Predictors of ambulatory recovery and walking proficiency in community-dwelling stroke survivors: a cross-sectional study

Bulletin of Faculty of Physical Therapy

Table 5 Standard multiple regression analysis model for the prediction of functional ambulatory profile

Dependent variable	Predictors	SE	β	t-val.	p-value	95% CI
Functional ambulatory profile	(Constant)	156.19		4.09	<0.001	330.32–947.48
	LEMF	2.46	0.075	1.14	0.255	−2.05–7.68
	MBL	2.81	0.105	1.66	0.099	−0.89–10.21
	PSTPL	1.66	0.042	0.56	0.575	−2.35–4.21
	Gait speed	81.45	−0.181	−2.32	0.022*	−349.55–−27.72
	Cadence	0.86	−0.085	−1.22	0.224	−2.74–0.65
	PSWT	34.57	−0.177	−2.58	0.011*	−157.52 to −20.91
	PSTT	31.54	0.142	1.72	0.087	−8.01–116.59
	PIDLST	37.67	0.223	3.03	0.003*	39.73–188.59
	Balance	1.93	−0.489	−6.21	<0.001*	−15.77 to −8.16
	Cognition	1.36	−0.071	−1.09	0.278	−4.16–1.21

Regression equation: Y = a + bx; thus, functional ambulatory profile = 638.90 +bx where Y = value of dependent variable, a = constant, b = regression coefficient of each predictor variable, and x = value of each predictor. Values of b and x will be added to the equation continuously depending on the number of predictor variables
Key: LEMF Lower extremity motor function, MBL Mobility level, PSTPL Paretic step length, PSWT Paretic swing time, PSTT Paretic stance time, PIDLST Paretic initial double-limb support time, B unstandardized coefficient, SE Standard error, β standardized coefficient, t-val. t-statistics, p-value significance level, 95% CI 95% confidence interval, Part C. part correlation, Tolr. tolerance, df degree of freedom, R² coefficient of determination, F ANOVA value
Final regression model for predicting functional ambulatory profile (R² =0.529, F_{4, df=10,} =17.098, p ≤ 0.001)
*Significant at p ≤ 0.05