Skip to main content

Table 3 Study characteristics in cohort studies in high impact journals and in low impact journals

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

Study characteristics

Cohort studies in High impact journals (n = 75)

Cohort studies in low impact journals (n = 75)

p-value*

Sample size, Median(Interquartile range)

4176(1274–23,452)

415(138.5–2586)

<0.001

Length of follow up (years), Median(Interquartile range)

5(2–10)

2.1(1–5)

0.002

Date of publication

  

<0.001

 2000–2013

57(76)

13(17)

 

 2014–2016

17(24)

62(83)

 

Journal type, n(%)

   

 High impact factor journals (HIJ)

75(100)

 

 Low impact factor journals (LIJ)

75(100)

 

Study area, n(%)

  

0.81

 Interventional

23(30.7)

27(34)

 

 Pharmacological

46(61.3)

43(57)

 

 Other

6(8)

5(9)

 

Source of funding, n (%)

  

0.15

 Industry

51(68)

49(65)

 

 Other

18(24)

12(16)

 

 NA

6(8)

14(19)

 

Type of primary outcome, n (%)

  

0.001

 Time to event

44(58.7)

21(28)

 

 Binary

19(25.3)

31(41)

 

 Continuous

12(16)

23(31)

 

Main effect of primary outcome, n (%)

  

0.02

 Statistically significant

55(73)

41(54.7)

 

 Not statistically significant

20(27)

34(45.3)

 

Newcastle-Ottawa Quality Assessment Score, n(%)

  

0.03

 0–3

1(2)

4(5)

 

 4–6

46(61)

42(56)

 

 7–8

28(37)

29(39)

 

HTE reporting

55(73)

33(44)

<0.001

Analysis of predictive factors on side effects, n(%)

2(4)

0(0)

0.53

Prespecification of HTE, n(%)

54(98)

30(91)

0.15

Type of HTE analysis, n(%)

  

0.44

 Subgroup

1(2)

1(3)

 

 Adjusted

40(73)

27(82)

 

 Propension

14(25)

5(15)

 

Predictive variables studied, n(%)

   

 Age

23(42)

7(21)

0.06

 Sexe

12(22)

3(9)

0.15

 Social level

2(4)

2(6)

0.63

 Genetics

2(4)

1(3)

0.03

 Treatment

44(80)

22(67)

0.21

 Comorbidities

23(42)

8(24)

0.11

 Severity of disease

9(16)

4(12)

0.76

Prognostic variables studied, n(%)

27(36)

24(73)

0.7

Analysis of prognostic factors on side effect, n(%)

1(1)

1(3)

1

Interaction test, n(%)

21(28)

6(8)

0.06

  1. *to determinate p-value we used the Wilcoxon rank sum test for the analysis of continuous data and the chi square/fisher test for binary data