The case report form is a work of art.

The first principle of CRF design is to stop thinking about the form as the unit of design. The unit is the question. Every individual field — its label, its unit of measurement, its default value, the order of items in its picklist, whether it permits free text, whether it requires confirmation — is a small UX decision. A form with 200 fields is 200 design decisions. The form's overall quality is the sum.

This sounds tautological, but the consequences are operational. Most CRF tooling encourages designers to think at the form level — "the demographics page," "the labs page," "the AE page" — and the tooling presents fields as a flat list. The result is forms where the labels are inconsistent across pages, the units shift between metric and imperial without warning, the picklists are alphabetical when they should be frequency-ordered, and the defaults are blank when they should be the most common answer. Each individual choice would be easy to fix if anyone were looking at it. No one is, because the convention is to look at the form.

Designing at the question level requires a different posture. It requires sitting with each field, asking what the coordinator will see and do at that field, and making the small decision well. It also requires the willingness to be slow on the field-level work — and the trust that the slowness pays back across thousands of entries. The slow work is the work.

The default value of a CRF field is the single most underestimated design decision on the form. Every time a coordinator opens a field, the default tells them what answer the CRF expects. If the default is wrong, the default is incorrect training delivered at every visit, on every subject, for the life of the study. Coordinators correct defaults less often than designers assume, especially under time pressure between patient appointments.

This is why blank defaults — fields that load empty — are usually wrong. A blank field forces the coordinator to make a positive choice, which sounds rigorous but in practice means the choice is made with whatever cognitive resources are available in that moment. A field defaulted to the most common answer (or to "n/a" with explicit override) channels the coordinator's attention to where their judgment is actually needed, and away from the fields where their judgment is just confirming the obvious. The good default is the one that makes the rare choices visible by contrast.

The third principle is the one most often ignored. Modern EDC platforms support field-level validation logic — range checks, conditional logic, cross-field consistency rules — and most of this capability goes unused. The reason, usually, is that the validation logic was deferred to "the SOP" or to "monitoring queries." Both deferrals are mistakes.

SOPs are read once at training and consulted irregularly thereafter. Coordinators do not pause every entry to consult the SOP. Monitoring queries arrive weeks after the entry, by which time the patient has gone home and the source data may or may not be retrievable. The window where bad entries can be caught at zero cost is the moment of entry. After that moment, the cost of catching the same error rises monotonically — first to a query, then to a deviation, then to a missing data point at lock.

Validation logic in the form catches errors at the moment they occur and gives the coordinator the chance to correct them while the source is still in front of them. This includes simple range checks (lab values that fall outside biological possibility), cross-field logic (visit dates that precede consent dates), and conditional logic (fields that should not be answered for the cohort the subject is enrolled in). None of this is exotic. All of it is undone, on most CRFs, because the work was deferred to a system that runs after the entry.

The fourth principle is a position. CRO templates — for CRFs, for monitoring plans, for SOPs — are conventionally treated as proprietary IP. They are guarded as commercial advantage, and CROs that have invested in them tend to keep them out of sight of clients and competitors alike.

We think this is wrong. The template is not the advantage. The advantage is the discipline that produced it and the discipline to keep it sharp. A CRO that publishes a strong CRF template gives away the artifact and keeps the muscle. A competitor who copies the template inherits the form but not the practice — and forms that are not maintained against new study contexts decay quickly.

The deeper reason is that the publication identity of a serious specialty CRO is built on the work it shows, not on what it withholds. Showing the template signals that the team has thought about CRFs at the question level, that the discipline is real, and that the team can be trusted to bring this discipline to a sponsor's study. Withholding it signals the opposite. Confidence is a marketing message that secrecy can never produce.