Professors from the Donald Bren School of Information and Computer Sciences (ICS) are teaming up with researchers from UCI’s Alzheimer’s Disease Research Center (ADRC) on a new five-year multi-PI grant from the National Institutes of Health (NIH). The $1.87 million grant, “Statistical Methods for Alzheimer’s Research,” is co-led by Chancellor’s Professor Bin Nan in Statistics (contact principal investigator) and Chancellor’s Professor and Statistics Department Chair Daniel Gillen. The work is a collaborative effort with ADRC Clinical Core Director David Sultzer and ADRC 90+ Core Director Maria Corrada-Bravo, both grant co-investigators.

The collaboration first started when Nan and Gillen received an ICS internal “exploration award” research grant of $75,000, which supported their efforts to develop the NIH proposal and obtain pilot data. Based on their research into aging and Alzheimer’s disease (AD), they were looking for ways to resolve statistical issues encountered in cohort studies of underrepresented and overall U.S. aging populations.

“The long-term objective of our proposed research,” explains Nan, “is to produce reliable inference and estimation procedures for characterizing AD and related dementia rates across [a person’s] lifespan.” Gillen adds that they also hope to find ways to identify “risk factors for AD among underrepresented populations using multiple longitudinal cohorts in addition to electronic health records [EHR] data that have imperfect collection protocols.” Their ADRC collaborators will be serving as subject matter experts and will be providing access to — and helping analyze — a wide range of datasets.

The team will produce novel predictive models for recurrent events when there are possible missing events. They will also produce a new survival analysis methodology that is robust to common assumptions and differential missingness patterns. As outlined in the grant abstract, these new methods will let researchers validly “investigate AD events during the entire lifespan, which provides a clearer picture of the relationships among AD onset, death, other life events, and risk factors [and] thus a better understanding of AD etiology and prevention.”

The team plans to develop publicly available statistical software for dissemination and generalization. The methods will likely apply to many other research fields for different diseases captured in EHR data.

— Shani Murray