Table 2. Cohort specific sample size calculation.

Cohort	Hypothesis	Critical value	Time frame	Assumptions	Loss to follow-up	Statistical method proposed	Sample size
Parkinson’s Disease	Reduced gait speed is associated with increased fall risk	a coefficient of 1 for a 1 m/s decline in walking speed	24 months [57]	(i) a standard deviation of real‐world walking speed of 0.11 m/s or higher [58, 59], (ii) a 2‐y rate of falls of 1.6 or higher [59]	20%	Poisson regression model	600
Multiple Sclerosis	Reduced RWS is associated with fall frequency	a coefficient of >0.5 for a 1 m/s decline in walking speed	24 months	(i) a standard deviation of real‐world walking speed of 0.13 m/s or higher (19), (ii) a 2‐y rate of falls of 5 or higher [60, 61]	10%	Poisson regression model	600
Chronic Obstructive Pulmonary Disease	RWS is associated with COPD exacerbations	an odds ratio of 1.43 for a 0.1 m/s decline in walking speed [62]	12 months	(i) a standard deviation of real‐world walking speed of 0.115 or higher [55]l, (ii) a proportion of exacerbations up to 0.6 [24, 62, 63]	30% [24, 62, 63]	Logistic regression model	600
Proximal Femoral Fracture	RWS is associated with care home admission	odds ratio of 2.55 in the risk of admission on slow RWS vs normal/high RWS [64]	6 months	(i) a ratio of low RWS vs normal/high RWS of 1:2, (ii) a six month proportion of admissions to care home of 10 to 30% [46, 65]	30% [46, 65]	Logistic regression model	572 (≈600)

RWS, real world walking speed.