Nonetheless, despite
what we might call nonscientific origins of these professions,
they’re very popular with patients who have back pain. These are results of a survey
that Consumer Reports did just a few years ago with almost 14,000
respondents who had back pain, asking who did they see and how
pleased were they with the care that they got. And you can see that chiropractic
care topped the list, and physical therapy
came in close behind, and acupuncture came
in close behind that. Medical specialists were
much farther down the list, and primary care physicians like
myself were even much farther down the list. So you can see that the public
sort of votes with its feet, and they seem to
appreciate these treatments. So if these treatments
are so popular, you might assume they must work. How could it be otherwise? But I’ll spend a couple of
minutes digressing about some of the reasons why it’s
so easy to be tricked, in a sense, by improvement
that occurs after someone has a treatment. The first of those
reasons is simply the natural history of disease. It turns out that acute
back pain at least has a very favorable prognosis. And if you look at
everyone with acute back pain, most are actually substantially
better within a month. So you can appreciate that if
somebody gets an acupuncture treatment and then gets better,
it might be because of the acupuncture, but it might also
be just because of the natural history of the condition. And so this is sort of what the
attorneys would refer to as the post hoc, ergo
propter hoc fallacy. You assume that just because one
thing came after the other that it was caused by
that other thing, and it ain’t necessarily so. Then there’s this concept of
regression to the mean that the statisticians would
remind us about. The idea is simply that patients
may seek care when their symptoms are most extreme, and
then when we follow them up, they’re back towards closer to
their average level of pain or symptoms and therefore it
appears they’ve improved– so plenty of room
to be fooled there. Placebo effects–I think
everybody sort of understands or at least
acknowledges them, although we probably underestimate
how powerful they may be at least in terms of
pain management. And finally something that I
think is different still from all of those things above is
just the care and attention and warmth and concern that
comes from a good clinician, that itself seems to
have an important effect. And I think these cumulatively
are some of the reasons why I think it’s so important that we
really study these treatments in the form of randomized trials
that give us a chance at least to control for some
of these factors. And let me just
elaborate a little bit on a couple of
these things. This may be old hat to many, but I think it’s still
easy to get fooled. Just a little riff here
about the problem of regression to the mean. I think many of us imagine
that chronic conditions are something like this. You see a patient with symptoms
up here at some high level, and then we intervene and then
they get better and we measure the outcome sometime
later and they’re better. But I would argue that in
fact the natural history of most chronic illnesses and
most chronic symptoms looks more like this. They have good
days and bad days. This person has pain every day. This is certainly chronic pain. But like other conditions,
they have good days and they have bad days. And when are they most likely to
go to the doctor or seek care or get enrolled in a
clinical trial? It’s probably here where
their symptoms are the worst. So they come in to
see us at their peak. And then if we follow
them up later sometime, just by random variation,
they’re likely to be back closer to their average level of pain. And it’s easy to make the false
assumption that we’ve seen a therapeutic effect when in fact
we’ve really just seen sort of like a return to
their average level. And then the
placebo effects, I think, are easy to underestimate. There have been some analyses
of trials that compared placebo therapy to
nothing, doing nothing. And it seems in those
comparisons that the placebo effects probably do not
affect patient survival or hard physiological measures. But it seems that they do have
an impact on pain perception and reports of pain, and probably most people
respond to placebo therapy. There’s sort of this widely
cited 30 percent response rate to placebos. But in fact when you look at
a variety of clinical trials, you see up to 85 percent of
people may respond to placebo therapy, and I suspect that
we all actually are placebo responders at one
time or another. And in fact, placebos
have a lot of properties that look like active therapy. So, for example, two pills
appear to be better than one pill when you’re
taking placebos, and a more expensive placebo
appears to be more effective than a less expensive placebo. And placebos are often
associated with a variety of sort of non-specific
side effects– drowsiness and headache
and nausea and so forth. And there’s this budding
literature at least that suggests that, while we know
pills have a placebo effect, they may actually have a weaker
placebo effect than injections, and injections may in turn have
a weaker placebo effect than open surgery. So it may be that surgery
actually has what’s been dubbed a mega-placebo effect
that’s only observable when you look at trials of
sham surgical interventions of which there are a few. And then there’s this issue of
just the care and attention that you get from a clinician, and is
that more than just a placebo? There’s this nice clinical
trial that Ted Kaptchuk did. Ted may be familiar to many of
you because he’s been involved at Harvard in the study of
complementary and alternative treatments, and he devised a
randomized trial for patients with irritable bowel syndrome. So this is not back pain but
another painful condition. And he randomized
patients to three groups. First, there was a group that
was randomly allocated to a waiting list, and about a
quarter of them improved over the span of 3 weeks. Then he had a group
that got sham acupuncture. This is with sort of a
specialized acupuncture needle that was specially designed
to mimic real acupuncture. But this was delivered by
therapists who were instructed not to talk very
much with a patient, don’t offer much in the way of
reassurance or concern or that sort of thing–just do
the sham acupuncture. And with that
placebo intervention, 44 percent of patients
were improved at the 3-week follow up. And then in his third arm, he
had the sham acupuncture again but lots of concern and
confidence and attention from the provider, and
there was a boost again. Now 62 percent of
patients had improved by the 3-week follow up. So a suggestion here that
concern and attention from a good clinician really have
another impact above and beyond the conventional placebo effect. So those kinds of things
create some real methodologic challenges when you try to
go to study CAM therapies.