Your Easy Guide to Clinical Studies

A statistical hypothesis is an assumption about a population parameter that is subject to empirical testing.

 

Example: a study assesses the effect of a new medication on blood pressure. The parameter in question is the mean blood pressure of the population.

 

Researchers establish a statistical hypothesis by formulating two competing hypotheses:

• A null hypothesis - also referred to as H0:  represents a default assumption (i.e. no treatment effect exists. Based on the example above, no difference in blood pressure between the intervention and control group is detected)
• An alternative hypothesis - also referred to as Ha: suggests a difference or an effect (i.e. a treatment effect does exist. Based on the example above, blood pressure is higher in participants taking a placebo than in participants taking the medication under investigation)

 

Statistical analyses are used to assess the data collected in a study, enabling researchers to conclude whether H0 may or may not be rejected in favour of Ha.  

As a SP-INV, think about the statistical framework of your study. As described in the sample size calculation topic, there are two possible frameworks:

• The precision-based framework
• The hypothesis testing framework

 

If the study aims at estimating a quantity with a certain accuracy, you are in the precision-based framework and therefore do not need to define a statistical hypothesis. Example: a study aims at estimating the proportion of deaths within 30 days of a myocardial infarction.

 

If you need to conduct a statistical test to address your research question, you are in the hypothesis testing framework, and need to define H0 and Ha.

 

Example: a study aims at testing the effect of an intervention (e.g. blood pressure lowering drug) between a placebo and intervention group.

References

ICH Topic E9 – see in particular

• 5.5 Estimation, confidence intervals and hypothesis testing

Swiss Law

ClinO – see in particular article

• Art. 2b Definition intervention

ClinO-MD – see in particular article

• Art. 2a Definition of clinical intervention
• Art. 2a Definition of performance study

HRO – see in particular article

• Art. 3a Definition of research