General Information
Conflict of Interest
Request Clinical Trials
Associated Trial(s):- NCT00679627 - A Randomized, Double-Blind, Placebo-controlled Trial of Long-term (2-year) Treatment of Galantamine in Mild to Moderately-Severe Alzheimer's Disease
- NCT00253214 - Placebo-Controlled Evaluation of Galantamine in the Treatment of Alzheimer's Disease: Safety and Efficacy of a Controlled-Release Formulation
- NCT00253188 - Efficacy, Tolerability and Safety of Galantamine in the Treatment of Alzheimer's Disease
- NCT00253227 - Galantamine in the Treatment of Alzheimer's Disease: Flexible Dose Range Trial
- NCT00253201 - Efficacy, Tolerability and Safety of Galantamine in the Treatment of Alzheimer's Disease
- NCT00236431 - A Randomized Double-Blind Placebo-Controlled Trial to Evaluate the Efficacy and Safety of Galantamine in Patients With Mild Cognitive Impairment (MCI) Clinically at Risk for Development of Clinically Probable Alzheimer's Disease
- NCT00236574 - A Randomized Double Blind Placebo-Controlled Trial to Evaluate the Efficacy and Safety of Galantamine in Patients With Mild Cognitive Impairment (MCI) Clinically at Risk for Development of Clinically Probable Alzheimer's Disease
- NCT00261573 - The Safety and Efficacy of Galantamine in the Treatment of Vascular and Mixed Dementia
- NCT00574132 - A Phase 3, Multicenter, Randomized, Double-Blind, Placebo-Controlled, Parallel-Group, Efficacy and Safety Trial of Bapineuzumab (AAB-001, ELN115727) In Patients With Mild to Moderate Alzheimer's Disease Who Are Apolipoprotein E4 Non- Carriers
- NCT00575055 - A Phase 3, Multicenter, Randomized, Double-Blind, Placebo-Controlled, Parallel-Group, Efficacy and Safety Trial of Bapineuzumab (AAB-001, ELN115727) In Patients With Mild to Moderate Alzheimer's Disease Who Are Apolipoprotein E4 Carriers
Request Clinical Trials
Data Request Status
Status: OngoingResearch Proposal
Project Title: Retrospective Trial-to-Cohort Study of AD/MCI Trial Generalizability for Functional Decline and Mortality
Scientific Abstract: Background: Alzheimer disease (AD) and mild cognitive impairment (MCI) trials often enroll selected participants, but clinicians must apply trial evidence to older adults with multimorbidity, functional limitation, and variable cognitive vulnerability. Objective: To evaluate how well selected YODA AD/MCI trial populations represent real-world older adults and to identify baseline cognitive and functional profiles associated with clinically relevant outcomes. Study design: Retrospective secondary analysis of participant-level randomized trial data linked analytically, not physically, with harmonized longitudinal aging cohorts. Participants: Adults enrolled in requested YODA AD/MCI trials, and comparable older adults from harmonized cohort datasets used as external target-population comparators. Main and secondary outcome measures: The primary outcome will be functional decline or functional limitation progression when available. Secondary outcomes include cognitive change, death or dropout, adverse-event burden, and trial eligibility or representativeness metrics. Statistical analysis: We will harmonize baseline cognition, function, demographics, and comorbidity variables; describe trial-to-cohort differences; estimate trial eligibility prevalence in cohorts; model outcome risks using regression and time-to-event or repeated-measures methods as appropriate; and use weighting or transportability analyses to assess how trial results may generalize to broader older populations. Analyses will be de-identified, aggregate-reported, and observational for generalizability questions.
Brief Project Background and Statement of Project Significance:
Randomized trials provide high internal validity, but their clinical value also depends on whether participants resemble the patients seen in routine care. This issue is especially important for AD and MCI, where age, frailty, multimorbidity, baseline functional limitation, caregiver availability, and cognitive severity may affect both eligibility and outcomes. Older adults with cognitive vulnerability are frequently heterogeneous, and trial exclusion criteria may limit the applicability of findings to populations most affected by dementia-related disability.
The proposed project will use YODA participant-level trial data to create new, generalizable knowledge in three ways. First, it will quantify the baseline cognitive, functional, demographic, and comorbidity profiles of AD/MCI trial participants. Second, it will compare these profiles with harmonized longitudinal aging cohorts representing broader older populations. Third, it will evaluate whether baseline cognitive vulnerability and functional status predict functional decline, cognitive change, death/dropout, and adverse-event burden within trial data. This is not a re-analysis designed only to confirm original efficacy claims. Its purpose is to clarify external validity: to whom the trial evidence most directly applies, which patient groups may be underrepresented, and what design features could make future AD/MCI trials more inclusive and clinically interpretable.
The requested data are necessary because published aggregate reports cannot support participant-level eligibility mapping, baseline phenotype harmonization, or outcome prediction across cognitive and functional subgroups. Results could guide clinicians interpreting AD/MCI trial findings for older adults with disability and multimorbidity, and help investigators design future trials with eligibility criteria and outcomes that better reflect public-health needs. Relevant prior work has emphasized the importance of external validity in randomized trials and documented how restrictive eligibility criteria can reduce representativeness, including in Alzheimer disease and related dementia trials (Rothwell, Lancet 2005; Van Spall et al., JAMA 2007; Lam et al., 2024).
Specific Aims of the Project:
Aim 1: Describe the baseline characteristics of participants in selected YODA AD/MCI trials, including age, sex, education when available, cognitive severity, functional status, comorbidity burden, and key eligibility-related variables.
Hypothesis 1: Trial participants will differ from real-world older adults with comparable cognitive vulnerability, particularly in multimorbidity, baseline function, and severity profiles.
Aim 2: Quantify the external representativeness of YODA AD/MCI trials by applying major trial eligibility features to harmonized aging cohorts and estimating the proportion and characteristics of cohort participants who would be trial-like.
Hypothesis 2: Common eligibility features will exclude a meaningful proportion of older adults with cognitive vulnerability, especially those with functional limitation or comorbidity.
Aim 3: Evaluate whether baseline cognitive and functional profiles predict functional decline, cognitive change, death/dropout, and adverse-event burden within the requested trial datasets.
Hypothesis 3: Baseline cognitive vulnerability combined with functional limitation will identify participants at higher risk of clinically relevant decline and attrition, supporting the need for functional and multimorbidity-aware interpretation of AD/MCI trial evidence.
Study Design: Meta-analysis (analysis of multiple trials together)
What is the purpose of the analysis being proposed? Please select all that apply.: Participant-level data meta-analysis Meta-analysis using data from the YODA Project and other data sources Research on clinical prediction or risk prediction
Software Used: R, STATA, Open Office
Data Source and Inclusion/Exclusion Criteria to be used to define the patient sample for your study:
YODA trial data: The requested sample will include all randomized participants in selected YODA trials of Alzheimer disease (AD) or mild cognitive impairment (MCI) for whom participant-level baseline data are available. We will include both intervention and comparator/placebo groups because the main purpose is representativeness and prognosis, not replication of the original efficacy analysis. Inclusion criteria for the analytic sample will be: enrollment in an eligible AD/MCI trial; available baseline age and sex; available baseline cognitive measure or diagnosis/stage variable; and at least one baseline functional, comorbidity, or safety variable relevant to eligibility mapping. For outcome analyses, participants must also have the relevant post-baseline outcome or censoring/dropout information.
Exclusion criteria: participants without usable baseline demographic information; participants not belonging to the AD/MCI target condition for the requested trial; participants with no baseline cognitive/stage information; and participants with no usable follow-up for the specific outcome model. We will not exclude participants by treatment arm, race/ethnicity, age, comorbidity, or baseline functional limitation unless those data are unavailable for the specific analysis.
Additional data: We will use harmonized aging cohort data from SHARE, CHARLS, ELSA, HRS, KLOSA, LASI, and MHAS as external target-population comparators. These datasets are obtained from their official public or approved-access repositories under each cohort's terms of use. We will not link individuals across YODA and cohort datasets. YODA participant-level analyses will be conducted only within the secure YODA platform. External cohort analyses will be conducted as aggregate target-population summaries, or de-identified derived datasets will be uploaded only if permitted by cohort data-use agreements and approved by YODA.
Primary and Secondary Outcome Measure(s) and how they will be categorized/defined for your study:
Primary outcome: Functional decline during trial follow-up. This will be defined from each trial's prespecified functional measure, using the protocol, data dictionary, CSR, and published endpoint definitions. If a continuous functional scale is available, such as ADL/IADL, ADCS-ADL, Disability Assessment for Dementia, or a comparable measure, the outcome will be change from baseline to the main endpoint or final available visit. If item-level function data are available, functional decline will be defined as new functional limitation among participants without baseline limitation, or worsening functional burden among those with baseline limitation.
Secondary outcomes: (1) Cognitive change, defined as change from baseline to follow-up on the trial's prespecified cognitive scale, such as ADAS-Cog, MMSE, CDR-SB, or another available measure. Repeated visits will be modeled when available; otherwise baseline-to-endpoint change will be used. (2) Attrition, defined as death, withdrawal, loss to follow-up, or early discontinuation, analyzed separately when possible and as a composite when needed. (3) Adverse-event burden, defined as serious adverse events, any adverse events, and discontinuation due to adverse events when consistently available. (4) Representativeness outcomes, defined as eligibility prevalence in external aging cohorts, standardized baseline differences between trial and cohort populations, and trial-likeness or sampling-score overlap.
Outcome definitions will be finalized before analysis and aligned with original trial documents when possible. Any final publication changes will be limited to unavailable or non-comparable variables and will be explicitly reported.
Main Predictor/Independent Variable and how it will be categorized/defined for your study:
Primary independent variable: Baseline cognitive-functional vulnerability. This will be defined a priori from baseline cognitive severity and baseline functional status. Cognitive vulnerability will be classified using trial diagnosis/stage variables and baseline cognitive scale severity where available, for example MCI, mild AD, moderate AD, or scale-based impairment categories. Functional vulnerability will be defined by baseline ADL/IADL limitation, dependence, disability score, or clinically comparable functional scale impairment.
The main exposure classification will be a four-level phenotype: (1) lower cognitive vulnerability without functional limitation; (2) cognitive vulnerability without functional limitation; (3) functional vulnerability without higher cognitive vulnerability; and (4) combined cognitive and functional vulnerability. If trial data do not support all four categories, the primary exposure will be simplified to combined cognitive-functional vulnerability versus no combined vulnerability, with the exact rule documented before modeling.
For eligibility and representativeness analyses, the main independent variables will also include trial-like eligibility features: age range, cognitive-stage range, baseline function threshold, major comorbidity exclusions, medication or contraindication exclusions when measurable, and baseline severity requirements. Treatment assignment will not be the main predictor for the primary generalizability aims, but exploratory analyses may examine whether treatment effects or outcome trajectories differ by vulnerability phenotype when allowed by sample size and trial design.
Other Variables of Interest that will be used in your analysis and how they will be categorized/defined for your study:
Other variables will be selected a priori based on clinical relevance, trial design, and a causal diagram for baseline vulnerability, trial participation, and outcomes. Demographic variables will include age, sex, race/ethnicity when available, country or region when available, education, and living arrangement or caregiver availability if recorded. Clinical variables will include baseline diagnosis or cognitive stage, baseline cognitive score, baseline functional score, body mass index if available, vascular risk factors, diabetes, hypertension, heart disease, stroke, kidney disease, depression or psychiatric history, number of chronic conditions, and concomitant medication classes when available.
Trial design variables will include trial identifier, treatment arm, randomization stratum, study visit time, follow-up duration, and protocol-defined eligibility features. Safety variables will include baseline risk factors relevant to adverse events when available. External cohort variables used for target-population comparison will be harmonized to the closest common definitions: age, sex, education, cognitive status, functional limitation, multimorbidity, vascular disease, depression, follow-up time, death, and functional progression.
We will not use stepwise selection or adjust for all statistically significant variables. Primary adjustment sets will be prespecified: age, sex, baseline cognitive severity, baseline function, comorbidity burden, and trial/study indicators as appropriate. Expanded models will add education, race/ethnicity, depression, and vascular disease when available and conceptually justified.
Statistical Analysis Plan:
All YODA participant-level analyses will be performed within the secure YODA platform. The analysis will begin with data inventory and harmonization using trial data dictionaries, protocols, CSRs, and annotated case report forms when available. We will create a reproducible variable map for cognition, function, demographics, comorbidity, safety, treatment arm, follow-up, and eligibility criteria.
Descriptive analyses will summarize baseline characteristics of trial participants overall and by trial, treatment arm, cognitive stage, and cognitive-functional vulnerability phenotype. We will report means or medians for continuous variables and counts and percentages for categorical variables. Missingness will be summarized by variable, trial, and outcome sample.
For Aim 1, we will describe the baseline cognitive, functional, demographic, and comorbidity profile of YODA trial participants. Between-group comparisons will use standardized mean differences rather than relying only on P values. Trial-level estimates will be reported separately and, where appropriate, summarized across trials using random-effects meta-analysis.
For Aim 2, we will compare YODA trial participants with harmonized external aging cohorts. We will apply measurable trial eligibility features to cohort data to estimate eligibility prevalence and to describe the older adults excluded by common criteria. We will estimate trial participation or trial-likeness scores using prespecified baseline variables: age, sex, cognitive severity, baseline function, multimorbidity, and major vascular or psychiatric conditions. These scores will be used diagnostically to quantify overlap between trial and target populations. If sufficient overlap exists, we will use standardization or inverse probability of sampling weights to estimate how selected trial outcomes or treatment contrasts might generalize to the broader target population. If overlap is poor, we will report nontransportability rather than force a weighted estimate.
For Aim 3, functional decline will be analyzed using linear regression or mixed-effects models for continuous functional scales and modified Poisson, logistic, or time-to-event models for binary or incident functional outcomes, depending on trial measurement structure. Cognitive change will be analyzed using repeated-measures mixed models or change-score models aligned with original trial endpoints. Death, dropout, and time-to-event outcomes will be analyzed using Kaplan-Meier summaries and Cox models when dates are available, or logistic/Poisson models when only interval status is available. Adverse-event outcomes will be summarized as risks or rates and modeled with generalized linear models when appropriate.
Primary multivariable models will adjust for prespecified covariates, not data-driven variable selection. Potential confounders include age, sex, baseline cognitive severity, baseline function, education, comorbidity burden, vascular disease, depression, and trial identifier. Conceptually, baseline cognitive and functional vulnerability may influence outcomes directly and may also modify the applicability of treatment evidence. We will therefore examine interactions between vulnerability phenotype and treatment arm, age group, sex, and baseline comorbidity if sample size permits. Subgroup analyses will include cognitive stage, age group, sex, baseline functional limitation, and multimorbidity burden. Sensitivity analyses will test alternative definitions of cognitive vulnerability, alternative functional decline thresholds, complete-case versus missing-indicator or multiple-imputation approaches if feasible, placebo-arm-only prognostic models, and exclusion of participants with very short follow-up. Results will be reported as estimates with 95% confidence intervals, emphasizing magnitude, uncertainty, and generalizability rather than statistical significance alone.
Narrative Summary: This retrospective secondary analysis will use participant-level data from selected YODA Alzheimer disease and mild cognitive impairment trials and harmonized longitudinal aging cohorts. We will compare trial participants with real-world older adults who have similar cognitive vulnerability, quantify how trial eligibility represents the broader population, and examine whether baseline cognition and function predict functional decline, cognitive change, and mortality. We will not attempt to re-identify participants. The project can clarify how well trial evidence applies to older adults with multimorbidity and functional limitations, helping clinicians and policy makers interpret AD/MCI trial results and design more inclusive future studies.
Project Timeline:
Estimated project start: within 2 weeks after YODA data access approval and completion of platform onboarding.
Months 0-2: Review protocols, CSRs, data dictionaries, and case report forms; build the harmonized variable map; define trial-specific eligibility and outcome algorithms.
Months 2-4: Complete data cleaning, missingness assessment, baseline phenotype derivation, and trial-to-cohort variable harmonization.
Months 4-7: Conduct descriptive, representativeness, eligibility, prognostic, and sensitivity analyses within the secure platform.
Months 7-9: Prepare tables, figures, internal quality checks, and interpretation of trial-specific and pooled findings.
Months 9-10: Draft manuscript and supplementary materials.
Months 10-11: Complete coauthor review, revision, and journal submission.
Month 12: Submit required report, analysis documentation, and publication status update to the YODA Project. If additional analyses are requested by reviewers, we may request an extension under the data use agreement.
Dissemination Plan:
Expected products include one primary peer-reviewed manuscript, a supplementary methods and harmonization appendix, and conference abstract submissions focused on AD/MCI trial generalizability, functional decline, cognitive change, attrition, and adverse-event burden in older adults. The main target audiences are clinicians caring for older adults with AD or MCI, dementia trialists, neurologists, geriatricians, epidemiologists, health-services researchers, public-health stakeholders, and investigators designing more representative AD/MCI clinical trials.
The primary manuscript will report participant-level YODA trial analyses, trial-to-cohort representativeness comparisons, and sensitivity analyses in a transparent observational framework. The supplementary appendix will include trial-specific variable harmonization, outcome definitions, eligibility mapping rules, missingness assessment, and additional subgroup or sensitivity results.
Potential journals for submission include JAMA Network Open, Alzheimer's & Dementia, Neurology, Age and Ageing, Journal of the American Geriatrics Society, BMC Medicine, or a comparable peer-reviewed journal depending on final scope and findings. We will not share participant-level YODA data outside the secure platform. Publications and presentations will acknowledge the YODA Project and requested trials according to YODA requirements, and any public communication will occur only after compliance with the data use agreement.
Bibliography:
1. Rothwell PM. External validity of randomised controlled trials: “to whom do the results of this trial apply?” Lancet. 2005;365:82-93.
2. Van Spall HGC, Toren A, Kiss A, Fowler RA. Eligibility criteria of randomized controlled trials published in high-impact general medical journals: a systematic sampling review. JAMA. 2007;297:1233-1240.
3. Lam J, et al. Analysis of eligibility criteria in Alzheimer’s and related dementias clinical trials. 2024.
4. Cole SR, Stuart EA. Generalizing evidence from randomized clinical trials to target populations: the ACTG 320 trial. Am J Epidemiol. 2010;172:107-115.
5. Stuart EA, Cole SR, Bradshaw CP, Leaf PJ. The use of propensity scores to assess the generalizability of results from randomized trials. J R Stat Soc Ser A Stat Soc. 2011;174:369-386.
6. Hernán MA, Robins JM. Causal Inference: What If. Boca Raton: Chapman & Hall/CRC; 2020.