Table 3.
Univariate random effects meta-regression (methods of moments) and subgroup analyses for AGE
| Subgroup analyses | Results from meta-regression | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Results | Heterogeneity | ||||||||
| Variable | Subgroup | Number of studies | OR | 95% CI | P-value | I 2 | Coefficient | SD | P-value |
| Outcome measurement | Mixed | 2 | 0.986 | 0.959–1.013 | 0.037 | 77% | Reference | ||
| Formal | 9 (12 different samples) | 0.969 | 0.946–0.992 | 0.000 | 89% | −0.0168 | 0.0251 | 0.504 | |
| Determinant measurement | Dichotomous | 1 (2 different samples) | 0.485 | 0.385–0.611 | 0.248 | 25% | Reference | ||
| Continuous | 10 (12 different samples) | 0.983 | 0.969–0.996 | 0.000 | 78% | 0.7122 | 0.1014 | 0.000 | |
| Proportion of volunteers (%) in baseline study sample | Continuous | 10a (13 different samples) | 0.970 | 0.950–0.991 | 0.000 | 88% | 0.0000 | 0.0001 | 0.591 |
| 0–100% | 9 (11 different samples) | 0.985 | 0.971–0.999 | 0.000 | 79% | Reference | |||
| 0% | 1 | 0.440 | 0.343–0.565 | N.A. | N.A. | −0.8053 | 0.1292 | 0.000 | |
| 100% | 2 | 0.747 | 0.444–1.256 | 0.001 | 91% | −0.0500 | 0.0292 | 0.087 | |
| Mean age at baseline | Continuous | 10b (12 different samples) | 0.983 | 0.969–0.996 | 0.000 | 78% | −0.0000 | 0.0000 | 0.200 |
| ≤ 55 years | 5 (6 different samples) | 0.991 | 0.975–1.007 | 0.000 | 79% | Reference | |||
| > 55 years | 6 (8 different samples) | 0.944 | 0.904–0.986 | 0.000 | 89% | −0.0296 | 0.0198 | 0.135 | |
| Continent | USA | 6 | 0.978 | 0.959–0.998 | 0.001 | 77% | Reference | ||
| Europe | 5 (8 different samples) | 0.966 | 0.933–1.000 | 0.000 | 91% | −0.0026 | 0.0210 | 0.900 | |
| Year of baseline measurement | Continuous | 11 (14 different samples) | 0.989 | 0.984–0.995 | 0.000 | 87% | 0.0006 | 0.0014 | 0.686 |
| < 2006 | 8 (10 different samples) | 0.970 | 0.948–0.993 | 0.000 | 91% | Reference | |||
| ≥ 2006 | 3 (4 different samples) | 0.975 | 0.959–0.991 | 0.388 | 1% | 0.0110 | 0.0230 | 0.631 | |
| Risk of bias items | |||||||||
| Study participation | Unclear/high risk of bias | 9 (12 different samples) | 0.975 | 0.956–0.995 | 0.000 | 88% | Reference | ||
| Low risk of bias | 2 | 0.858 | 0.623–1.192 | 0.000 | 93% | −0.0099 | 0.0315 | 0.754 | |
| Study confounding | Unclear/high risk of bias | 1 | 1.000 | 0.981–1.020 | N.A. | N.A. | Reference | ||
| Low risk of bias | 10 (13 different samples) | 0.970 | 0.950–0.991 | 0.000 | 88% | −0.0303 | 0.0331 | 0.360 | |
aThe study of Ajrouch et al. (2014) is not included in this analysis, because the proportion of volunteers (%) in the baseline study sample is not reported
bThe study of Hank & Erlinghagen (2010) is not included in this analysis, because the mean age at baseline is not reported