy is for sleep consolidation, which seems to help ramp down cataplexy (though the “why” of that is unknown).

    I don’t have a strong opinion one way or the other about approving it for Fibro — if it seems to really help people, fine, if it doesn’t, fine. I don’t really have a dog in that fight, except for some basic sympathy for people with Fibro. But when I get the sense that it’s not being approved at least largely because of some OMG CLUB DRUGS! thing, I do have to wonder whether that’s an ok reason to keep it from being an approved option for doctors and patients to try in the right situations — lots of very helpful drugs have some abuse potential, and defining “abuse potential” is, as far as I’m concerned, kind of a weird thing generally anyway.

    I’m on modafinil for extreme daytime sleepiness, which is supposedly favored because of its low abuse potential, but I’ve got to tell you — I think they’re going to find over time that they’re wrong about that. It’s very helpful for me and I take the minimum I can to stay awake during the day, but I can absolutely see why some students I know used to take it recreationally, both for studying late and for partying. Defining “abuse potential” is not as simple, IMO, as “is it possibly euphoric?”

    As for the “date rape drug” thing — crap is salty as hell, I can’t imagine that it would be all that terribly common in reality, though I haven’t looked into actual statistics, to be fair — if your cocktail tastes like somebody spiked it with soy sauce, don’t drink it. The most commonly used “date rape drug” in this country is alcohol, by the way.

  13. spit

     /  September 1, 2010

    Oh, Vicki, I couldn’t agree with you more. Subjective patient experience is so important. It really bothers me that there are so many moves by _some_ doctors and researchers to discount patient accounts as useless — I’ve had a few of those doctors, and I’ve switched away from them very quickly. One can sometimes sense the immediate suspicion.

    IMO, the vast, vast majority of patients are trying to give the most straightforward account they can of their experience. Unless there’s reason to think they’re bullshitting or drug-seeking or hugely exaggerating, I think they should be taken at face value, trying to honestly find answers to their health concerns. Not that it’s always necessarily enough to do more than find some directions to look in or some things to try for symptom management, but that’s a start.

%d bloggers like this: