Table 8. Association of the BMQ-PL subscales with self-reported measure of drug adherence.
Drug adherence was assessed with the use of Polish version of the Adherence to Refills and Medications Scale. Statistically significant results are presented in bold. The extent of cell shading reflects the extent of correlation. Results of equivalent non-parametric tests for the univariate associations were presented in S1 Table and their similarity to parametric ones documented in S1 Fig.
| Group of cardiovascular patients | Difference between in- and outpatients** | |||||
|---|---|---|---|---|---|---|
| Inpatients | Outpatients | |||||
| Raw estimate | Adjusted estimate* | Raw estimate | Adjusted estimate* | Raw p-value | Adjusted p-value* | |
| Specific-Necessity | -0.04 p = 0.71 | -0.02 p = 0.84 | 0.29 p = 0.0053 | 0.22 p = 0.059 | p = 0.031 | p = 0.078 |
| Specific-Concerns | -0.21*** p = 0.038 | -0.24 p = 0.036 | -0.24 p = 0.024 | -0.24 p = 0.021 | p = 0.85 | p = 0.71 |
| Necessity-minus-Concerns | 0.14 p = 0.18 | 0.17 p = 0.13 | 0.40 p = 0.0001 | 0.33 p = 0.0018 | p = 0.069 | p = 0.38 |
| General-Overuse | 0.08 p = 0.45 | 0.10 p = 0.39 | -0.27 p = 0.0091 | -0.27 p = 0.0095 | p = 0.024 | p = 0.049 |
| General-Harm | -0.11 p = 0.27 | -0.11 p = 0.35 | -0.30 p = 0.0030 | -0.25 p = 0.015 | p = 0.19 | p = 0.27 |
* adjusted for age, sex, education, place of residence and number of drugs used.
** General Linear Model–variables included to the model: group, BMQ subscale and their two-way interaction.
*** equivalent non-parametric Spearman’s rho test yielded non-significant result (p = 0.063).