Recovering From a Broken Hip

Right now, sit at the front of a firm chair with your arms folded, and stand up and sit back down five times as fast as you safely can. Time it. That one number tracks the leg strength that decides whether someone walks independently again, and it is exactly what a good rehabilitation programme is supposed to move.

The Verdict

The surgery fixed the bone. What decides the next year is the muscle you rebuild afterwards.

Think of the hip like a shelf bracket that snapped off a wall. The surgeon puts the bracket back, and it holds weight straight away, so the metal is not what is stopping you. What is stopping you is that the wall around it crumbled. Weeks of pain, surgery and lying still strip strength out of legs that had very little spare to begin with. The bracket is fixed on day one. Rebuilding the wall is a training job, and it takes months.

What this actually is: the bone is repaired and able to take weight within a day or two, so what holds recovery back is lost leg strength and balance, not unhealed bone.
What most people get wrong: doing the same exercises for twice as long. Sixty minutes a day of standing exercise was tested head to head against thirty minutes and produced no extra strength and no extra walking speed. Making the exercises harder is what worked.
The one change that matters: keep going after the physical therapy appointments stop. Two separate trials, one of them starting four months after surgery, found people were still improving, and that is exactly the point where most people are discharged to nothing.

Best for

Someone recovering from hip fracture surgery who has been cleared to put weight through the leg, and the family or coach helping them do it.

Skip if

You have new hip pain and cannot weight-bear, a hot or leaking wound, sudden confusion, or a painful swollen calf. Those need same-day medical assessment, not exercise.

Want the full evidence? Keep scrolling

What Works

Cinematic rendering of lower limb musculature under load

1. Progressive strength training, as the core of the programmeStrong

Not an add-on and not a late addition. Four independent evidence reviews across nine years agree on this. The clearest single analysis pooled 13 trials and 1,903 people: exercise overall gave a small effect, and adding progressive resistance shifted that effect by 0.58 and explained 60% of the difference between trials. In a separate review, resistance work gave the largest effect of any component while aerobic exercise on its own gave nothing measurable.

Evidence: STRONG for the direction. The successful trials used loads around 60 to 80% of a person's maximum, but no trial has ever compared one load against another, so treat that as the tested territory rather than a proven number.

2. Balance training, programmed separatelyStrong

Eight trials and 752 people. Clear improvements in balance, walking, leg strength, daily activities and quality of life. Do not assume the strength work covers it.

Evidence: STRONG. Expect changes over about three months.

3. Carrying on after formal rehabilitation endsStrong

The most under-delivered treatment on this page. One trial gave people simple home exercises after their formal rehabilitation had already finished and they kept improving for six months, with the gain still there at nine. Another started a structured walking and balance programme four months after surgery and improved walking speed at no extra cost to the health service, still measurable at twelve months.

Evidence: STRONG, from two independent trials at two different time points. The researchers on the first describe the size of the improvement as modest, and that caveat travels with it.

4. Getting up and moving on day oneModerate

Unless the surgeon has restricted weight-bearing. It is safe, it is what every guideline recommends, and intensified daily physical therapy in hospital caused no adverse events in the trial that measured it.

Evidence: STRONG for function, WEAK for the dramatic survival figures often quoted. See The Debate below.

Tier 2 and Tier 3 — moderate and emerging

Weight-bearing exercise once past the first fortnightModerate
Four months after usual care finished, the weight-bearing group improved balance and function by 30 to 40% more than the alternative. In the first two weeks, though, it made no difference which you used.

Referral for osteoporosis assessmentModerate
81% of these patients met the criteria for bone-protection medication before the fracture and were not on it.

Referral for nutrition assessmentModerate
Ten trials, 1,119 people: nutritional therapy alongside rehabilitation reduced deaths. It did not significantly reduce complications, despite that review's own summary suggesting otherwise.

Home-based programmes where hospital rehabilitation is not neededModerate
Comparable to in-hospital rehabilitation for daily activities, and both beat doing nothing structured.

Remote or video-delivered rehabilitationEmerging
Seven trials. The effect on mobility is genuinely uncertain, and the researchers describe the effect on daily activities as too small to matter clinically. Possibly useful for confidence about falling. Not currently a substitute.

What doesn't work

Adding session minutes instead of adding difficulty. Sixty minutes a day of standing exercise was randomised against thirty minutes a day of seated exercise, in 160 people, and produced no difference in leg strength or walking speed. This is the clearest actionable contrast in the whole field.
Extending the outpatient block on the assumption that more weeks means more benefit. Twelve weeks was not better than six. Both groups improved by roughly three times the amount considered clinically meaningful.
Five sessions of anything. A trial adding strength training to hospital physical therapy averaged five completed sessions, and the main result was null. The lesson is not that early strength work fails, it is that a handful of sessions is not a dose.
Memory and thinking exercises as a route back into community life. Twenty trials: only physical retraining improved getting back out into the world.
Testosterone added to exercise in older women. A properly powered, blinded trial across eight sites found it added nothing to walking distance.

Exercise Prescription

The home programme below is drawn from what the successful trials actually prescribed. The progression rule matters more than any individual exercise.

Exercise	How	Sets × Reps	Frequency
Sit-to-stand	Firm chair, feet back under the knees. Stand up, sit down slowly over 3 seconds. Drop the hands as soon as you can	3 × 8-10	Daily
Step-ups	Bottom stair, holding the rail. Up with the operated leg first, down slowly	3 × 8 each leg	Daily
Standing hip abduction	Hold a worktop. Leg out to the side, toes pointing forward, no leaning the other way	3 × 10 each leg	Daily
Heel raises	Hold a worktop, rise onto the toes, lower slowly over 3 seconds	3 × 12	Daily
Standing balance	Feet together, holding a worktop. Progress to one finger, then no hands, then one foot slightly ahead	3 × 30 sec	Daily
Sideways and backwards walking	Along a worktop, 10 steps each direction	3 × 10 steps	Daily

The progression rule: every two weeks, make the exercises harder rather than longer. Add a hand weight or a loaded rucksack, use a lower chair or a higher step, take away hand support on the balance work. Effort in the thigh muscles is the point. Sharp hip pain is not.

Red Flags

Cinematic anatomical rendering of the hip and proximal femur

Any of these means stop the exercises and get medical assessment the same day.

New hip, groin or thigh pain with loss of the ability to put weight through the legEspecially after a period of improvement. Assume the fixation has failed or the bone has broken around it until an X-ray says otherwise.
Wound discharge, spreading redness, fever, or deep pain at rest and at nightPossible deep surgical or joint infection.
Sudden confusion, muddled thinking or unusual drowsinessDelirium. It is a medical emergency, it independently predicts worse recovery, and it is not the patient failing to try. In this population it rises from around 12% to 45% as cognitive impairment worsens.
One calf becoming painful and swollenPossible clot in the leg.
New pain somewhere else, especially the middle or lower backPossible second fragility fracture.
No bone-protection medication in placeNot an emergency, but the referral most often missed. Breaking a hip in a fall from standing height means osteoporosis is already diagnosed, and 81% of these patients were untreated before the fracture.

Refer to: the surgical team for pain and wound signs, the medical or older-adult team for confusion, and the family doctor or fracture liaison service for bone protection.

Return to Training

Concrete criteria, not "when it feels ready."

Cleared by the surgical team, with weight-bearing status confirmed and unrestricted
No red flag symptoms present
Walking indoors with the same aid used before the fracture, or none
Five sit-to-stands from a standard chair without pushing off with the hands
Knee and hip strength on the operated side within 10 to 15% of the other side
Six-minute walk distance improved by at least 55 metres from the starting point
A full flight of stairs, one foot per step, holding a rail
Osteoporosis assessment done or documented as declined
Falls risk addressed: medication review, home hazards, vision
For anyone returning to weight training: pain-free loading on both legs together before any single-leg or free-weight progression, and no flare 24 to 48 hours afterwards

Conviction

Moderate Stratified, because the different claims on this page are not equally well supported.

Progressive strength training is the active ingredientModerate to High
Four independent research teams over nine years, pointing the same way, with one analysis identifying it as the ingredient that explains the variation rather than merely pooling it.

Benefit continues after formal rehabilitation endsModerate
Two independent trials at two different time points, with effects still present at nine and twelve months.

Which setting or delivery model is bestLow
The only network analysis grades every comparison as very low certainty, with none reaching significance.

The optimal load, number of sets, or programme lengthLow
No trial has randomised any of them.

Early movement reduces death ratesLow as a cause-and-effect claim
A strong association from observational data, in which the people who get moving on day one are the people well enough to.

What would change my mind

On the strength training claim. A trial of at least 400 adults over 70, within six weeks of surgery, comparing a heavy version of the same programme against a light version, with function measured at twelve months. If the two arms came out the same, the active ingredient would turn out to be practising the movements rather than the load, and this page would change substantially.

Nothing like this exists. Every trial to date compares exercise against usual care or against a different kind of exercise, never against a different amount of load.

On who gets offered rehabilitation. The same trial, deliberately enrolling people with mild to moderate dementia. A positive result in that group would overturn the current practice of offering them less, and would make an overlooked finding from 2009 the most important result in this field.

Next Step

Someone you know will break a hip. When they do, the difference between a good year and a bad one is decided by whether anyone keeps the training going after the appointments stop. We publish free, evidence-checked protocols like this one every week.

Join The Verdict, free

The Full Picture — Anatomy, Diagnosis & Evidence

What's Actually Going On

Cinematic anatomical rendering of the proximal femur and hip joint

A fragility fracture of the femoral neck is a low-energy injury. A fall from standing height or less, onto bone that no longer tolerates it. That makes the fracture a symptom of two separate problems, and rehabilitation has to answer both: the bone was weak, and the person fell.

What the fracture and the operation then do is remove another slice of capacity from someone who had very little in reserve. The average age across these trials is 78 to 83. Surgery, pain, anaesthetic, days of reduced movement and often an episode of confusion all subtract from thigh strength, hip strength, balance and walking speed at the same time.

Almost nothing about the healing bone is the limiting step, because both of the common operations, pinning it or replacing the ball of the joint, are designed to take weight immediately. The useful way to hold this: the rehabilitation problem is a strength and balance problem wearing a fracture costume. Which is exactly why the treatment with the best evidence is not fracture-specific at all.

How to Identify It

Cinematic clinical rendering of hip and lower limb assessment

There is nothing to identify about the fracture itself. It arrives already X-rayed and already fixed. What a clinician is actually assessing is a patient who is not progressing.

On special tests, the honest answer is that there are none. All 124 research papers gathered for this page were searched for the standard measures of how well a physical test catches or rules out a condition. Zero appeared, because diagnosis here is done by X-ray before anyone reaches rehabilitation. An empty table would read like a page that got cut short, so this states the gap instead 0 tests found in 124 papers

What replaces testing is measurement, repeated over time:

Cumulated Ambulation Score, daily on the ward, targeting the level they had before the fracture
Timed Up and Go, and five sit-to-stands, every four weeks
Six-minute walk test, with 55 metres as the change that counts as real
Berg Balance Scale, at the start and at three months
A formal memory and thinking screen, on admission and whenever participation changes

The single most important thing to record is the one most often skipped: what could this person do before the fracture? Which walking aid, could they get outdoors, could they manage stairs? Everything else is scored against that, not against normal.

The differential worth keeping in mind is not for the fracture. It is for the patient who was improving and then stopped: failure of the metalwork, deep infection, delirium, a clot, or a second fracture at another site.

The Debate

Older position

Cochrane, 2011, 19 trials and 1,589 people: "insufficient evidence from randomised trials to establish the best strategies for enhancing mobility after hip fracture surgery." A 2010 Cochrane review said the same about psychological and social interventions.

Recent evidence

Four pooled analyses between 2016 and 2025 found consistent benefit from exercise, with progressive resistance identified as the ingredient that mattered, explaining 60% of the variation between trials.

Both are correct, because they asked different questions. Cochrane asked which strategy is best and found the trials too small and too varied to rank them. The later work asked whether exercise beats no exercise. "We cannot tell you which programme" and "exercise works" are perfectly compatible answers. Follow the later work on content, and Cochrane on humility.

The pooled analyses

Significant benefit from community rehabilitation programmes, with effect sizes in the small to moderate range.

2026 network analysis

Twelve trials, 1,510 people, and the first to formally grade certainty: no comparison between programmes reached significance, the network was substantially inconsistent, and certainty was rated very low for every primary comparison.

This is the ceiling on how confidently anyone can rank one rehabilitation model against another. It is not a refutation of exercise. It is a refusal to crown a winner.

The headline number

Early movement after hip fracture surgery is associated with far lower 30-day death rates, across 13 studies and 297,435 patients.

What that number is

Pooled observational comparisons, and the review states plainly that a direct cause-and-effect link "remains to be demonstrated." When the question is asked with randomised designs, the effects shrink dramatically, and older-adult co-management shows no significant reduction in one-year deaths.

The people who get up on day one are the people well enough to get up on day one. Mobilise early because it is safe, recommended and functionally useful. Do not present that survival figure as an effect of physical therapy, and do not convert it into a percentage.

Honest Limitations

The trials exclude the patients with the worst outlook

The finding: progressive strength training improves function after hip fracture.

The gap: one major trial enrolled only "cognitively intact" adults. Another required people to have been walking independently before the fracture. Meanwhile around 28% of a real caseload has dementia, and across that spectrum twelve-month deaths run from 21% to 54%, and the proportion still walking at six months falls from 84% to 42%. The evidence is built on the healthier half of the people who actually get this injury.

The adjustment: exclusion from a trial is not evidence of futility. The one signal pointing either way points the other way: in 2009, participants with impaired memory and thinking gained more from the higher-dose programme, not less. That was an after-the-fact look at a subgroup in a trial that found nothing on its main measures, so it is a hypothesis and nothing more. Treat it as a reason not to ration, not as a promise.

Sticking to the exercises is not the bottleneck

The finding: community programmes often underdeliver compared with trial results, and the intuitive explanation is that people do not do the exercises.

The gap: across 17 trials and 1,850 people, adherence pooled at 88%, which is high, and doing more of the programme was not linked to better outcomes. Programmes running longer than six months had worse adherence.

The adjustment: spend the effort on what is in the programme and whether it is getting harder, rather than on chasing compliance. Keep programmes under six months. This is a trial-level pattern, so it cannot say an individual's effort is irrelevant, but it does undercut the usual assumption.

The impressive numbers belong to the hospital, not the therapist

The finding: early movement and specialist older-adult care carry the largest effects in this field.

The gap: both are observational or non-randomised at their core. When restricted to randomised and prospective evidence, specialist co-management improves independent walking and shortens hospital stay by about 1.4 days, with no significant effect on one-year deaths.

The adjustment: argue for these pathways on function and length of stay, which the good evidence supports. Leave the survival claim alone.

The Nuance

Cinematic anatomical rendering of the hip region

Surgery versus no surgery. Surgery is the default and, unusually for this library, that is not controversial. Across 185 studies covering 10.76 million patients in 44 countries, 8.4% of hip fractures were managed without an operation, with wide variation that patient characteristics do not explain. Those patients had a higher risk of death at every time point, though the reviewers note the obvious confounder: not operating is often chosen for people who are dying or too unwell for anaesthetic.

The debate that dominates every other fracture is largely settled here. What is not settled, and what actually decides the patient's year, is what happens in the twelve weeks after the operation. That question has been asked properly, dozens of times, and the answer keeps coming back the same: load the muscle progressively, train balance, and do not stop when the funded block stops.

What recovery honestly looks like. At twelve months, in a trial that included people with impaired memory and thinking, 56% of the intervention group and 58% of the control group had regained or improved on their walking ability from before the fracture. A widely repeated claim that "up to 75% never recover" traces back to an unreferenced opening line in a paper describing a study's design, which reported no outcomes at all. The measured figure is the one worth quoting.

One number to be careful with. A commonly cited review of nutritional therapy alongside rehabilitation reports reductions in both deaths and complications. The deaths result holds. The complications result does not: its range of uncertainty crosses the point of no effect. That correction is made here rather than passed along.

Sources

Diong et al., 2016, British Journal of Sports Medicine. Meta-analysis with meta-regression, 13 trials, N=1,903. Structured exercise SMD 0.35 (95% CI 0.12-0.58); adding progressive resistance changes SMD by +0.58 (95% CI 0.17-0.98), adjusted R²=60%. PMID 26036676.
Latham et al., 2014, JAMA. Randomised trial, N=232. Home exercise after formal rehabilitation ended: function +0.8 points (95% CI 0.4-1.2), sustained at 9 months. PMID 24549550.
Taraldsen et al., 2019, PLoS One. Randomised trial, N=143 randomised at 4 months post-surgery. Gait speed +0.09 m/s (95% CI 0.04-0.14), still +0.07 m/s at 12 months, cost-neutral. PMID 31738792.
Zhang et al., 2022, Disability and Rehabilitation. Meta-analysis, 12 trials, N=1,198. Progressive resistance SMD 0.48 (95% CI 0.27-0.69); aerobic alone null. PMID 34101525.
Lin et al., 2025, Journal of Orthopaedic Science. Meta-analysis, 7 trials, N=665. High-intensity strength training beat general exercise on balance and 6-minute walk. PMID 39721850.
Lee et al., 2019, Journals of Gerontology Series A. Meta-analysis, 8 trials, N=752. Balance training: overall function SMD 0.390 (95% CI 0.114-0.667). PMID 30517613.
Moseley et al., 2009, Age and Ageing. Randomised trial, N=160. 60 min/day standing vs 30 min/day seated: no difference. Post hoc only: greater benefit in participants with cognitive impairment. PMID 19017676.
Overgaard et al., 2022, Journals of Gerontology Series A. Randomised trial, N=100. 12 weeks not superior to 6; both exceeded the 55 m threshold threefold. PMID 34460897.
Liang et al., 2026, BMC Geriatrics. Network meta-analysis, 12 trials, N=1,510. No comparison significant; GRADE certainty very low throughout. PMID 41486115.
Handoll et al., 2011, Cochrane Database of Systematic Reviews. 19 trials, N=1,589. Insufficient evidence to rank strategies. PMID 21412873.
Agarwal et al., 2024, Musculoskeletal Care. Meta-analysis, 13 studies, N=297,435. Early mobilisation 30-day mortality OR 0.35 (95% CI 0.31-0.41). Pooled observational; causal effect undemonstrated. PMID 38353351.
Tarazona-Santabalbina et al., 2015, Geriatrics & Gerontology International. Observational, N=1,258. Dementia 28.1%; 12-month mortality 21.2% to 53.5% by severity. PMID 25164866.
Karlsson et al., 2016, JAMDA. Randomised trial, N=205, included cognitively impaired participants. 56.3% and 57.7% regained pre-fracture walking at 12 months. PMID 26975205.
Yau et al., 2022, Archives of Physical Medicine and Rehabilitation. Meta-analysis, 17 trials, N=1,850. Adherence 0.88 (95% CI 0.78-0.95), not associated with functional improvement. PMID 35143745.
Winfield et al., 2025, Anaesthesia. Systematic review, 172 cohorts, N=10,763,994. Non-operative management 8.4% (95% CI 7.2-9.7). PMID 40855691.
Böck et al., 2026, Archives of Orthopaedic and Trauma Surgery. Observational, N=145. 81% met osteoporosis treatment criteria before the fracture and were untreated. PMID 42397594.
Takahashi et al., 2020, JAMDA. Meta-analysis, 10 trials, N=1,119. Nutritional therapy mortality RR 0.61 (95% CI 0.39-0.93); complications RR 0.67 (95% CI 0.44-1.03), not significant. PMID 32859299.
Binder et al., 2025, JAMA Network Open. Randomised trial, N=129 women. Testosterone added nothing to supervised exercise. PMID 40372752.

Full 38-source evidence trail with per-source annotation is held in the clinical record for this condition. Two authoritative guidelines, NICE CG124 and the 2021 American Physical Therapy Association guideline, are not indexed in the databases searched and did not appear in any sweep. No claim on this page rests on either, and their recommendations are independently supported by the sources above.