Table 3 Best-fit model* from backward stepwise linear regression for predicting average daily step-count assessed by accelerometry
VariablesDifference in average daily step-count from baseline group (95% CI)p Value
Age
65–690
70–7422 (−884 to 929)0.96
75–79−842 (−1855 to 171)0.10
80 or more−1845 (−2950 to −741)0.00
Sex
Female0
Male338 (−371 to 1047)0.34
Pedometer worn
No0
Yes793 (29 to 1557)0.04
Disability score
0 (none)0
1–5 (mild)−471 (−1338 to 395)0.29
>5 (moderate/severe)−1363 (−2646 to −79)0.03
General health
Very good0
Good−969 (−1768 to −169)0.02
Fair/bad−1758 (−2908 to −606)0.00
Presence of diabetes
No0
Yes−1394 (−2387 to −402)0.01
Body mass index
Normal weight0
Overweight−1252 (−2140 to −364)0.01
Obese−1724 (−2831 to −617)0.00
Exercise self-efficacy
Low0
Moderate1108 (267 to 1949)0.01
High1885 (139 to 3631)0.04
Control over exercise
Low0
Moderate1284 (535 to 2034)0.00
High1380 (326 to 2436)0.01
  • *Final best-fit model based on 201 observations and adjusted for 158 household clusters