Evaluating social and clinical risk factors for the progression from Mild CognitiveImpairment to Alzheimer’s Disease and Related Dementias using Real World ClinicalData

PROJECT SUMMARY Uncovering factors influencing progression from Mild Cognitive Impairment (MCI) to Alzheimer’s Disease and Alzheimer’s Disease Related Dementias (AD/ADRD) represents a valuable opportunity to prevent clinical AD/ADRD and prolong functional independence of people along the AD/ADRD continuum. Identifying this

progression is particularly difficult due to the diverse clinical criteria used to diagnose MCI and extensive underdiagnosis leaving a nebulous path for researchers attempting to learn from its phenotypic characteristics. Leveraging informatics tools to identify MCI cases in Electronic Medical Records (EMR) offers a great

opportunity to mitigate the heterogeneity and underdiagnosis of MCI. Risk factors exacerbating the progression from MCI to AD/ADRD have been previously studied in the context of clinical measures. However, few have included neighborhood context measures as risk factors, which may be key to understanding the higher

AD/ADRD incidence rates among underrepresented populations. The objective of this F31 proposal is to equip me with the necessary skills to pursue a career as an independent investigator using clinical data and other large, complex information systems to identify opportunities to prevent and delay progression to AD/ADRD. To

achieve this goal, I will use longitudinal data from a large, diverse EMR system to identify MCI cases using structured and unstructured data to evaluate whether neighborhood measures of resource availability and indicators of cardiovascular risk factors predict time to progression from MCI to AD/ADRD. Aim 1 will identify

patients with MCI diagnosis via diagnostic codes and MCI early symptoms in clinical notes using EMR data from the University of California, San Francisco (UCSF) through the construction of a specific natural language processing (NLP) algorithm. This will be validated by comparing association with future AD/ADRD diagnosis

and ground-truth chart review on a selected sample. Aim 2 evaluates the independent and mediated effects of neighborhood deprivation indices and cardiovascular disease (CVD) risk factors on time to progression from MCI to AD/ADRD. Independent effects will be evaluated with time-varying-models of repeated measures of

CVD factors and neighborhood deprivation indices. Mediation will be used to determine whether place-level deprivation effects are mediated by CVD risk burden on time to progression from MCI to AD/ADRD. The proposed research contributes to the NIA goals to improve assessment of MCI and AD related conditions and

understand the effects of societal factors and health disparities on AD/ADRD. Under the guidance of a highly qualified mentorship team, my training aims will augment my research aims by enhancing my understanding of MCI and AD/ADRD clinical processes and phenotypes, provide training on causal inference methods using

longitudinal data with integration to machine learning models. With my highly committed mentorship team and training environment at UCSF, I will succeed in my F31 proposed research, training plan, and acquire the skills to emerge as an accomplished, independent clinical translational researcher in the field of AD/ADRD.