Table 2

Logistic regression models for predicting PA level 1 year poststroke

CoefficientBSEWald's testdfpOR (95% CI)
(a) Dependent variable of mostly inactive (n=73)
 Constant−6.522.159.1710.0020.001
 Age0.060.034.1810.0411.07 (1.00 to 1.13)
 mRS at discharge1.950.717.4310.0067.01 (1.73 to 28.43)
 Prestroke PA (mostly inactive)2.010.816.1010.0147.46 (1.51 to 36.82)
 TestΧ2 dfp
 Likelihood ratio test32.593<0.001
 Hosmer and Lemeshow9.6680.290
(b) Dependent variable of low level of PA (n=77)
 Constant−8.122.2513.031<0.001<0.001
 Age0.130.0313.521<0.0011.13 (1.06 to 1.21)
 mRS at discharge1.290.614.4110.0363.62 (1.09 to 12.04)
 TestΧ2 dfp
 Likelihood ratio test30.472<0.001
 Hosmer and Lemeshow3.2870.858
  • Dependent variable coded as (a) mostly active=0, mostly inactive=1; (b) moderate/high PA=0, low PA=1; Cox and Snell R2(a)=0.360; (b)=0.327 Nagelkerke R2(a)=0.489; (b)=0.453.

  • PA, physical activity; mRS, Modified Rankin Scale.