Skip to main content

Table 3 Methods used in adjusted analysis

From: Reporting on covariate adjustment in randomised controlled trials before and after revision of the 2001 CONSORT statement: a literature review

Year of Publication

2000

(n = 84)

2006

(n = 113)

Continuous data

65 (77.4%)

81 (71.7%)

   ANOVA/ANCOVA

50 (76.9%)

56 (69.1%)

   Multiple regression method*

7 (10.8%)

19 (23.5%)

   Stratified analysis

1 (1.5%)

0

   Other†

6 (9.2%)

6 (7.4%)

   Not mentioned

1 (1.5%)

0

Binary data

13 (15.5%)

19 (16.8%)

   Logistic regression

6 (46.1%)

11 (57.9%)

   Stratified analysis (Cochrane-Mantel-Haenszel test)

5 (38.5%)

5 (26.3%)

   Other‡

2 (15.4%)

3 (15.8%)

Ordinal data

1 (1.2%)

3 (2.6%)

   Stratified analysis (Cochrane-Mantel-Haenszel test)

1 (100%)

1 (33.3%)

   Nonlinear mixed effect model

0

1 (33.3%)

   Ordinal logistic regression

0

1 (33.3%)

Time to event data

5 (5.9%)

10 (8.9%)

   Cox proportional hazard

5 (100%)

9 (90.0%)

   Stratified log rank test

0

1 (10.0%)

  1. * Including random effect and mixed effect models
  2. † Including GEE, GLM, ANCOVA for rank data, Zellner seemingly unrelated regression, Poisson model, Van Elteren test
  3. ‡ Including non-parametric Generalized mixed effect model, GEE, non-parametric ANCOVA