Balance Training That Actually Prevents Falls

Falls are the thing that turns an independent older person into a dependent one, and exercise is one of the few interventions with high-certainty evidence behind it. The catch is that it has to be the right exercise, at a dose most programs never reach.

What the evidence says, in numbers#

The Cochrane review pooled 108 randomised trials across 25 countries, 23,407 community-dwelling older people with a mean age of 76.

Two things about that table. The first two rows are as good as evidence gets in this field. The fracture row is rated low certainty and its confidence interval nearly touches no effect, so it should not be quoted with the same confidence as the rest.

Which kind of exercise#

This is where most advice goes wrong.

Balance and functional exercise means work that deliberately makes standing harder: feet together, then in line, then on one leg; stepping over things; turning while walking; standing up from a chair without hands. It is unglamorous and it is the category with the best evidence in the review.

The walking result deserves its own paragraph, because "just get out and walk more" is the standard advice given to older people who are unsteady. Across the trials, walking programs did not reduce falls.

In a trial of 165 women recruited after an arm fracture, a progressive brisk walking program was associated with an excess of about 15 falls per 100 person-years compared with controls doing upper limb exercises. The reviewers' conclusion is unusually direct: walking should not be prescribed as a single falls prevention intervention, and high-risk older people should not be told to walk briskly.

Walking remains good for the heart, the blood pressure and the mood. It is not a falls intervention, and for someone already unsteady, more time walking at speed is more exposure to the thing that causes falls.

The dose is the whole game#

A meta-regression of the trials found two variables that between them explained 76 percent of the difference in results.

Programs delivering two hours a week or more showed no effect at all. Programs delivering three hours a week or more cut falls by 23 percent. Add a genuine challenge to balance, meaning exercises that actually threaten stability rather than gentle seated movement, and the reduction reached 39 percent. Programs with neither managed 10 percent.

Three hours a week sounds like a lot until you realise it can be twenty-five minutes a day, at home, in bare feet, holding a worktop. The Otago program splits it across the day deliberately.

One honest caveat: these are comparisons across trials rather than a trial that randomised people to different doses, so the usual confounding applies. Programs that run longer differ from programs that run shorter in other ways too.

What a real program looks like#

The Otago Exercise Programme is the best-documented version. Five lower limb strength exercises done with adjustable ankle weights: seated knee extension, knee flexion, hip abduction, calf raises, toe raises. Ten repetitions each, with the weight going up once two sets of ten are comfortable. People over 80 typically start with one or two kilos.

Then twelve balance exercises, progressed through four levels from two hands on support to nothing: knee bends, walking backwards, walking and turning in a figure of eight, sideways walking, standing with feet in line, walking heel to toe, standing on one leg, walking on heels, walking on toes, standing up from a chair, and stairs.

Strength three times a week, balance at least three times, about thirty minutes each and splittable across the day.

Pooled across seven trials and 1,503 people with a mean age of 81.6, it cut falls by 32 percent. Notably, it works better with age and frailty, not worse: the original trials found clear reductions in the over-80s, and mixed results in the 65 to 79 group.

The adherence figure is the honest part. At twelve months, only 36.7 percent were still exercising three times a week. Whatever program you pick, the version you will still be doing next spring beats the better one you will abandon in March.

The things that do not work on their own#

Vitamin D. The US Preventive Services Task Force recommends against supplementing to prevent falls in community-dwelling adults, and in 2024 broadened that to all doses and to men aged 60 and over. A trial of high monthly doses found more falls in the supplemented group. This does conflict with the CDC falls program, which still advises supplementing where someone is deficient, and that disagreement is worth knowing rather than glossing. Correcting a real deficiency is worth doing for bone and muscle. It is not a falls plan.

Medication review alone. Pooling the trials, medication review and deprescribing as a single intervention produced no significant reduction in falls in any setting, including a relative risk of 1.05 for the number of people falling in community settings. This is surprising, because sedatives, blood pressure medicines and drugs that cause dizziness plainly do cause falls. The finding is not that medicines do not matter; it is that a review on its own, without the rest, has not been shown to change the outcome. It belongs inside a broader assessment.

What a proper assessment includes#

If you or someone you look after has fallen in the past year, feels unsteady, or worries about falling, the assessment worth asking for covers more than exercise:

• Blood pressure lying and standing. A drop of 20 mmHg systolic or 10 mmHg diastolic on standing is the threshold.
• A timed up and go test. Twelve seconds or more to stand, walk three metres, turn and sit is the cutoff for increased risk.
• A four-stage balance test. Being unable to hold a full tandem stance for ten seconds flags risk.
• Vision. Worse than 20/40, or no eye test in over a year.
• Feet and footwear, home hazards, and a full medicine list including anything bought over the counter.

Where to start today#

Stand on one leg while the kettle boils. Hold a worktop, and take a finger off it when you can. Then try it with your eyes shut, which is the version that actually discriminates at any age under sixty: the average drops from about fifteen seconds in your thirties to under five past sixty.

Balance responds faster than almost anything else trained. Across roughly a hundred trials the gains are consistent, the usual course is about twelve weeks, and one small study found most of the resistance to a stumble appeared after a single session.

Test where you are with the one-leg balance test, and try getting off the floor while you are at it.