Randomised Controlled Trial

Randomised Controlled Trial

The gold standard is the double blind randomised controlled trial. It has two arms: a study group and a control group. As discussed, blinding is designed to reduce observational bias and randomisation to reduce selection bias.

The advantages of a randomised controlled trial are:

Evaluation of single variable
Prospective study
Reduces bias
Allows for meta-analysis by combining several randomised controlled trials

Steps in setting up Randomised Controlled Trial

The following steps are required in setting up a randomised controlled trial:

Hypothesis
- Null hypothesis: there is no difference as opposed to the alternate hypothesis (there is a difference)
Inclusion / Exclusion Criteria
- Describe the criteria for exclusion and inclusion in the study
Outcome Measure
- Select outcome measure and the type of data. Data could be nominal, ordinal or continuous. Depending or the type of data, a test statistic is selected. If continuous data are normally distributed, parametric statistics (i.e. t-test) can be used. Alternatively, we can use a non parametric statistical test.
Bias
- Assessment of bias; there could be selection bias, confounding bias or observational bias. Selection bias is reduced by randomisation. If there is a known confounding factor, stratification can be used to reduce confounding bias. Observational bias is reduced by blinding.
Power Analysis
- Power analysis is performed to estimate the number of patients required in the study. It is necessary to have an estimate of:
  - Difference desired to detect
  - Spread of data (variance / standard deviation)
  - Significance level (α) (type 1 error)
  - Power (type 2 error)
- A pilot study might be required to estimate these parameters.
Ethical Approval
Informed Consent
Collection of Data & Results
Analysis
- Use parametric or non parametric tests as appropriate. Usually computers (R or Python) are used in the analysis.
Conclusions
Publication and Presentation of Data