Skip to main content

Table 1 Binary logistic regression analysis

From: The risk of bias in randomized controlled trials in otorhinolaryngology: hardly any improvement since 1950

Time strata

Random sequence generation (n = 384)

Allocation concealment (n = 399)

Blinding participants and personnel (n = 137)

Blinding outcome assessment (n = 130)

Blinding (n = 227)

Incomplete outcome data (n = 345)

Selective reporting (n = 337)

Other bias (n = 273)

<1990

1.00

1.00

1.00

1.00

1.00

1.00

1.00

1.00

1990–1995

1.17 (0.58–2.35)

0.74 (0.34–1.62)

1.29 (0.48–3.50)

0.71 (0.26–2.00)

1.12 (0.44–2.84)

0.49 (0.24–0.99)

1.13 (0.56–2.26)

1.10 (0.49–2.49)

1996–2000

2.20 (1.07–4.50)

0.93 (0.42–2.06)

0.85 (0.26–2.80)

1.07 (0.34–3.35)

0.90 (0.36–2.29)

0.78 (0.37–1.63)

2.31 (1.06–5.01)

0.95 (0.40–2.26)

2001–2005

3.07 (1.69–5.57)

2.09 (1.13–3.88)

0.35 (0.11–1.15)

0.17 (0.05–0.62)

1.019 (0.50–2.09)

1.65 (0.86–3.19)

3.65 (1.90–7.00)

2.23 (1.12–4.42)

2006–2012

6.09 (3.11–11.95)

3.59 (1.87–6.90)

0.66 (0.22–2.00)

0.93 (0.42–3.91)

0.68 (0.31–1.48)

1.27 (0.63–2.54)

1.21 (0.63–2.31)

0.95 (0.49–1.93)

  1. Legend: The odds ratios for scoring a low RoB compared to RCTs published in the time stratum ‘<1990’ are presented per time stratum per item. We used RoB assessment “low” as the reference category in a binary logistic regression analysis in comparison with RoB assessments “unclear” and “high” taken together, and time stratum ‘<1990’ was the reference time stratum. Values in bold font are statistically significant (p < .05)