Skip to main content

Table 2 Summary of the accuracy of accelerometers worn on various anatomical locations and wheelchair during laboratory protocols. Comparison to criterion measures of oxygen uptake, energy expenditure and physical activity energy expenditure

From: Measurement of Physical Activity and Energy Expenditure in Wheelchair Users: Methods, Considerations and Future Directions

Study

Samplea

Criterion measure

Activity protocol

Device/outputs

Anatomical location

Results

Garcia-Masso et al. [38]

20 SCI

(T4-S1)

V̇O2

(COSMED K4b2)

Ten activities which included ADL, transfers, ACE and propulsion that covered a wide range of exercise intensities.

GT3X (36 features extracted from the second-by-second acceleration signals were used as independent variables)

Non-dominant wrist

r = 0.86, MSE = 4.98 ml kg−1 min−1

Dominant wrist

r = 0.86, MSE = 5.16 ml kg−1 min−1

Chest

r = 0.68, MSE = 10.41 ml kg−1 min−1

Waist

r = 0.67, MSE = 10.61 ml.kg−1.min−1

Learmonth et al. [39]

24 (9 F).

SCI (n = 10),

SB (n = 5),

MS (n = 4),

AMP (n = 2),

CP (n = 1),

Other (n = 2)

V̇O2

(COSMED K4b2)

Three wheelchair propulsion speeds (1.5, 3.0 and 4.5 mph) on a WT

PAC from GT3X ACC

Right wrist

r = 0.95

Left wrist

r = 0.93

Combined

r = 0.94

Washburn & Copay, [37]

21 (9 F).

SCI (n = 11), SB (n = 7), other (n = 3)

V̇O2

(Aerosport TEEM 1000)

Three timed pushes (slower than normal, normal, and faster than normal) over a rectangular indoor course

PAC from a CSA uniaxial ACC

Left wrist

r = 0.67, SEE = 4.99 ml.kg−1.min−1

Right wrist

r = 0.52, SEE = 5.71 ml.kg−1.min−1

Hiremath & Ding, [54]

24 SCI (5 F) (T3-L4)

IC EE (COSMED K4b2)

Resting and three activity routines; propulsion (performed on a WERG and flat tiled surface), ACE (20–40) and deskwork.

PAC from a RT3 tri-axial ACC and participant demographics

Waist

r = 0.66, SEE = 1.38 kcal.min−1. EE estimation errors ranged from 12.9 – 183.4%

Upper left arm (general equation)

r = 0.83, SEE = 1.02 kcal · min−1. EE estimation errors ranged from 14.1% – 113.7%

Upper left arm (activity specific equations)

r ranged from 0.63 (deskwork) to 0.91 (propulsion). EE estimation error ↓ to between 12.2 and 38.1%

Combined (Waist and upper arm)

r = 0.84, SEE = 1.00 kcal.min−1. EE estimation errors ranged from 14.1% – 116.9%

Kiuchi et al. [41]

6 SCI

(C6 – T9)

IC EE

(AR-1 Type-4)

Propulsion at three continuous speeds on a WT that elicited an RPE of 9 (2.5–3 km/h), 11 (3.5–4.0 km/h) and 13 (4.5–5.0 km/h)

Tri-axial ACC with gyro sensor. EE was predicted by incorporating acceleration, angular velocity and participant demographics

Left wrist

r = 0.93

Right wrist

r = 0.82

Left upper arm

r = 0.87

Right upper arm

r = 0.93

Hiremath et al. [60]b

45 SCI (6 F) (C5 – L5)

IC EE (COSMED K4b2)

Participants performed 10 activities from a list that included a range of activities and exercises of differing intensities

Gyroscope-based wheel rotation monitor (G-WRM) and tri-axial accelerometer; The Physical Activity Monitoring System (PAMS)

PAMS-arm

ICC = 0.82 (95% CI; 0.79 – 0.85)

M ± E = 9.82 ± 37.03%, MAE = 29.04%

PAMS-wrist

ICC = 0.89 (95% CI; 0.87 – 0.91)

M ± E = 5.65 ± 32.61%, MAE = 25.19%

Nightingale et al. [42]

15 (3 F).

SCI (n = 9), SB (n = 2), Other (n = 4)

IC PAEE (COSMED K4b2)

Five activities including deskwork and wheelchair propulsion at various velocities around an outdoor athletics track.

PAC from GT3X + ACC

Right wrist

r = 0.93, SEE = 0.80 kcal.min−1. Absolute bias ± 95% LoA: 0.0 ± 1.55 kcal.min−1

Right upper arm

r = 0.87, SEE = 1.05 kcal.min−1. Absolute bias ± 95% LoA: 0.0 ± 2.03 kcal.min−1

Waist

r = 0.73, SEE = 1.45 kcal.min−1. Absolute bias ± 95% LoA: 0.0 ± 2.82 kcal.min−1

Nightingale et al. [43]

17.

SCI (n = 10),

SB (n = 3),

CP (n = 1),

AMP (n = 1),

Other (n = 2).

IC PAEE (TrueOne 2400, ParvoMedics)

A wheelchair propulsion protocol across a range of treadmill velocities (3 – 7 km/h and gradients (1 – 3%) including load carriage (+8% body mass) and a folding clothes task

PAC from GT3X+ ACC

Right Wrist

r = 0.82, SEE = 0.91 kcal.min−1.

MAE = 0.69 kcal.min−1 (33.0%)

Right Upper Arm

r = 0.68, SEE = 1.16 kcal.min−1.

MAE = 0.86 kcal.min−1 (35.3%)

Raw acceleration (g · s−1) from GENEActiv ACC

Right Wrist

r = 0.88, SEE = 0.75 kcal.min−1.

MAE = 0.59 kcal.min−1 (21.0%)

Right Upper Arm

r = 0.87, SEE = 0.77 kcal.min−1.

MAE = 0.58 kcal.min−1 (20.4%)

  1. AB able-bodied, ACC accelerometer, ACE arm crank ergometry, ADL activities of daily living, AMP amputee, CP cerebral palsy, CSA computer science applications, EE energy expenditure, IC indirect calorimetry, LoA limits of agreement, MAE mean absolute error, MS multiple sclerosis, MSE mean square error, M ± E mean signed error, PAC physical activity counts, PAEE physical activity energy expenditure, SB Spina Bifida, SCI spinal cord injury, SEE standard error of estimate, ULAM upper limb activity monitor, V̇O 2 oxygen uptake, WERG wheelchair ergometer, WT wheelchair treadmill
  2. aAll male participants unless stated otherwise
  3. bNote to avoid confusion and make interpretation easier, + or − before M ± E statistics has been switched to reflect whether prediction method over or under predicted EE