21 CFR 812 describes three types of studies: significant risk (SR) device studies, non-significant risk (NSR) device studies, and exempt studies.

Significant Risk Device Studies are conducted for devices which present a potential serious risk to the health, safety or welfare of a subject. Sponsors of investigational SR device studies must obtain an approved IDE from FDA before starting their study. If FDA disapproves an IDE, they will describe the reasons for disapproval and allow for an IDE amendment to satisfactorily address their concerns. Once both FDA and the IRB have given their approval, the study may begin.

Non-Significant Risk Device Studies are conducted for devices which do not meet the definition of an SR device. If the IRB determines that a proposed study is an NSR study, there is no need to seek IDE approval from FDA before starting the study.

Exempt studies include certain cases which only need to comply with 21 CFR 812 and section 119, not the entire regulation. Studies which do not put subjects at risk, such consumer preference testing, are likely to be exempt.

If an exempt study is being conducted to collect data to support either a clinical investigation or a marketing application, then the study must comply with 21 CFR Part 50 and should comply with 21 CFR Part 56. 21 CFR 50.1(a), 21 CFR 50.20, 21 CFR 56.101(a), 21 CFR 56.103.