Table 3

Medians and quartiles of percentages of assessments for completeness and correctness of statistical aspects for AMG studies in comparison to non-regulated studies (higher values indicate a lower level of completeness and/or correctness)

In x(%) of the ethical proposals	Median (1.; 3. quartile), n=number of valid responses (excluding ‘non assessable’)
In x(%) of the ethical proposals	Regulated studies (AMG, MPG)	Non-regulated studies
(1)… wording of study aims, hypotheses and/or endpoints are inadequate or inconsistent.	10 (10; 20), n=45	50 (20; 80), n=45
(2)… specification of the significance level is missing.	10 (0; 10), n=44	30 (10; 50), n=45
(3)… sample size calculation is incomplete, inadequate or missing completely.	20 (10; 30), n=45	70 (50; 80), n=45
(4)… only general statements on statistical analysis methods are provided not fitting and addressing the specific study aim.	10 (10; 25), n=46	70 (50; 80), n=45
(5)… no clear differentiation between confirmatory and explanatory analyses is provided.	15 (10; 27.5), n=46	70 (50; 82.5), n=45
(6)… specification of one-sided or two-sided statistical testing is missing.	10 (10; 30), n=44	55 (30; 87.5), n=43
(7)… description of how to handle missing values is incomplete, inadequate or missing completely.	40 (22.5; 70), n=46	90 (80; 90), n=42
(8)… description of how to handle multiple comparisons problems is incomplete, inadequate or missing completely.	30 (20; 50), n=45	80 (70; 90), n=42
(9)… description of how and for which covariables adjustment is planned is incomplete, inadequate or missing completely.	25 (20; 50), n=46	80 (50; 90), n=42
(10)… specification of randomisation and stratification is incomplete, inadequate or missing completely.	10 (10; 30), n=45	65 (50; 80), n=36
(11)… no study biometrician is specified.	20 (10; 50), n=43	70 (40; 90), n=40
(12)… description of statistical methods is not sufficiently specified.	20 (10; 40), n=45	80 (60; 90), n=42

AMG/MPG, German Medicines Act/German Act on Medical Devices.