Overview of the multilevel regression models
| Model | Aim | Dependent factor | Potential predictors |
|---|---|---|---|
| A | To determine the difference between experimental groups regarding the within group change in patients' evaluations of care over time, taking into account clustering of data within GPs | Change of patient evaluations per patient (post-intervention score minus baseline score) | GP factor; group allocation (intervention v control group) |
| B | To determine the proportion of variation associated with systematic differences between GPs regarding post-intervention patients' evaluations of care | Post-intervention patients' evaluations per patient | GP factor |
| C | To determine the difference between experimental groups regarding post-intervention patients' evaluations of care, taking clustering of data within GPs into account (“straightforward intervention effect”) | Post-intervention patients' evaluations per patient | GP factor; group allocation (intervention v control group) |
| D | To determine the difference between experimental groups regarding post-intervention patients' evaluations of care, taking into account the clustering of data within GPs as well as the pre-intervention scores (“conditional intervention effect”) | Post-intervention patients' evaluations per patient | GP factor; group allocation (intervention v control group); pre-intervention score of patient evaluation (mean score per GP) |
| E | To determine the difference between groups defined by GP reported actions on the relevant dimensions regarding post-intervention patients' evaluations of care, taking into account the clustering of data within GPs as well as the pre-intervention scores | Post-intervention patients' evaluations per patient | GP factor; group allocation defined by GP reported action on the relevant dimension (yes v no); pre-intervention score of patient evaluation |