Identifying Site‐Level Data Integrity Risks in Alzheimer’s Trials via Subject‐Level Heuristics and a Novel Machine Learning Framework
Bibliographic record
Abstract
BACKGROUND: Undetected data integrity issues in Alzheimer's disease (AD) trials often appear as "paradoxical" participant profiles with atypical or implausible profiles that can distort efficacy signals and threaten the validity of study conclusions. A unique mathematically-augmented machine learning (MAML) framework was used to identify high-risk study sites based on patterns in subject-level medical history data, despite a small sample size of trial sites. METHODS: Using data from a trial evaluating treatment for mild to moderate AD, 1855 subject-level variables were transformed into 7816 site-level features by computing distributional summaries of subject medical data (min, mean, max) across 36 sites. Sponsor-provided site-risk labels were used to train classification models including large language models and several standard ML methods such as Random Forest and XGBoost (LLM/ML) in three scenarios: (1) LLM/ML on all raw features (2) LLM/ML on MAML-derived features (3) LLM/ML on MAML-derived features and learned site relabeling. Leave-one-out cross-validation (LOOCV) assessed accuracy, sensitivity, specificity, and AUC due to the small sample size constraints. RESULTS: Scenario 3 achieved the highest predictive performance:XGBoost yielded an AUC of 0.99 (accuracy 0.89, sensitivity 0.93, specificity 0.86). In contrast, Scenario 1 trained with all features peaked at an AUC of 0.25, underscoring the value of MAML's data-driven relabeling to uncover latent site-level risks. The MAML framework grouped all five known high-risk sites and nine additional sites into a high-risk group defined by 10 key variables, predominately anxiety-related. High-risk sites consistently displayed statistical deviations from the canonical AD symptomatology structure. CONCLUSION: Pre-randomization, baseline subject data, can effectively flag site integrity risks in AD clinical trials by detecting deviations from the standard spectrum of symptomologies via MAML. This deviation can be used to better evaluate clinical trial sites. Incorporating these insights into ongoing and future AD trials may enhance data quality, mitigate bias, and reinforce confidence in trial outcomes. IMPACT: Due to AD's clinical heterogeneity and overlap with neuropsychiatric symptoms, our scalable approach offers early detection of anomalous trial behavior, supporting more trustworthy efficacy assessments across AD and related psychiatric trials.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.026 | 0.040 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.002 | 0.002 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".