Perturbation training

Perturbation training means practising controlled slips and trips, under supervision, until catching yourself becomes automatic. It reliably cuts falls provoked in the lab and may cut real-world falls by nearly a quarter. But that evidence is low certainty, on its own it did not significantly cut real-world falls, the case for adding it to ordinary exercise rests on one study, and its safety has barely been measured.

If a supervised slip-and-trip programme with a safety harness is available to you, use it as an add-on to balance and strength work, not a replacement. At home, borrow the idea: practise quick recovery steps and reaches, not only still stances.

What it is

Most real falls come from an unexpected slip or trip. Ordinary balance practice mainly trains the slow, anticipated control of holding a position, not the fast reflex step that stops a fall already under way.

Perturbation training (also called perturbation-based balance training) targets that gap. You walk or stand on a treadmill, a moving platform or a walkway while the equipment, or a trainer, repeatedly throws you off balance without warning, until the recovery response becomes automatic. It is done in a lab or clinic, with supervision and safety rigging.

Strong in the lab, more modest in daily life

The main evidence is a review of 25 randomised trials in about 2,660 older adults living at home.^[1]

Falls provoked in the lab dropped by 58%, from nine of the trials. This is the reactive skill the training targets.
Real-world falls dropped by a more modest 23%, from 11 trials in about 1,230 people.
Falls causing injury dropped by about 24%, from only four studies.

The review rated the real-world and injury results as low-certainty evidence. They are also fragile: leaving out any one of six trials made the real-world result lose statistical significance. The lab result showed signs of publication bias, meaning missing small null studies, and was no longer significant once three outlying studies were removed.

An add-on, not a replacement

On its own, perturbation training did not significantly reduce real-world falls. It looked much stronger combined with ordinary exercise, but that comparison rests on a single study. Against no exercise at all it cut falls by 45%; against exercise without perturbations, by only 13%, which was not significant. The authors' own reading is that it complements conventional exercise rather than substituting for it.^[2]

It is also narrow. It did not improve walking speed, step length, standing balance, a timed get-up-and-walk test or fear of falling. Exercise in general does ease fear of falling by a small-to-moderate amount, at least in the short term, on low-certainty evidence, without increasing falls.^[3]

Dose and method are unsettled

Most programmes were short, averaging 1.6 hours over three to six sessions. Only programmes of more than six hours significantly reduced falls, by about a third, though the formal comparison between short and long programmes was not statistically conclusive.

By method, only walkway-based training significantly reduced falls; differences from treadmill or hands-on methods remained unclear. Whether skills learned on a treadmill carry over to slips on real ground is unsettled, with trials pointing both ways.^[4]

The most striking single trial

In a trial of 212 adults aged 65 and over, one group had a single lab session of 24 unannounced slips while walking; the control group had just one slip in the same setting. Over the next year, 15% of the trained group had a real-world fall, against 32% of the controls. The benefit was most visible in people who had fallen before.^[5]

The trial had no group that was never slipped, so it cannot say what slip-naive people would have done. It is also one of the six trials the review's real-world result depends on.

Safety: largely unmeasured

Only 8 of the 25 trials reported adverse events at all. That is missing safety data, not reassuring safety data. Being repeatedly knocked off balance is the point of the training, which is why it belongs in a supervised setting with a harness, not in your living room.

What to do

Build the base first: progressive balance work and strength training, as set out in Training mobility and balance.
If you are at risk of falling and a supervised programme is available, add it on top, knowing that the extra benefit over exercise alone was small and not significant in the review. Longer programmes (more than six hours in total) looked better than the typical one or two hours.
At home, don't stage your own slips. Practise fast recovery steps and reaches near something solid, and consider step training, which has its own trial evidence.
Don't expect it to improve walking speed or fear of falling. For fear of falling, ordinary exercise is the better-tested route.