Optimizing the efficiency and effectiveness of data quality assurance in a multicenter clinical dataset
Bibliographic record
Abstract
OBJECTIVES: Electronic health records (EHRs) data are increasingly used for research and analysis, but there is little empirical evidence to inform how automated and manual assessments can be combined to efficiently assess data quality in large EHR repositories. MATERIALS AND METHODS: The GEMINI database collected data from 462 226 patient admissions across 32 hospitals from 2021 to 2023. We report data quality issues identified through semi-automated and manual data quality assessments completed during the data collection phase. We conducted a simulation experiment to evaluate the relationship between the number of records reviewed manually, the detection of true data errors (true positives) and the number of manual chart abstraction errors (false positives) that required unnecessary investigation. RESULTS: The semi-automated data quality assessments identified 79 data quality issues requiring correction, of which 14 had a large impact, affecting at least 50% of records in the data. After resolving issues identified through semi-automated assessments, manual validation of 2676 patient encounters at 19 hospitals identified 4 new meaningful data errors (3 in transfusion data and 1 in physician identifiers), distributed across 4 hospitals. There were 365 manual chart abstraction errors, which required investigation by data analysts to identify as "false positives." These errors increased linearly with the number of charts reviewed manually. Simulation results demonstrate that all 3 transfusion data errors were identified with 95% sensitivity after manual review of 5 records, whereas 18 records were needed for the physician's table. DISCUSSION AND CONCLUSION: The GEMINI approach represents a scalable framework for data quality assessment and improvement in multisite EHR research databases. Manual data review is important but can be minimized to optimize the trade-off between true and false identification of data quality errors.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.052 | 0.028 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.002 |
| Insufficient payload (model declined to judge) | 0.000 | 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; both teacher heads agree on what is shown here.
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".