A senior physiotherapist’s perspective for residential aged care management.
Just because a resident can walk in some fashion, does that mean they should?
Few clinical decisions in residential aged care generate more discomfort than the one this article addresses: at what point should a resident stop walking. Not “stop walking forever,” and not as a punitive or convenience-driven restriction — but the clinical judgement that continued unsupervised or under-supported ambulation has become a greater risk to the resident than the loss of that specific activity.
This is uncomfortable territory, and it should be. It sits close to autonomy, identity and dignity. But discomfort is not a reason to avoid the clinical question, and avoiding it has a cost: residents sustaining serious injury from falls that a structured, evidence-based assessment would have flagged well in advance.
Capacity is not the same question as safety
The starting clinical distinction is this: can this person produce stepping movements, and should this person be walking independently or with the current level of support, are two entirely different questions — answered by two entirely different parts of the nervous system.
This isn’t a rhetorical point. It’s neurophysiology.
A newborn infant, held upright with their feet touching a surface, will produce a coordinated, alternating stepping pattern — heel strike, stance, swing, repeat. This is the primitive stepping reflex, and it is not a cortical achievement. It is generated by spinal locomotor circuits — central pattern generators (CPGs) — that produce rhythmic, alternating limb movement with minimal input from the brain above the brainstem. The evidence for this is unambiguous: infants without a functioning cerebral cortex still produce a stepping pattern when supported, and decades of spinal cord injury research have confirmed that isolated spinal circuits, given appropriate afferent input (such as body-weight-supported treadmill loading), can generate rhythmic stepping in the complete absence of any voluntary or cortical drive.
No competent parent responds to this reflex by letting go. Every parent understands, correctly, that the presence of a stepping pattern says nothing about the infant’s ability to bear weight independently, to correct for a perturbation, or to protect themselves from a fall. That capacity — postural control — develops separately, and later, through an entirely different set of neural systems: vestibular integration, visual-spatial processing, proprioceptive feedback, and the cortical and subcortical structures that generate anticipatory postural adjustments before a threat to balance even occurs.
This is precisely why human motor development follows a fixed, cephalocaudal and proximal-to-distal sequence: head and trunk control, then rolling, then sitting with postural support, then crawling, then pulling to stand, then cruising, and only then, independent walking. Each stage depends on the postural and balance competencies established in the one before it. A child does not walk before they can stand; they do not stand before they can sit with control; they do not sit with control before they have trunk and head stability. The sequence is not arbitrary — it reflects the order in which the underlying neural control systems mature.
The same sequence, running in reverse
This developmental sequence matters to your falls program because, in progressive neurocognitive and neurodegenerative disease, there is substantial clinical evidence that functional decline mirrors this sequence — in reverse.
The most rigorously documented example is Reisberg’s retrogenesis model in Alzheimer’s disease, operationalised through the Functional Assessment Staging Tool (FAST) — a 7-stage, 16-substage clinical instrument with well-established validity and reliability. In the terminal stages of the FAST, the sequence of functional loss is specific and ordered: stage 7c is loss of the ability to walk without personal assistance; stage 7d, immediately following, is loss of the ability to sit up without lateral support; stage 7e is loss of the ability to smile; stage 7f is loss of the ability to hold the head up independently. This is developmental acquisition, unwinding in precisely the reverse order in which it was built. Independent studies applying the FAST have found a statistically significant correlation between FAST stage and estimated developmental age (Spearman’s r = −0.85, p<0.01), lending quantitative support to a model that might otherwise sound purely descriptive.
The clinical implication is direct: loss of independent ambulation is not a random, late symptom of severe dementia. It is a specific, staged marker in a validated clinical tool, occurring at a defined point in a well-documented sequence, and the loss of safe sitting balance follows shortly after it.
If your facility is relying on informal observation rather than a structured functional stage marker to make this call, you are working with less clinical precision than already exists, freely available, in the literature.
Parkinson’s disease provides a second, distinct evidence base for the same underlying principle: gait persists, in some form, well after the postural control required to make it safe has been lost. The Hoehn and Yahr staging system marks the clinically significant transition from Stage 2 to Stage 3 specifically at the point postural instability emerges — evidenced by an abnormal retropulsion (pull) test, where the person cannot recover balance independently following a backward perturbation and requires assistance to avoid falling. Biomechanical studies confirm this isn’t subtle: the force required to destabilise a person with Parkinson’s disease to their limit of stability is measurably and progressively reduced from the earliest stages of the disease, even before postural instability is clinically obvious. Add freezing of gait — a phenomenon largely independent of overall disease severity that can appear without warning mid-stride — and the picture is clear: in Parkinson’s disease, the motor pattern for walking frequently outlasts the postural and executive control systems needed to walk safely.
The clinical marker I use as a gatekeeper
Across both presentations — and in general age-related mobility decline without a specific neurological diagnosis — I use one functional marker as a non-negotiable gatekeeper in my own clinical reasoning: the ability to sit-to-stand safely, even with physical assistance.
Sit-to-stand is not a peripheral strength test. It requires anticipatory postural adjustment before the movement begins, dynamic weight transfer through a rapidly changing base of support, and integration of visual, vestibular and proprioceptive input in real time — the same systems required to walk safely, in a lower-risk, more controllable clinical setting. If a resident cannot complete this transfer safely with hands-on assistance from a trained carer, there is no clinically coherent basis for permitting a higher-demand, higher-consequence task — independent or supported walking — to continue. This isn’t a personal rule of thumb; it reflects the same principle underlying validated fall-risk instruments such as the Tinetti Performance-Oriented Mobility Assessment (POMA), where a low score is a strong, independent predictor of falls. In one multi-facility nursing home cohort study, a POMA score of 18 or below carried a hazard ratio of 2.13 (95% CI 1.61–2.81) for time to first fall — a stronger predictor than age or comorbidity.
Where cognition changes the calculation entirely
Everything above concerns motor and postural systems. Cognition is the variable that determines whether a resident can compensate for motor decline, and its absence is frequently the deciding factor.
Safe ambulation in the presence of emerging postural instability depends on judgement: recognising fatigue, requesting assistance, avoiding hazardous transfers, following a redirection from staff, remembering to use a prescribed aid. A resident with intact insight and executive function can partially offset a genuine physical deficit through appropriate care-seeking behaviour. A resident with moderate-to-severe cognitive impairment cannot reliably do this — not through non-compliance, but because the neural systems required for that judgement are precisely the ones dementia and delirium compromise first. This is why cognitive status is now listed as a core domain — not a peripheral consideration — in the Australian Commission on Safety and Quality in Health Care’s 2025 Falls Guidelines for Residential Aged Care Services. A resident with borderline motor function and intact cognition, and a resident with identical motor function and moderate dementia, are not the same clinical risk. Treating them as though they are is where I most often see this decision go wrong.
This is not about restricting rights — and it must never be treated that way
I want to be direct about what this article is not proposing. Ceasing a resident’s independent or under-supervised walking is not the same as restraint, and it must never be implemented as one. Under the Aged Care Quality Standards and the Serious Incident Response Scheme, a restrictive practice is any intervention that limits a resident’s freedom of movement, and it may only be used as a last resort, in the least restrictive form, for the shortest necessary time, with informed consent from the resident or their representative, and only after alternatives have been trialled and documented.
A properly conducted clinical recommendation is different in kind, not just in degree. It is: based on a structured, validated assessment; discussed with the resident and family wherever cognition allows meaningful participation; framed around dignity of risk rather than risk elimination; and — critically — focused on how the resident continues to mobilise safely, not simply on stopping them. Ceasing independent walking might mean transitioning to assisted walking with two staff, a gait aid with supervision, a wheelchair for longer distances with standing transfers preserved, or a structured standing program to maintain what function remains. It is a change in the mode and support level of mobility, not the abolition of it — and it should be documented and reviewed with the same rigour as any other clinical intervention, not left as an informal, unrecorded judgement call made differently by every shift.
The question for your organisation
If your facility relies on a physiotherapy service, ask this directly: is this decision currently being made against a structured functional marker — sit-to-stand capacity, POMA or Berg Balance Scale score, FAST stage, Hoehn and Yahr stage, cognitive status — reviewed on a defined schedule and documented as part of the resident’s clinical record? Or is it, in practice, being made only retrospectively, after a serious fall has already occurred, or arbitrarily, based on which staff member is rostered on and how much time they have?
Both failure modes carry real cost — one clinical, one ethical. The evidence to make this decision well already exists. The question is whether your current provider is actually using it.
Sources
Sources referenced include Reisberg’s retrogenesis model and the Functional Assessment Staging Tool (FAST) for Alzheimer’s disease, foundational and contemporary central pattern generator research in human and animal locomotion (Forssberg; Yang; Minassian et al.), Hoehn and Yahr staging and retropulsion/pull test biomechanics research in Parkinson’s disease, a multi-facility nursing home cohort study of the Tinetti Performance-Oriented Mobility Assessment as a falls predictor, the Australian Commission on Safety and Quality in Health Care’s 2025 Falls Guidelines for Australian Residential Aged Care Services, and Australian Government guidance on restrictive practices in aged care under the Aged Care Act and Serious Incident Response Scheme.
This article is general clinical information for operational planning purposes and does not replace individual clinical assessment, multidisciplinary case review, or facility-specific policy.
