Findings, implications and lessons learned

1. The above study protocol is typical of protocols that need to be prepared before any study is undertaken. A number of details are included deliberately in this protocol to stimulate further discussion on possible improvements that might be made. The first point to note about the protocol is that many aspects to experimental design are concerned with scientific rather than statistical method. This is important. There is no point designing a good statistical design if the scientific conduct of the experiment is poor. A biometrician should be able to follow what is proposed without necessarily understanding the details.

2. A protocol will start with an introduction that describes the background to the study, previous work that has been done and reasons for approaching the study in the way proposed. Here we have separated the introduction from the protocol and included it as the Background at the beginning of this case study. The introduction/background to the study has also been intentionally abbreviated to simplify it for the purpose of this case study. Usually the introduction will contain a literature review referring to work already done and into which the proposed study can be put into context.

3. The objectives should then be clearly stated. Researchers sometimes find these difficult, but they are essential for ensuring that an appropriate study is designed to meet the specified objectives and that the results can be suitably analysed to fulfil the study expectations.