Self-made CRF: what do and what avoid to do

I review and generate CRF on a daily basis, this leading me to spot some best practices that I would like to share with my network:

1- to each his own job: the researcher must necessarily define the variables and it would be nice that he set the draft of the CRF, but the final version must be modelled by an expert CSC (having seen more than 5 different CRFs), a central data manager or a statisticians, because they know the best way to do it;

2- temporal steps= logical structure: the flow-chart is the foundation of the CRF, spending time doing it lead you to a better CRF and to gain time later (i.e. MELD score should be placed after haematological data);

3- comfort does not mean efficiency: capturing a date with a single field is convenient, but if most of the times the day is missing, it is more useful to acquire day, month and year separately; another example, set an open field is easier with respect to thinking about all the potential options, but if the same information is described in 10 different ways, will you be able to reconcile all of them?

4- the sections help to structure data: to put together the laboratory data with those of hemogas surely reduces the pages of the CRF, but it reduces the readability too and rather than to invest hours in the data cleaning, maybe it would be better to invest half an hour for sections drawing;

5- patient identity is not an option: a clinical trial is not a registry, do not collect Name, Surname and/or patient initials: patient identity is not useful for meet study objectives, only subject features matter (for a discussion about patient Vs subject, read this old article of mine: http://bit.ly/32fc4CU);

6- no consent no party: ICH GCP E6 state that "a subject voluntarily confirms his or her willingness to participate in a particular trial, after having been informed of all aspects of the trial that are relevant to the subject's decision to participate. Informed consent is documented by means of a written, signed and dated informed consent form" (ICH GCP E6 1.28) and moreover "Freely given informed consent should be obtained from every subject prior to clinical trial participation" (ICH GCP E6 2.9)

7- dignity to the date of the event/visit: if the study is composed of several time points/events/visits, it is important to dedicate a section to the collection of that single data, to make it easily recognizable (for those who insert the information) and accessible (for Data managers and Biostatisticians);

8- less is best: to be sure to collect only useful information ask yourself: can I meet the study objectives without this? If the answer is yes, remove that field. This will lead to fewer data to input/ to insert in CRF, less work for data entry operators, more focused operators, fewer mistakes, a shorter cleaning phase, faster results and happier PI and promoters too!

9- multiple Vs mutually exclusive choices: radio buttons and checkboxes are not interchangeable, they collect two different information: use radio buttons or dropdown menus (two way to display the same thing) when you can have only one possible answer, instead you must use checkboxes when you can have more than one answer for that question

10- saving space and losing time: I saw too much dropdown menus (DDM) in the CRF, why do you use a DDM to list 3 answers? Is it useful? I do not think so, because you need time to open the DDM whereas with radio buttons you can find the answer, at first sight, leading to saving time, efforts and concentration.