Project Grant R01AG097722

STATISTICAL METHODS FOR INTEGRATING EVIDENCE AND DATA FOR PRECISION HEALTH MANAGEMENT IN HIGH-RISK ALZHEIMER'S DISEASE POPULATIONS - PROJECT SUMMARY: ALZHEIMER'S DISEASE (AD) IS A MULTIFACTORIAL SYNDROME CHARACTERIZED BY SIGNIFICANT HETEROGENEITY IN ITS PATHOLOGY, CLINICAL PRESENTATION, AND RISK FACTORS. CURRENT PREVENTION AND TREATMENT TRIALS OFTEN FAIL DUE TO INEFFECTIVE TREATMENT, LATE INTERVENTIONS, AND THE INABILITY TO ADDRESS DISEASE HETEROGENEITY, HIGHLIGHTING THE URGENT NEED FOR A PARADIGM SHIFT IN DISEASE PREVENTION STRATEGIES. WE PROPOSE STATISTICAL METHODS FOR ACTIONABLE HEALTH MANAGEMENT STRATEGIES IN HIGH-RISK POPULATIONS BASED ON AVAILABLE TREATMENTS (E.G., ANTIHYPERTENSIVES, ANTIDIABETIC THERAPIES) OF AD MODIFIABLE RISK FACTORS AND COMORBIDITIES (E.G., HYPERTENSION, DIABETES). EFFECTIVE MANAGEMENT OF COMORBIDITY RISK FACTORS, WHICH SHARE BIOLOGICAL PATHWAYS WITH AD, COULD IMPROVE COGNITIVE FUNCTION AND SLOW DISEASE PROGRESSION WELL BEFORE AD BECOMES CLINICALLY APPARENT, THEREBY IMPROVING THE OVERALL EFFECTIVENESS OF DISEASE MANAGEMENT STRATEGIES AND QUALITY OF LIFE. WE WILL LEVERAGE REAL-WORLD DATA (RWD) AND STATE-OF-THE-ART COMPUTATIONAL APPROACHES TO IDENTIFY HETEROGENEOUS TREATMENT EFFECTS (HTES) OF INTERVENTIONS TARGETING THESE MODIFIABLE RISK FACTORS. WE WILL FOCUS ON INTEGRATING EVIDENCE WITH INDIVIDUAL-LEVEL DATA. THERE ARE THREE SPECIFIC AIMS. IN AIM 1 (EVIDENCE SYNTHESIS), WE WILL CONDUCT A SEMI-AUTOMATED SYSTEMATIC REVIEW AND META-ANALYSES OF PUBLISHED STUDIES. WE WILL SYNTHESIZE EVIDENCE ON PROGNOSTIC FACTORS, MODIFIABLE RISK FACTORS, COMORBIDITIES, AND SUBGROUP TREATMENT EFFECTS AND HTES FOR AD COMORBIDITIES. THIS PROCESS WILL GENERATE A HARMONIZED DATABASE OF PROGNOSTIC FACTORS, SUBGROUP-SPECIFIC EFFECTS, AND TREATMENT MODERATORS TO GUIDE SUBSEQUENT ANALYSES. IN AIM 2 (DYNAMIC RISK MODELING), WE WILL DEVELOP CONTINUOUS-TIME HIDDEN MARKOV MODELS (CTHMM) TO CHARACTERIZE THE DYNAMIC PROGRESSION OF AD AND ITS COMORBIDITIES. USING SYNTHESIZED EVIDENCE AND DATA FROM REPRESENTATIVE COHORTS, WE WILL INTEGRATE DIAGNOSTIC MARKERS, CLINICAL MEASURES, AND AUXILIARY MARKERS TO REFINE RISK PROFILES AND IDENTIFY HIGH-RISK INDIVIDUALS FOR TARGETED INTERVENTIONS. IN AIM 3 (DISCOVERING TARGETED STRATEGIES), WE WILL LEVERAGE RWD FROM AD-FOCUSED STUDIES SUCH AS ADNI, NACC, WHICAP, ROSMAP, AND POPULATION-BASED STUDIES SUCH AS ALLOFUS, TO ESTIMATE HTES OF INTERVENTIONS FOR AD COMORBIDITIES IN HIGH-RISK POPULATIONS. WE WILL USE ADVANCED DOUBLE MA-CHINE LEARNING TECHNIQUES AND INCORPORATE SYNTHESIZED EVIDENCE. BY USING MULTIETHNIC COHORTS, WE WILL ENHANCE GENERALIZABILITY TO BROADER POPULATIONS. IN EACH AIM, WE WILL ESTABLISH THEORETICAL PROPERTIES USING MODERN EMPIR-ICAL PROCESS THEORY. BY ADDRESSING THE COMPLEXITY AND VARIABILITY OF AD THROUGH A PRECISION MEDICINE APPROACH, THIS WORK WILL ADVANCE THE IDENTIFICATION OF TAILORED MANAGEMENT STRATEGIES AND IMPROVE CLINICAL DECISION-MAKING. OUR MULTIDISCIPLINARY TEAM, COMPRISING EXPERTS IN BIOSTATISTICS, NEUROLOGY, AND INFORMATICS, IS UNIQUELY POSITIONED TO TRANSLATE THESE FINDINGS INTO ACTIONABLE HEALTH MANAGEMENT STRATEGIES.