Table 2

Factors influencing harm reported in the abstract

cPR (CI95%)

Funding*

drug or device companies

1

public institutions

0.77 (0.60 to 0.97)

Placebo

other control

1

placebo

0.97 (0.81 to 1.13)

Sample size

less than 20

1

equals or more than 200

1.14 (0.96 to 1.33)

Direction of effect

beneficial to intervention

1

neutral or against the intervention

1.00 (0.83 to 1.19)

Harm in text

no harm in text

1

harm in text

8.71 (5.90 to 12.01)

Clinical significance of harm

no mortality nor composite in text

1

mortality or composite in text

1.43 (1.22 to 1.65)

Harm in text statistically significant

not statistically significant

1

statistically significant

1.70 (1.47 to 1.92)

cPR: Prevalence Ratio estimated by logistic regression, and corrected by applying the method proposed by Zhang [24]; CI: Confidence Intervals with error type I equals 5%; *363 articles except for funding in which sample size was n = 226

Bernal-Delgado and Fisher BMC Medical Research Methodology 2008 8:14   doi:10.1186/1471-2288-8-14