Skip to main content

Table 2 Prevalence estimates with 95% confidence intervals of any potential pitfalls in meta-regression-analyses

From: Most published meta-regression analyses based on aggregate data suffer from methodological pitfalls: a meta-epidemiological study

Pitfall Total (N = 81) Year 2002 (N = 29) Year 2012 (N = 52)
Ecological fallacy 53 (65%, 55 to 75%) 20 (69%, 51 to 83%) 33 (63%, 50 to 75%)
Overfitting 14 (17%, 11 to 27%) 6 (21%, 10 to 38%) 8 (15%, 8 to 28%)
Meta-regression on risk of the analysed outcome 5 (6%, 3 to 14%) 2 (7%, 2 to 22%) 3 (6%, 2 to 16%)
Any potential meta-regression pitfall 57 (70%, 60 to 79%) 21 (72%, 54 to 85%) 36 (69%, 56 to 80%)