Rehabilitation robots, first introduced in the 1990s, are just what they sound like – robotics used to aid in regaining function through rehabilitation following an injury. The idea sounds compelling, and the technology has been advancing steadily. But still we have to ask ourselves the question – do they actually help, and what is the evidence?
A recent comprehensive meta-analysis and systematic review is likely to spark some discussion and even controversy over this question. The review looked specifically at the use of upper extremity (UE) robotics for post-stroke rehabilitation. Their overall conclusion was this:
“The small significant effects found at the level of motor impairment do not show generalization to clinically meaningful effects at the level of UL-capacity. Meta-regressions suggest that selected participants with some potential of UL-recovery may benefit most from UL-RT, especially earlier poststroke. The robustness and consistency of our findings suggest that the development of the next generation of UL-RT needs to be guided by a better mechanistic understanding about assumed underlying interaction effects between motor learning and motor recovery poststroke.”
They did find statistically significant improvement in some aspects of UE function correlating with the use of rehabilitation robots, but not in other measures of function. But critically the improvements were small (3% in strength, for example) and did not translate into a clinically meaningful difference for the patient. They left the door open, however, for the most favorable cases if robotic intervention is early enough. They also point out (elsewhere, not in that quote) that robotics that use brain-machine interface appear more promising.
Study leader, Gert Kwakkel, is also quoted in the press release as saying:
“Our findings provide robust evidence. This evidence is independent of the type of robot used, the number of weights the robot checks, the cost, and additional features – such as virtual game displays – that have been added to the robot arm. With the data of more than 4,000 patients and thousands of treatments, we would have been able to demonstrate the presence of a positive effect of the arm robots. The fact that this did not happen suggests that the arm robots are more of a hype than an actual solution.”
But in the article and in interviews the researchers are not all negative. The study, they emphasize, does not mean the technology is dead or useless. Rather, the modest results likely reflect the fact that we still do not understand enough about the underlying mechanisms of brain recovery following a stroke, and the rehabilitation factors that enhance recovery.
As a neurologist who has treated many patients and who has followed the rehab literature for decades I have another thought as well. It’s possible that there is simply not that much room for improvement. Specifically, following a stroke for example, brain recovery of function is going to happen mostly as a function of time. Age and health matter, but these are variables that cannot be manipulated by rehab. It does seem to matter that patients engage in some physical and mental activity, and not be sedentary. But if you look at the totality of evidence it is reasonable to conclude that the benefits of activity peak quickly, with further or more specific activity having little room for improvement. In other words, neurological recovery is going to happen as long as there is a minimum of activity, and there’s not much we can do to enhance that.
Perhaps, as the authors suggest, if we understood the actual mechanisms of recovery at a reductionist level we could intervene in a clinically significant positive way, but we are not there yet. Perhaps as robotics improves further, with greater brain-machine interface, we can make recovery quicker and move complete. The authors are not suggesting robotic rehab is a dead-end, rather that we have to realize we are not there yet. We have to think carefully about the path forward, and a lot more research has to be done.
Proponents of robotic rehab point out that there are other advantages to consider. Robotic rehab is often used as an extender of the physical therapist, allowing for far greater rehab time and repetition of activity. Therefore, there may not be much difference between ideal rehab with and without robotics, but robotics allows for more people to get sufficient rehab to maximize recovery. It also saves the physical therapists from extended tedious activity, which may have its own risks, such as repetitive motion injury. Further, the incremental improvements (while not clinically significant) are there, providing a proof-of-concept that we can build on for the future of robotic rehab.
These are all good reasons to continue with the research and development of robotic rehab, and to use them in the appropriate setting. I do think, however, we need to be realistic about the potential for improvement. There may be inherent biological limits on the speed and amount of neurological recovery possible from certain kinds of injuries, and no amount of rehab is going to change those limits.
Unfortunately, we also have to consider the cost effectiveness of such high-tech interventions. Using advanced expensive technology for small incremental clinical improvements may simply not be sustainable, given the challenges of rising health care costs. So future research and development must also consider the cost effectiveness of robotic interventions as well.
Having said all that, and in the context of this recent research, I remain cautiously optimistic for this technology. This is mainly because the underlying technologies – robotics, brain-machine interface, and artificial intelligence – are advancing rapidly and have a tremendous amount of room for improvement. The relevance of past research, therefore may rapidly become obsolete. This creates its own challenge, however, that we have seen in other areas of medicine involving technology – incremental technological advances constantly stay ahead of the research (which takes years to conduct), providing an endless cycle of excuses for past failures and hype over future potential. At some point we need evidence that we are on the right track and that the technology is worth it.
As I often do, I have to point out as a side issue (but one core to science-based medicine) that I think all this hand-wringing about the evidence is healthy and is exactly what we should be doing. But I am often struck by the stark contrast between how we treat evidence in mainstream medicine and how it is treated in the world of “alternative” medicine. If robotic rehab were “alternative” the small incremental improvements would be hailed as proof that the technology not only works, but is entirely worth it. Further, it would be used to justify extreme speculation about underlying mechanisms of action, without any independent evidence. Researchers would have already shifted to studies about how best to implement their treatment, without directly addressing the question of whether or not it actually works.
Alternative medicine proponents have certainly used much weaker evidence to promote much more speculative and even fanciful and “magical” treatments, and never seem to do any hand-wringing about the quality of their evidence or the limitations of their basic science understanding. It’s a different world, with a different culture, and a different set of standards and rules. The alternative medicine world is one we should reject, because it doesn’t work, and because it puts the needs of gurus, manufacturers, and practitioners above patients and the public.
With all the limitations I outlined above, robotic rehabilitation is a plausible technology that does seem to work and provide some incremental improvement. But we still need to determine if it is ultimately worth it, or can be improved enough or better targeted to be worth it.
