Under the control of the study coordinator, patients are assigned ids and each is randomly included in one of the placebo or the drug groups. Then, the corresponding groups are prepared and encoded by id. To highly reduce the observer bias, patients and researchers ignore to which group a patient belongs until the end of the study. This is why the trial is called a double-blind experiment.
Data is collected from the appropriate devices or input by hand. Digital data are always preferred because this contributes to the credibility of the protocol.
Each data block is registred by patient id, time of the observation, time of the record and type of data. It can be sent and received by many ways depending on how data is input. Frequent calibrations data might be recorded to adjust the measures between different devices.
After the end of the collection campaign, the official raw data file is ready and the api refuses any new input for this project.
In fact, not only double-blind experiments are possible. Many variants like triple blind and randomised controlled trials protocols, eventually with more groups can be implemented as is.