The falls benefit that didn’t last

LONGEVITY LATEST ISSUE 32 · 14 OCTOBER 2026

LONGEVITY LATEST

The Evidence-Based Edge on Living Longer and Better

Issue 32 · The Fall and the Floor · 14 October 2026

Welcome

For a few months, the exercise seemed to work. PreFIT enrolled 9,803 people over 70 at 63 English GP practices, and between months four and eight the exercise group fell 21% less than people sent advice by post. Over the full 18 months: no fewer falls, and no fewer fractures.

Delivery was thin. Only about one in five people in the exercise arm took up the classes, typically starting more than three months in. The trial can’t tell us whether the benefit faded because people stopped or because so few began. Either way, it sits awkwardly beside the best-evidenced finding in this field: balance training, done properly, cuts falls by about a quarter.

The short version. If you’re over 65, do exercise that challenges your balance on most days, building towards about three hours a week. If a fall has hurt you, or you’ve fallen more than once this year, ask for a falls assessment. And don’t buy a vibration plate for your balance. The plan is in Biohacking Corner; the grades follow.

SPOTLIGHT

Stopping a fall and stopping a fracture are different jobs

My reading of the middle column: what works is aimed at people likely to fall and asks something specific of them, whether that’s training, fixing the stairs or wearing the pad. What disappoints is broad and light-touch. The top right-hand cell is a finding rather than a reading: even exercise hasn’t shown at high certainty that it prevents fractures, mostly because trials are too small to count them.

Sherrington et al. Cochrane (2019) · Clemson et al. Cochrane (2023) · Santesso et al. Cochrane (2014) · Bhasin et al. STRIDE. NEJM (2020)

Aimed and done. The programmes with the best results reach people likely to fall, and those people actually take part.

THIS WEEK’S ANALYSIS

Top 3 Interventions Under the Microscope

The grades judge evidence that a strategy prevents falls in people living at home, not how sensible it sounds.

Falls-evidence grades: A = many randomised trials show fewer falls, high certainty; B = randomised trials show fewer falls, but fewer trials or lower certainty; C = early, mixed or null in the trials that matter; D = claims outrun the data.

1. Balance and strength training, at dose | Grade A

If you take one thing from this issue, take this, and keep at it. No other falls strategy has evidence this deep.

The evidence. The 2019 Cochrane review pooled 108 randomised trials, 23,407 people with an average age of 76. Exercise cut the rate of falls by 23% (rate ratio 0.77), with high certainty: about 195 fewer falls per 1,000 people a year. Balance and functional work, such as sit-to-stand, stepping and heel-to-toe walking, did most of it, and adding resistance training looked better still (0.66). Tai Chi earns its place too. In a 670-person Oregon trial, Tai Ji Quan: Moving for Better Balance produced 58% fewer falls than stretching and 31% fewer than a standard exercise class. Walking on its own? Uncertain.

The dose. A 2017 meta-analysis of 88 trials found that programmes which genuinely challenged balance, for more than three hours a week, cut falls by 39%. Gentle seated exercise is unlikely to be enough. A, firmly, for falls.

Sherrington et al. Cochrane (2019) · Sherrington et al. BJSM (2017) · Li et al. JAMA Intern Med (2018)

2. Slip-and-trip training | Grade C

Clever idea, wobbly evidence. Wait for a better trial before paying for it.

Most falls start with a trip or a slip, and what saves you is the reactive step, the quick unplanned stride that catches your weight. Perturbation training goes after that reflex directly. You walk on a treadmill or walkway in a safety harness while the machine trips or slips you without warning, again and again.

I was ready to give this a B. A 2025 meta-analysis, posted as a preprint, pooled 11 trials with falls data, 1,231 people: 23% fewer falls and 24% fewer injurious falls, both low certainty, with a clear effect only in programmes longer than six hours. Then came ReacStep, from Yoshiro Okubo’s Sydney lab, one of the field’s leaders: 88 people, six weekly sessions. Laboratory falls didn’t drop, and daily-life falls over the next year ran higher in the trained group (rate ratio 1.48), though not significantly.

Why C, not B. One small, non-significant, unreviewed trial can’t overturn pooled data, and ReacStep doesn’t. The grade rests on the pooled evidence itself: low certainty, inconsistent results across small trials, and a clear effect only in longer laboratory programmes few people can reach. A B would need a larger trial showing fewer falls in daily life. The first trial since the review didn’t supply it.

Don’t improvise a home version. The harness is there for a reason.

Perturbation training meta-analysis. medRxiv (2025) · Sharma et al. ReacStep. medRxiv (2026)

3. Falls assessment and referral | Grade C

Worth it for the right person, flat as a blanket programme. The value is in what happens after the assessment.

The trials. STRIDE put nurse falls-care managers into 86 US primary-care practices: 5,451 people aged 70 and over had their medicines, blood pressure, feet, vision, home and strength assessed, then got a tailored plan. Serious fall injuries came in at 4.9 against 5.3 per 100 person-years (hazard ratio 0.92), not a significant difference; self-reported fall injuries were modestly lower. Even deprescribing, the obvious fix, disappointed: in a 2,367-person trial more people came off sedating drugs, but medically treated falls didn’t drop (hazard ratio 1.11).

Where it works. Specific fixes for specific people. In a 306-person trial, women given their first cataract operation within weeks, rather than after a year’s wait, fell 34% less. Home hazard visits help people at higher risk, as the Spotlight shows. My reading: an assessment sorts people, it doesn’t treat them, and in the big programmes too little treatment followed. C as blanket policy. If you’ve had a fall that injured you, or more than one this year, guidelines say you should be offered one. Ask.

Bhasin et al. STRIDE. NEJM (2020) · Phelan et al. JAMA Netw Open (2024) · Harwood et al. Br J Ophthalmol (2005)

HYPE CHECK

“Ten minutes a day for better balance”

The claim. Vibration plates went viral as a lymphatic-drainage gadget, and the pitch is moving older. Roundups of the “best vibration plates for seniors” promise help with balance and bone, and one guide says they cut fall risk by about 30%. Consumer plates from Lifepro, Merach and Zelus cost $120 to $400.

At home, the evidence is thin but not empty. A 2017 BMJ Open meta-analysis found a third fewer falls (rate ratio 0.67), from four trials and 746 people who knew what they were getting. Fracture data came from one trial, with no clear effect, and bone density didn’t change overall. A June 2026 conference analysis, from a team including Cathie Sherrington, who led the Cochrane exercise review, saw a similar signal in three community trials.

In care homes, the same analysis found more falls with vibration than without: rate ratio 1.48, from two trials. The authors advise against it there.

In two care-home trials, the people on vibration platforms fell more often.

Why it misleads. The trials used clinical platforms on set protocols, sometimes alongside exercise; a $120 plate from an affiliate roundup is a different machine. The 30% rests on a handful of small, unblinded studies, while balance training has more than a hundred trials behind it. None of that proves vibration is useless. It means the evidence is too small, too mixed and too machine-specific to justify buying one for balance.

Jepsen et al. BMJ Open (2017) · Kwok et al. ANZSGM abstract (2026) · The Daily Beast. Vibration plate prices (2026)

Our verdict. Skip it as a falls-prevention tool. The researchers behind the 2026 analysis advise against vibration in care homes, and for people living at home they recommend an individual assessment and professional guidance first. If you already use one, that’s the bar: ask a physiotherapist whether it suits you. For balance itself, the Otago programme, a home routine tested in several trials, needs a kitchen chair.

IN BRIEF

Fewer falls on semaglutide, or healthier patients? A records study in Osteoporosis International this July matched adults over 65 with type 2 diabetes starting semaglutide or tirzepatide against similar people starting DPP-4 inhibitor tablets. Over a year, falls were about a third lower (hazard ratio 0.66) and femoral fractures roughly halved on either drug. A welcome counterweight to Issue 30’s bone-loss worries, but the authors list the catches themselves: residual confounding, better recording of falls in people who see doctors more, and smaller effects in their intention-to-treat analysis. A hypothesis, not a reason to start a drug.

Chen et al. Osteoporos Int (2026)

Village doctors as balance coaches. A useful contrast with PreFIT’s low uptake, from rural China. The FAMILY trial, published in JAMA last year, randomised 128 villages and 2,610 adults over 60. Village doctors ran monthly balance classes and encouraged home practice four times a week; class attendance topped 80%. Over 12 months, 29.7% fell, against 38.3% given education alone, and fall injuries dropped from 21.6% to 15.2%. The exercises weren’t unusual. The uptake was.

Peng et al. FAMILY. JAMA (2025)

THIS WEEK’S DEEP DIVE

It’s the Landing, Not Just the Fall

In a Cambridge study of people over 90, eight in ten of those who fell couldn’t get up unaided after at least one fall. Three in ten lay on the floor for an hour or more.

The companion is about what happens once you start to go down: which landings break hips, whether falling more safely can be learned, and how to get up, or when not to try. It opens with a two-minute balance check, with the age norms to judge your score.

Read the companion: It’s the Landing, Not Just the Fall →

BIOHACKING CORNER · THE PLAN LIVES HERE

Aim for exercise that genuinely tests your balance on most days, with strength work alongside. Three hours a week is where the biggest effects clustered in the 2017 meta-analysis, so treat it as a target, not an entry fee: in a 503-person trial, the StandingTall app cut injurious falls by 20% over two years at about two hours a week. Three routes: a Tai Chi or strength-and-balance class twice a week, plus short sessions at home; the Otago routine (sit-to-stand, heel-to-toe walking, one-leg stands holding the kitchen counter, step-ups) for about half an hour a day; or an app such as StandingTall.

Make it harder as it gets easier: feet closer together, eyes closed for a few seconds, a cushion underfoot, a head turn as you walk. If you never wobble, it probably isn’t challenging your balance.

My rule, and it’s a judgement rather than something any trial isolated: balance work goes in the diary like an appointment, because the programmes with the best results were the ones people kept attending. Unsteady, or fallen recently? Start supervised, and ask your GP about local strength-and-balance classes.

Delbaere et al. StandingTall. BMJ (2021)

READER PULSE

This week, one question: in a typical week, how much balance training do you do? None, under an hour, one to three hours, or more than three? If most of you say under an hour, the next home-exercise piece will be a step-by-step progression plan.

CLOSING

Pick the plan you’ll still be doing next year

Balance responds to practice. The best programme is the one you’ll still be doing next October, not the hardest one you can manage this month.

Next week: hearing. Hearing loss is linked to falls as well as to dementia, and the ACHIEVE trial tested whether hearing aids slow cognitive decline. The answer was more interesting than yes or no.

Stay curious and stay healthy!
Christian Thomsen, Editor

Longevity Latest is published weekly by FrontWave Media Ltd. Educational content, not personal medical advice. Don’t stop or change a prescribed medicine to reduce falls without your prescriber; some need tapering. If you’ve fallen recently, feel unsteady or have osteoporosis, start balance training with professional advice.

Sources and reading notes

Evidence checked through 9 October 2026. PreFIT uptake: 697 of 3,279 in the exercise arm, median start 14 weeks after randomisation. Fracture estimates for every strategy here are low or very low certainty. The perturbation meta-analysis and ReacStep are unreviewed preprints; the 2026 vibration analysis is a conference abstract; the GLP-1 study is observational. Prices are US, September 2026.

Sherrington et al. Exercise for preventing falls in older people living in the community. Cochrane Database Syst Rev (2019)

Sherrington et al. Exercise to prevent falls in older adults: updated systematic review and meta-analysis. BJSM (2017)

Li et al. Tai Ji Quan, multimodal exercise or stretching for fall prevention. JAMA Intern Med (2018)

Lamb et al. Screening and intervention to prevent falls and fractures in older people (PreFIT). NEJM (2020)

Bhasin et al. A randomized trial of a multifactorial strategy to prevent serious fall injuries (STRIDE). NEJM (2020)

Clemson et al. Environmental interventions for preventing falls in older people living in the community. Cochrane Database Syst Rev (2023)

Santesso et al. Hip protectors for preventing hip fractures in older people. Cochrane Database Syst Rev (2014)

Perturbation-based balance training reduces falls and fall injuries in older people: systematic review and meta-analysis. medRxiv preprint (2025)

Sharma et al. ReacStep reactive balance training: randomised controlled trial. medRxiv preprint (2026)

Phelan et al. Reducing CNS-active medications to prevent falls and injuries: cluster randomized trial. JAMA Netw Open (2024)

Harwood et al. Falls and health status after first eye cataract surgery: randomised controlled trial. Br J Ophthalmol (2005)

Jepsen et al. Whole-body vibration exercise in preventing falls and fractures: systematic review and meta-analysis. BMJ Open (2017)

Kwok et al. Should whole body vibration be used for falls prevention in older people? ANZSGM Annual Scientific Meeting abstract (2026)

The Daily Beast. Vibration plates for lymphatic drainage, product prices (updated September 2026)

Chen et al. Fall and femoral fracture risks with semaglutide and tirzepatide vs DPP-4 inhibitors. Osteoporos Int (2026)

Peng et al. A fall prevention program integrated in primary health care in rural China (FAMILY). JAMA (2025)

Delbaere et al. StandingTall home-based balance training: randomised controlled trial. BMJ (2021)

© 2026 FrontWave Media Ltd · Longevity Latest