When were IIDP participants enrolled?
IIDP participants were enrolled at two different time points (1992 and 2001). In 1992, 2,212 African Americans and 2,494 Nigerians were enrolled in the study and underwent longitudinal follow-up evaluations every 2 or 3 years. In 2001, 1,892 additional African Americans and 1,939 Nigerians were enrolled in the study and were longitudinally followed.
- What is two-stage sampling?
Two-stage sampling is a sampling method often used to identify disease cases when the rate of the disease is relatively low and when clinical evaluations are intensive. The first stage is a screening stage where a screening instrument with high sensitivity for detecting the disease is administered to all individuals. A stratified random sampling is used at the second stage with greater sampling probability for individuals whose screening results indicate greater probability of disease.
- How was two-stage sampling carried out in the Indianapolis-Ibadan Dementia Project?
At each evaluation wave, all participants were first evaluated with the CSID administered by trained interviewers at participants’ homes. A cognitive score and an informant score based on interviews with a family member, if available, were calculated based on results from the screening interview. Participants were divided into three groups based on screening performance: good, intermediate and poor. All individuals (100%) in the poor group were invited to undergo extensive clinical assessment; 50% (75% for some waves) in the intermediate group were invited for clinical assessment; 5% in the good group were randomly selected to undergo clinical assessment.
For sampling plans used at different study waves, please see “Sampling Plan” in Documentation.
- Did the methods of data collection and data entry systems change from one wave to another?
The methods of data collection did not change throughout the years. All data were collected on paper by trained interviewers, and verified for completion at the conclusion of each interview by the interviewer and the project supervisor.
The data entry systems evolved throughout the years. Secure, password protected desktop custom applications were developed using various software. Initial database was implemented in Fox Pro2.0 in 1992 and was migrated to Visual FoxPro in 1997.
- Are there changes in the information being collected?
The majority of study data including cognitive assessment and functional assessment remained the same throughout the study allowing longitudinal analyses of these outcomes. Thus, the majority of variables’ names and formats remained consistent from one wave to another. If a question needed to be rephrased and the meaning changed slightly, or if a new item was introduced to the questionnaires, then a new variable was created to reflect accurately the meaning of the data intended to collect.
- How does each dataset relate to the data collections forms posted online?
Datasets and data collection forms are grouped into three categories: screening, clinical assessment and neuropsychological tests. Most of the data collection forms are marked by the year of evaluation. Datasets are combined across all waves of data with a variable “wave” to indicate the wave of data collection.
- How similar are the values of the variables included in the downloadable datasets to the answer options listed in the data collection forms?
In the process of aggregating data from each wave into a longitudinal dataset, we had to make decisions that allowed us to harmonize the data. Some questions were removed due to privacy concerns, other items needed to be re-formatted. Each variable included in the dataset is well documented in the data dictionary.
- Do the datasets include any derived variables that I cannot find on data collection forms?
Each dataset includes some derived variables and these are documented in the data dictionary. The reasons for creating these variables vary, and they are typically one of the following:
De-identification of the dataset (i.e.: age at visit)
Harmonization of the data (i.e.: wave the denote data collected across waves)
Calculation of Summary Scores needed for analysis and for making decisions about the cognitive impairment of subjects and their ability to continue participation in the study. (i.e.: cogscore)
- Can I link this dataset with genetic data? How do I do that?
Genome-Wide Association Study (GWAS) from 1,345 African Americans was completed by the Broad Institute yielding 823,561 SNPs. These data have been included in the Alzheimer Disease Genetic Consortium (https://www.niagads.org/resources/related-projects/alzheimers-disease-genetics-consortium-adgc-collection) along with phenotype data for comparing AD cases and normal controls. GWAS for the Ibadan samples were completed with funding from the Center for Inherited Disease Research providing genotyping for 5 million SNPs. The Ibadan GWAS data were included in dbGAP (https://www.ncbi.nlm.nih.gov/gap/). Whole genome sequencing has been conducted for the African American samples and is currently underway for the Yoruba samples.
