Subject Characteristics

There are a number of tables that record non-clinical information about participants. Broadly, these tables can be broken down into the following categories:

  • Participant demographics, including race/ethnicity, educational attainment, and many other measures

  • Family history of AD and dementia

  • Self-reported medical history and medications log

In this section we offer a detailed description of the participant demographics table and provide some general usage notes on the other tables in this category. The tables discussed here can be found on the IDA under the ‘Subject Characteristics’ heading:

Family History of AD/Dementia

As of ADNI4 there are 4 tables that contain information on family history of AD.

There are several differences in the way that this data was stored and organized, the exact information collected in the questionnaire, and the naming conventions for the fields and tables involved.

The data here is divided between phases, with information from ADNI1, ADNIGO, and ADNI2 contained in the Family History Questionnaire (FHQ) table – which records information about parental history of AD – and the Family History Questionnaire Sub-table (RECFHQ) - which records information about siblings.

For the ADNI3 and ADNI4 phases, the family history questionnaire is divided up into two tables - one for information about parents (FAMHXPAR), and one on siblings (FAMHXSIB).

The granularity of information also varies between these two blocks of phases, with the tables in newer phases capturing additional information, including reported age of onset for parent/sibling AD symptoms.

Participant Demographics

The Participant Demographics table (PTDEMOG.csv) is, as the name would suggest, the central repository for demographic information on ADNI participants.

The naming conventions in this table generally follow a predictable format: they are abbreviated portmanteaus, which may seem obtuse at first but there is some logic behind it. For example, participant ethnic identity is captured in the field PTETHCAT, which is short for Participant (PT) Ethnic (ETH) Category (CAT).

Note that all demographic information is self-reported by ADNI participants, and as such there may be some degree of inconsistency. For instance, there are a small number of participants whose self-identified race/ethnicity changes over time.

The most basic demographic information is captured in the PTGENDER, PTRACCAT, PTETHCAT, and PTDOB fields – which give participant gender, self-identified race, self-identified ethnicity (Hispanic/Latino or not), and month/year of birth, respectively.

Note that the coding for the PTRACCAT variable has differed between some phases. In addition, more granular information on racial/ethnic self-identification was made available starting in ADNI4.

There are two fields that record a participant’s preference for English or Spanish, the difference between the two could probably benefit from some clarification:

PTTLANG (ParTicipant Testing LANGuage) records the language used for testing the participant. Note that certain assessments in the ADNI test battery may not have had a Spanish language translation available at the time when a visit was conducted, or the participant’s site may not have had the capacity to administer testing in Spanish.

PTPLANG, on the other hand, records the language that the participant considers to be their primary language.

The participant educational attainment variable (PTEDUCAT) is also deserving of some attention:

<UNDER CONSTRUCTION>

Inithealth

This is the table that records medical history for participants in the ADNI3 and ADNI4 phases. It is listed in the IDA as ‘Initial Health Assessment’, as indicated in the figure below:

At present, the IHDESC (description of condition) and IHCOMM (clinician comments) fields are redacted from the data set pending review for potential PII.

The IHSYMPTOM field can be used to get a general idea of the pre-existing condition, however in general the utility of this table is highly limited by the lack of the free text fields.

MEDHIST and RECMHIST

These two tables record information on self-reported medical history for participants in ADNI1, ADNIGO, and ADNI2, and their location in the IDA is indicated below:

The Medical History (MEDHIST) table contains much of the same information as the IHSYMPTOM field in the INITHEALTH table for later phases, spread across multiple columns rather than coded into a single field.

One notable feature of MEDHIST compared to INITHEALTH is more granular recording of information surrounding participant history of smoking. Information on frequency and duration of smoking is captured in MEDHIST, allowing for the calculation of pack/years and other widely used metrics.

The recent medical history details log (RECMHIST) is currently missing most of the information captured by the table but may prove useful in the future when the free text fields present in the table are reviewed and restored.

Anti-amyloid treatment information

Following the FDA approval of disease-modifying treatments for AD, participants have been allowed to enroll in ADNI while undergoing these treatments. This table identifies those participants and gives specific information on their treatment course.

Since these disease-modifying treatments are formulated to directly affect AD biomarkers, it should be noted that

This table is critically important for researchers who are interested in studying the natural pathology of Alzheimer’s disease.

Other demographics tables

ADNI4 introduced three new instruments that are intended to measure social determinants of health that are not captured in the main demographics table. The results of these instruments are contained in three tables that only apply to participants in the ADNI4 phase, and include:

  • Abbreviated Multidimensional Acculturation Scale (AMAS)

  • RUCC/RUCA measures of rurality

  • Area Deprivation Index (ADI)

Further information on these tables and their contents can be found in the data dictionary.