Enrollment

One of the first tasks that users will need to take on in any study involving ADNI data is the selection of a participant cohort. The tables present under the ‘Enrollment’ subheading are particularly useful for this purpose. This section is also home to the registry, a particularly noteworthy table for users who are working with data drawn from multiple disparate files.

Participant identifiers are assigned at screening, however not every participant who undergoes a screening visit ends up enrolled in the ADNI study. There are a few reasons why this could be this case, including:

Failure to meet all inclusion criteria
Meeting some exclusion criteria
Satisfying all eligibility criteria for a particular diagnostic arm for which the site is not actively enrolling participants
Loss to follow-up between screening and baseline

Inclusion/exclusion information for ADNI2 and ADNIGO is stored in the clinician verification (CLIELG) table, and the study visits summary (STUDYSUM) table houses that information for ADNI3 and ADNI4.

Note that these general inclusion/exclusion criteria should not be confused with the cohort-specific inclusion/exclusion criteria detailed [here
][screening visits and diagnostic cohorts]Excerpts from the procedures manuals detailing the specific inclusion/exclusion criteria for each phase are included below:

A great deal of data, including demographic information and the results of many basic clinical evaluations, are collected at screening and made available regardless of whether a participant enrolls in the study and completes a baseline visit. This should be taken into account depending on the nature of the projects, especially in demographic summary statistics.

In general, there is no straightforward and consistent way to determine what happens to a participant between screening and baseline. The most direct criteria for determining whether a participant is enrolled in the study is the existence of a baseline visit in the registry. For more granular information, the exact recording and formatting of data has varied substantially between phases.

ADNI1 Inclusion and exclusion (INCLUSION and EXCLUSION)

There are two separate inclusion and exclusion tables for ADNI1 (INCLUSION and EXCLUSION), as well as a unique VISCODE for identifying screen fails.

Identifying screen fails in the ADNI1 phase is straightforward. Screening visits where participants were deemed eligible to enroll are associated with the VISCODE entry ‘sc’, while screening visits that resulted in a screen failure are identified with an entry of ‘f’ in the VISCODE field.

Clinician verification (CLIELG)

For every phase of the study, the CENROLL field of this table indicates whether or not a participant was deemed eligible for enrollment. Note that some participants may still be allowed to enroll via protocol exception, and such cases can be identified by cross-referencing the registry for the existence of a baseline visit.

For the ADNIGO and ADNI2 phases specifically, this table contains additional, more specific information on inclusion/exclusion criteria in the INCLUSION, EXCLUSION, FAILINCLU, and FAILEXCLU fields. These fields were not used in ADNI3 or ADNI4.

Note that this table may not be completely comprehensive, and some participants may not be accounted for.

Study visits summary (STUDYSUM)

The fields from the clinician verification table that track inclusion/exclusion criteria for the ADNIGO and ADNI2 phases are largely mirrored here.

Early discontinuations and withdrawals (TREATDIS)

This table contains data from the ADNI1, ADNIGO, and ADNI2 phases. It is a log of discontinuations and withdrawals, which can capture some cases of participants choosing to withdraw from the study between screening and baseline.

The consent tracking table (CONSENTS) serves a similar function for ADNI4.

Registry