Doing Research That Might Not Work

When I was a little guy I thought research science type people were pretty much all about power. Big hair, white coat, potions that bubble and a maniacal laugh. That was the cardboard cutout version of a science person I would have pinned on my wall.

Then I got a tiny bit older and got a tiny bit more of a clue and I realised that mostly these science people who do research were on a mission to discover massive things that would change everything about everything. Or at least everything about one thing. It seemed like something as cool as magic that wasn’t actually magic. Well, I did say it was a tiny bit more of a clue.

Decades later and I find myself trying to do research and I finally understand more that researchers are generally just people trying to answer questions. In a way that involves a bit of method to get you there. And that also involves all the magic of repeatedly walking into a rake.

And that occasionally means researching stuff that might not work.


Simple Questions

Just in case you haven’t read all the background on this research project (I’m guessing that’s everyone) it’s about brains and lights.

It started with an observation. Whether I’m doing work in prehospital medicine or helping kids snooze at the hospital I am very interested in the brain. When we go to accidents in particular we see our fair share of people who have injuries to the brain. This is a big deal.

With minor injuries to the brain, it might well be that you are briefly unwell and then you recover. Some brain injuries leave you with permanent problems in just about any function you can name. Speech, thought, movement, sensation, anything.

What might be less obvious is that the injury doesn’t just happen when your head gets rattled. The injury triggers off evolving processes that can be worsened significantly by further insults over the following period. Things like low oxygen or not enough blood flow to tissues or too much pressure in the head.

So when we are looking after these patients you can probably imagine we’d like to make sure we give the brain all the things it wants to start healing (or at least not keep getting worse) as quickly as possible. We put as many monitors as we can on to do this. Strangely, this doesn’t include anything to look at the bit we care so much about – the brain.


Before My Time

Back in 1977, a researcher by the name of Franz Jöbsis described a technique where you could shine a light through brain tissue, look at the light that made it out the other side and figure out stuff about the levels of oxygen and metabolism happening deep in that brain tissue. This was the start of tissue spectroscopy.

Sounds like the perfect fit for the problem, right? So why isn’t this standard 38 years later? Well actually there’s all sorts of challenges that held it up.

For starters, Jöbsis first tried it out on cats. Cat lovers might tell you many things about the miraculous brilliance of those brains but it is fair to say they are smaller than humans. So in humans what we tend to do is shine light into the brain tissue and pick up what bounces back. But it doesn’t all bounce back.

Picture it like this. You give a big group of squirrels a whole lot of speed (well this analogy is struggling already). Then you release them into the woods and tell them to come back and tell you what they saw. Some will just head in and come pretty much straight back and tell you their story. But some of them run and never stop running (well just imagine how twitchy a squirrel on speed would be). And some bounce off 10, 20 or 30 trees in all directions before they finally stumble back with their own unique story about the woods. Now you have to put all the stories together and say something sensible about the bit between the trees. Messy.

Is he pre or post stimulants? Who can tell? [Photo by Corey Seeman via flickr and shared unchanged under CC 2.0]

Is he pre or post stimulants? Who can tell? [Photo by Corey Seeman via flickr and shared unchanged under CC 2.0]

And that’s part of the reason it has taken so long to figure out what to do with that technique from Jöbsis. To get to here we’ve ended up using devices that use particular different wavelengths of light in the near-infrared range which have been tested under different conditions so we hopefully know a bit about how the light is absorbed and reflected in the tissues (how many squirrels come back and how many run). Most systems then display a number between 0% and 100% which is supposed to tell you a bit about the oxygen delivery and use under the sensor.

But even that isn’t that simple. There are different manufacturers, and each one tries to figure out how best to do it slightly differently so you can’t really directly compare any of them. And they won’t tell you exactly how they do it.

And the number between 0 and 100%? Well it’s not really measured against a gold standard. They make some assumptions there too. Plus you’re only sampling a small area so what about the rest of the brain?



Trying Things

So why bother? Well we’re going to try something slightly different and see how we go. For starters, we won’t be using the sensors on a single site, but a couple of sites on the head and a comparison sensor checking circulation in the body to try and put more information together.

More importantly we’re not just going with the number. Part of the analysis will look at the number that’s all about oxygen stuff, as well as another one that’s about how much blood is in the area. But it will look at the patterns in how these change in all 3 spots over time and compare them to the other observations we already take. Because if there’s hope for this monitoring to show something new, it would seem like the best bet is to think that it might pick up a change as it is happening, rather than relying on a single number not tested against any real gold standard.

And that might work. Or it might not and that would be absolutely fine. There are already companies out there telling people these monitors add vast amounts of knowledge to managing a brain injury. That’s not really true. But they have to sell units and saved brains sound like a pretty good story.

Of course it might just work and tell us new things about what is happening in the brain in real time. Which would be fairly sensational since we could then start figuring out how to treat patients using that information to hopefully stop all those evolving injuries. And less brain injuries would mean more people getting back to the lives they planned.

Or we might find something cool that’s unexpected and that would be a bonus. It’s just as likely though that the story from the speed-addled squirrels will be pretty confusing and we’ll find it’s not useful. Which would also be a great result since ruling something out of the calculations still brings us closer to finding things that work and not exposing patients to things that don’t.

So now there’s one more challenge if I’m to use this post to inspire me back into it. We don’t have squirrels in Australia.


3 thoughts on “Doing Research That Might Not Work

  1. That’s ok Andy. We mightn’t have squirrels, but we have quolls, numbats, dunnarts, antechinus, phascogales… Or even frill-necked lizards if our mammals don’t do it for you.
    Glad it’s just an analogy!

  2. Pingback: The Number | The Flying PhD

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s