A hybrid automated event adjudication system for clinical trials
Bibliographic record
Abstract
INTRODUCTION: In clinical trials, event adjudication is a process to review and confirm the accuracy of outcomes reported by site investigators. Despite efforts to automate the communication between a clinical-data-and-coordination center and an event adjudication committee, the review and confirmation of outcomes, as the core function of the process, still fully rely on human labor. To address this issue, we present an automated event adjudication system and its application in two randomized controlled trials. METHODS: Centrally executed by a clinical-data-and-coordination center, the automated event adjudication system automatedly assessed and classified outcomes in a clinical data management system. By checking clinically predefined criteria, the automated event adjudication system either confirmed or unconfirmed an outcome and automatedly updated its status in the database. It also served as a management tool to assist staff to oversee the process of event adjudication. The system has been applied in: (1) the Cardiovascular Outcomes for People Using Anticoagulation Strategies (COMPASS) trial and (2) the New Approach riVaroxaban Inhibition of Factor Xa in a Global trial versus Aspirin to prevenT Embolism in Embolic Stroke of Undetermined Source (NAVIGATE ESUS) trial. The automated event adjudication system first screened outcomes reported on a case report form and confirmed those with data matched to preset definitions. For selected primary efficacy, secondary, and safety outcomes, the unconfirmed cases were referred to a human event adjudication committee for a final decision. In the New Approach riVaroxaban Inhibition of Factor Xa in a Global trial versus Aspirin to prevenT Embolism in Embolic Stroke of Undetermined Source (NAVIGATE ESUS) trial, human adjudicators were given priority to review cases, while the automated event adjudication system took the lead in the Cardiovascular Outcomes for People Using Anticoagulation Strategies (COMPASS) trial. RESULTS: Outcomes that were adjudicated in a hybrid model are discussed here. The COMPASS automated event adjudication system adjudicated 3283 primary efficacy outcomes and confirmed 1652 (50.3%): 132 (21.1%) strokes, 522 (53%) myocardial infarctions, and 998 (59.7%) causes of deaths. The NAVIGATE ESUS one adjudicated 737 cases of selected outcomes and confirmed 383 (52%): 219 (51.5%) strokes, 34 (42.5%) myocardial infarctions, 73 (54.9%) causes of deaths, and 57 (57.6%) major bleedings. After one deducts the time needed for migrating the system to a new study, the automated event adjudication system helped to reduce the time required for human review from approximately 1303 to 716.5 h for the Cardiovascular Outcomes for People Using Anticoagulation Strategies trial and from 387 to 196 h for the New Approach riVaroxaban Inhibition of Factor Xa in a Global trial versus Aspirin to prevenT Embolism in Embolic Stroke of Undetermined Source trial. CONCLUSION: The automated event adjudication system in combination with human adjudicators provides a streamlined and efficient approach to event adjudication in clinical trials. To immediately apply automated event adjudication, one can first consider the automated event adjudication system and involve human assistance for cases unconfirmed by the former.
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 | 0.044 | 0.123 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.003 | 0.001 |
| Bibliometrics | 0.008 | 0.006 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.008 | 0.005 |
| Open science | 0.005 | 0.006 |
| Research integrity | 0.003 | 0.003 |
| Insufficient payload (model declined to judge) | 0.034 | 0.023 |
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".