Search authoritative bibliographic databases, such as PubMed and Scopus to find scientific papers about the experimental nutritional supplement that you want to test and/or about the ingredients that it contains. Look in particular for similar clinical studies. Bibliographic information that relates to your research will help you design a study that is valid and original
Decide what the inclusion and exclusion criteria for the volunteers in your study will be (more 2.7.1).
Your findings regarding the effectiveness of your experimental supplement will be no good unless you compare it to a placebo (that is, an inert supplement) or to an established (commercial) supplement as control. This is called a controlled study, and you need to decide what your control supplement will be. Next, you need to decide whether your study will be an acute or prolonged one (more 2.7.2). In additional, you need to decide whether your study will be a parallel or crossover one (more 2.7.3).
Search authoritative bibliographic databases, as mentioned above, about the dosage of your experimental supplement and the control supplement. Then, decide the formulation that you will administer to the volunteers. Generally, formulations are classified as solid (such as capsules) or liquid (such as drinks). In prolonged studies, decide the duration (that is, how many days, weeks, or months they will consume the supplement) and frequency (that is, how many times during the day they will consume the supplement).
Decide whether your study will be open-label or blinded. In case you implement a blinded study, you must decide whether it will be single blind or double blind (more 2.7.4).
Decide where you will look for volunteers. Ideally, you should perform a random selection from among the population (just like pollsters do when they perform opinion polls), but this is seldom done in biological research. Usually, we resort to a "convenience" sample, that is, volunteers that are easy to find through public announcements, acquaintances, clienteles etc. Either way, you should describe how you picked your sample when you prepare a report or publication of your study.
Decide how you will randomize the volunteers between groups (in the case of a parallel study) or order of treatments (in the case of a crossover study). Examples of randomization include assignment in order of recruitment in an alternating fashion, by drawing lots, or through a random number generator.
Carefully choose the measurements you will perform, the timing of measurements, the facilities you will use, the staff you will need, and the equipment. Also decide on the statistical tests you will use to analyze your data.
Prepare a detailed questionnaire to be filled in and a consent form to be signed by the volunteers (more 2.7.5).
Apply to an appropriate Institutional Review Board for approval of the study design by carefully fulfilling all requirements. Do this well in advance of the anticipated start of the study, considering that your study design might not be approved on first submission (much like a scientific paper!).