Stimulating spines

Spinal cord stimulators are small battery-powered devices that are implanted in the back and electrically stimulate the spinal cord to relieve chronic pain, often from failed spine surgery. These days they can be controlled remotely and can even do your step count. They are widely used and are very expensive. But we don’t know if they work or not – good evidence hasn’t been produced, and this problem doesn’t look like being corrected anytime soon. Why should it, when everybody is paying for the treatment now?

In theory (like with everything else that doesn’t work), these devices look promising. They send a signal to the spinal cord which is meant to modify the chronic pain signal coming from the back and legs, heading towards the brain. They have been used for decades and are a particular favourite of pain specialists who manage people with chronic back pain, often after failed spine surgery.

In older stimulators, the patients could feel the stimulation, but with modern (‘high frequency’) stimulators, the patient can’t feel if it is on or off. And because they can be switched on and off remotely, this seems like the perfect setting for a placebo-controlled trial where patients could act as their own controls. You could bring a patient back at regular intervals and switch the device on or off (according to a randomly generated schedule) and measure their pain during each interval. You would not need to recruit many patients to such a study to tell if the device worked or not. Yet for some reason, the proponents of these devices (including the manufacturers) don’t seem very keen to do this.

The evidence used to support spinal cord stimulation (SCS) comes from a handful of studies, all of which are ‘open label’ (unblinded) studies and are therefore at risk of bias. A recent review can be found here, and the two “high-quality” trials that they use to conclude that it works (at least after failed back surgery) are here and here. The first one compared stimulation to medical management (probably similar to what they had already tried) and was not blinded. The second one compared one (new, high frequency) type of stimulator to an older model – it was also not blinded. A third study often quoted is here. It is from 2005 and for a group of patients with failed back surgery, it compared spinal cord stimulation to (wait for it) more spine surgery! I think I could make a pretty good argument that placebo surgery would have been less harmful and just as effective as the repeat surgery.

As an aside, it is interesting how many people argue against the use of placebo surgery to determine effectiveness, yet are happy to commit patients to major surgery with all the attendant risks even when that surgery is not proven or may be harmful.

The reason why a high quality placebo study has not been done isn’t because it would be difficult – it would be one of the easiest procedural studies I could think of – it’s because it doesn’t have to be done: there is no incentive for anyone in this field of work to conduct such a trial (except to find out if the devices work or not). If the devices were shown to be no better than placebo, everybody in the industry would lose. Society and the patients may benefit from such a study, but the people who use it and sell it would lose. Money. The devices costs tens of thousands of dollars and the industry is worth billions annually.

The bottom line

This is another example of introducing a treatment based on theory and low quality observational evidence, having it approved and getting it established in practice before finding out if it really works or not, and then saying ‘it’s all too hard’ or ‘why rock the boat’ when someone asks for a properly conducted, blinded trial to determine effectiveness.

Addit (14 April 2018):
I have been made aware of 2 placebo studies using SCS, and found a third. I am drafting an update now and will post later today as a separate blog post (link here).