Metaregression to explore heterogeneity in estimate of venous thromboembolism prevalence
| Studies (n) | Beta (95% CI)* | Intercept† | P value | R2 (%) | |
| Number of patients‡ | 56 | −0.0082 (−0.0143 to −0.0021) | 0.1929 | 0.008 | 11 |
| Proportion of patients in ICU | 56 | 0.0017 (0.0008 to 0.0025) | 0.0816 | <0.0001 | 28 |
| MINORS score | 31 | 0.0066 (−0.0065 to 0.0197) | 0.1212 | 0.32 | 0 |
| D-dimers§ | 30 | 0.0027 (0.0004 to 0.0051) | 0.1183 | 0.02 | 14 |
| Proportion of patients receiving anticoagulation | 38 | 0.0016 (−0.0013 to 0.0044) | 0.0311 | 0.29 | 0 |
| Proportion of men | 40 | 0.0042 (−0.0002 to 0.0087) | −0.0833 | 0.06 | 5 |
R2 estimates the amount of heterogeneity accounted for by the moderators.
*Beta signification: prevalence increase or decrease for the augmentation of one unit of the variable tested.
†Intercept signification: rate of venous thromboembolism for a variable with null value.
‡Per increase of 100 patients.
§Per increase of 100 µg/L.
ICU, intensive care unit; MINORS, Methodological Index for Non-randomised Studies.