Does Your Step Count Still Mean Something If You Have Parkinson's, a Stroke, or Another Movement Condition?
Step-count validation studies mostly recruit people who walk the way the algorithm expects. Here's what happens when they don't.
This piece covers what published research says about wearable step-count and activity-tracking accuracy specifically in people with Parkinson's disease, stroke or acquired brain injury, Huntington's disease, and myositis. It does not cover treatment or rehabilitation decisions, and it is not a substitute for a clinician's assessment of any individual's device data.
The accuracy picture is genuinely mixed and depends on which condition, which metric, and where the device sits. Step counting has held up reasonably well in Parkinson's disease and in myositis, with researchers even establishing a specific clinically meaningful step-count threshold for Parkinson's. Energy expenditure tracking has fared worse, particularly in Huntington's disease. And for stroke and traumatic brain injury, a systematic review found that accuracy and validity simply have not been reported for many of the available studies, which is a different problem than accuracy being confirmed or denied.
The validation studies most people rely on were not built around atypical gait
Most of what gets cited as "wearables are accurate for steps" comes from research on healthy adults walking normally, the same population the underlying step-counting algorithm was designed and tuned around. Parkinson's disease changes stride length and cadence. Stroke and traumatic brain injury can alter gait asymmetry entirely. Huntington's disease affects movement control in yet another way. None of that is a footnote to normal walking, it is a different signal for the same accelerometer to interpret.
So the real question isn't whether trackers are accurate in general. It's whether the accuracy already established in healthy populations actually transfers to gait patterns those validation studies never included.
2 studies
- Establishing the minimal clinically important difference for daily step count in people with mild-to-moderate Parkinson's disease, using a consumer smartwatch worn for five consecutive days: a change of roughly 581 steps per day, about 10 percent of the group's mean daily step count, was identified as the threshold for a clinically meaningful change.
- A systematic review of wearable devices for tracking physical activity in the community after stroke or traumatic brain injury found that device adoption for stroke has begun but is not widespread, and has not yet started at all for traumatic brain injury in community settings, with no consensus on a preferred device or wear location.
Where step counting has actually been validated in movement disorders
In Parkinson's disease specifically, step counting is better tested than most people would guess. An earlier comparison of the Fitbit Charge HR and Garmin vivosmart HR in people with mild-to-moderate Parkinson's found low error, under 3 percent, and moderate to high consistency against a research-grade accelerometer, both indoors and on outdoor terrain. That same study found something worth separating out clearly: while step count itself held up, the devices' ability to reflect exercise intensity did not. Cadence measured by either tracker only weakly tracked heart rate, and did not reflect oxygen consumption or perceived exertion at all.
A comparable pattern shows up in myositis. Fitbit step counts compared to a research-grade ActiGraph monitor showed strong test-retest reliability over a one-month follow-up, and daily steps correlated with several standard disease-activity and functional measures across a six-month observational study. Step count is clearly not a uniformly weak signal across every movement condition. It depends on which condition and which output is being read.
2 studies
- Comparing Fitbit Charge HR and Garmin vivosmart HR step counts against a research-grade accelerometer in people with mild-to-moderate Parkinson's disease during indoor and outdoor walks: both trackers showed low error and high consistency for step count, but cadence only weakly reflected heart rate and did not reflect oxygen consumption or perceived exertion.
- Fitbit step counts, compared against a research-grade ActiGraph monitor over four visits across six months in people with inflammatory myopathies including polymyositis and dermatomyositis, showed strong test-retest reliability and moderate-to-strong correlation with several standard disease-activity and physical-function measures.
Where the accuracy breaks down: placement, assistive devices, and energy expenditure
Device placement turns out to matter more once gait is atypical than it does for a typical walker. In one study simulating altered and shuffling gait, the Apple Watch performed best overall, and leg placement, not the wrist, accounted for half of the top 20 device-and-position combinations, suggesting that larger movement amplitude at the leg helps the algorithm register slower or shuffling steps that a wrist sensor may under-register.
Assistive devices complicate the picture further, and not uniformly. In the same study, wrist, hip, and ankle placement all correlated well with actual step count during a cane trial and during a deviceless control trial. A two-wheeled walker was the one condition that broke wrist accuracy specifically, producing a 31.2 percent average error at the wrist, compared to about 1.5 percent at the ankle or hip. The walker, not a cane and not the underlying gait pattern alone, appears to be what disrupts wrist-based counting in that scenario.
Energy expenditure is where the research turns most clearly negative. In Huntington's disease, Fitbit significantly overestimated energy expenditure against a metabolic cart reference, while a research-grade ActiGraph underestimated it, and Fitbit's reliability was described as poor at every tested treadmill speed. That is a meaningfully different verdict than the step-count findings above, and a reminder that calorie and energy-expenditure estimates carry their own separate accuracy problems even before a movement condition enters the picture.
None of these findings generalize across conditions. A device shown to be accurate for step count in Parkinson's disease was not tested in Huntington's disease, and the Huntington's energy-expenditure findings say nothing about step-count accuracy in that same population. Each result here is specific to the condition, metric, and device combination the study actually tested.
What still hasn't been tested
The systematic review covering stroke and traumatic brain injury is the clearest signal that some of this ground is simply unmapped rather than settled. Community-based wearable use for stroke has started but remains limited, and for traumatic brain injury it has not yet begun in earnest. Accuracy and validity for most outcome metrics used in the community were not reported across many of the included studies. That is a different and more honest finding than either "accurate" or "inaccurate," it means the evidence to answer the question directly does not yet exist at scale for those two conditions.
A separate pilot protocol is currently underway to validate consumer and research-grade wearables specifically in people with lung cancer, a population with its own distinct mobility challenges and gait impairments. That study had not yet reported results at the time of this review, but its existence is itself a signal that condition-specific validation is still actively expanding rather than complete.
A validation gap is not the same claim as a validation failure. The stroke and traumatic brain injury research base being thin means the accuracy question hasn't been answered yet for those conditions specifically, not that existing devices have been tested and found wanting.
Common questions
Are step counts accurate for people with Parkinson's disease?
Two separate studies found step count from consumer devices like Fitbit and Garmin trackers showed low error and good consistency against research-grade accelerometers in people with mild-to-moderate Parkinson's disease. Researchers have also established a specific threshold, roughly 581 steps per day, for what counts as a clinically meaningful change in that population.
Does a fitness tracker measure exercise intensity accurately if I have Parkinson's?
Not as well as it measures step count. In the same Parkinson's comparison study, cadence from wrist-worn trackers only weakly reflected heart rate and did not reflect oxygen consumption or perceived exertion at all, indicating that step-count accuracy and intensity accuracy are separate questions with separate answers.
Does using a cane or walker throw off a fitness tracker's step count?
It can, depending on wrist placement specifically. One study found a 31.2 percent average error at the wrist during walker use, compared to about 1.5 percent at the ankle or hip. Without an assistive device, all three placements performed well in the same study.
Is calorie or energy expenditure tracking accurate for movement disorders?
The evidence here is weaker than for step counting. In Huntington's disease specifically, a consumer tracker overestimated energy expenditure and a research-grade device underestimated it, with reliability described as poor across every tested walking speed.
Has wearable accuracy been studied for stroke or traumatic brain injury?
Only in a limited way so far. A systematic review found that community-based wearable use has started but remains limited for stroke, has not yet begun in earnest for traumatic brain injury, and that accuracy and validity were not reported for many of the studies that do exist in this space.
Sources
- Accurate measurement of simulated slow and altered walking activity: Apple Watch best in class wearable devices.
- The Effect of Assistive Devices on the Accuracy of Fitbits in Healthy Individuals: A Brief Report.
- Wearable devices for tracking physical activity in the community after an acquired brain injury: A systematic review.
- Minimal Clinically Important Difference of Average Daily Steps Measured Through a Consumer Smartwatch in People With Mild-to-Moderate Parkinson Disease: Cross-Sectional Study.
- Accuracy of wearable physical activity trackers in people with Parkinson's disease.
- Validation of ActiGraph and Fitbit in the assessment of energy expenditure in Huntington's disease.
- Consumer-based activity trackers in evaluation of physical activity in myositis patients.
- Comparing the Accuracy of Different Wearable Activity Monitors in Patients With Lung Cancer and Providing Initial Recommendations: Protocol for a Pilot Validation Study.