|Association (regression coefficient beta/odd’s ratio (OR) and 95% confidence interval (CI)) of dose of PRO-FIT*advice and counselling with post-test LDL-C and multiple lifestyle behaviours, adjusted for baseline levels of the dependent variable, in the intervention group (n = 181)|
|LDL-C||MVPA1||Fat intake||Fruit intake||Vegetable intake||Smoking||Compliance to statin therapy|
|95% CI||95% CI||95% CI||95% CI||95% CI||95% CI||95% CI|
|Participants who had logged on at PRO-FIT*advice and completed at least one advice module:||−0.18|
|Participants who had logged on at PRO-FIT*advice and completed the module on:|
|Compliance to statin therapy||−0.11||1.09|
|Participants who had received face-to-face counselling||N/A2||N/A2||N/A2||N/A2||N/A2||N/A2||N/A2|
|Telephone booster calls delivered (mean, SD)||0.06||−0.04||0.26||−0.03||−4.66||1.00||1.02|
|Participants who had logged on, finished at least 1 module3, received face-to-face counselling and at least 1 telephone booster call (=complete intervention-package)||−0.18||0.10||−0.50||0.16||−6.87||0.11||0.90|
1 MVPA=moderate to vigorous physical activity. Due to skewed data, log-linear regression was conducted. Therefore, the beta should be interpreted as follows: a 1% increase of the independent variable is associated with a beta% increase in physical activity.
2 Due to minimal variation in dose delivered, no association between dose delivered and efficacy could be tested.
3 For LDL-C this means at least one module, for the lifestyle behaviours, this means the related advice module (e.g. for physical activity, the completion of the physical activity module).
Significant associations between dose and efficacy (p < 0.05) are printed in bold. Effect parameters (beta regression coefficient or odd’s ratio (OR)) either indicated a positive association if LDL-C/lifestyle behaviours improved when regressed to the process, or a negative association if vice versa.
Broekhuizen et al.
Broekhuizen et al. BMC Public Health 2012 12:348 doi:10.1186/1471-2458-12-348