Skip to main content

Table 2 Univariate et multivariable regression analysis of factors associated with reporting versus not reporting of HTE analysis

From: Reporting of heterogeneity of treatment effect in cohort studies: a review of the literature

 

Logistic univariate regression analysis

Logistic multivariable regression analysis

Study characteristics

Odds ratio (95% CI)*

P-value

Odds ratio (95% CI)*

P-value

Sample size

0.99(0.99–1.00)

0.11

1(0.99–1)

0.58

Length of follow up (years)

1.17(1.07–1.30)

0.02

1.1(0.99–1.24)

0.07

Date of publication (<2014 vs ≥2014)

0.5(0.25–0.97)

0.04

0.6(0.16–2.17)

0.42

Journal type (high impact factor)

3.5(1.78–7.5)

<0.001

0.4(0.1–1.55)

0.19

Source of funding(industry funded vs non industry funded)

1 .92(0.82–4.78)

0.85

 

0.37

Study area (pharmacological vs non pharmacological)

0.26(0.13–0.51)

<0.001

0.25(0.09–0.66)

0.007

Type of primary outcome

 Binary

1(reference)

1(reference)

 Time to event

3.05(1.41–6.80)

0.05

2.22(0.72–7.12)

0.17

 Continuous

0.91(0.38–2.17)

0.83

0.98(0.26–3.7)

0.98

Main effect of primary outcome (significant vs non significant)

1.22(0.62–2.40)

0.56

0.97(0.34–2.71)

0.95

Newcastle-Ottawa Quality Assessment Score, n(%) (continuous)

 0–3

 4–6

 7–8

ref

1[0.13–6.70]

0.79[0.09–5.14]

ref

0.95

0.81

ref

1.61(0.17–1.99)

1.75(0.12–2.3)

ref

0.70

0.66

  1. *in our model, we adjusted on the following variables: source of funding, study area, journal type, main effect of primary outcome, type of primary outcome, quality of the cohort