Assessing the quality of clinical and administrative data extracted from hospitals: The General Medicine Inpatient Initiative (GEMINI) experience
Bibliographic record
Abstract
Abstract Objective Large clinical databases are increasingly being used for research and quality improvement, but there remains uncertainty about how computational and manual approaches can be used together to assess and improve the quality of extracted data. The General Medicine Inpatient Initiative (GEMINI) database extracts and standardizes a broad range of data from clinical and administrative hospital data systems, including information about attending physicians, room transfers, laboratory tests, diagnostic imaging reports, and outcomes such as death in-hospital. We describe computational data quality assessment and manual data validation techniques that were used for GEMINI. Methods The GEMINI database currently contains 245,559 General Internal Medicine patient admissions at 7 hospital sites in Ontario, Canada from 2010-2017. We performed 7 computational data quality checks followed by manual validation of 23,419 selected data points on a sample of 7,488 patients across participating hospitals. After iteratively re-extracting data as needed based on the computational data quality checks, we manually validated GEMINI data against the data that could be obtained using the hospital’s electronic medical record (i.e. the data clinicians would see when providing care), which we considered the gold standard. We calculated accuracy, sensitivity, specificity, and positive and negative predictive values of GEMINI data. Results Computational checks identified multiple data quality issues – for example, the inclusion of cancelled radiology tests, a time shift of transfusion data, and mistakenly processing the symbol for sodium, “Na”, as a missing value. Manual data validation revealed that GEMINI data were ultimately highly reliable compared to the gold standard across nearly all data tables. One important data quality issue was identified by manual validation that was not detected by computational checks, which was that the dates and times of blood transfusion data at one site were not reliable. This resulted in low sensitivity (66%) and positive predictive value (75%) for blood transfusion data at that site. Apart from this single issue, GEMINI data were highly reliable across all data tables, with high overall accuracy (ranging from 98-100%), sensitivity (95-100%), specificity (99-100%), positive predictive value (93-100%), and negative predictive value (99-100%) compared to the gold standard. Discussion and Conclusion Iterative assessment and improvement of data quality based primarily on computational checks permitted highly reliable extraction of multisite clinical and administrative data. Computational checks identified nearly all of the data quality issues in this initiative but one critical quality issue was only identified during manual validation. Combining computational checks and manual validation may be the optimal method for assessing and improving the quality of large multi-site clinical databases.
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 | 0.126 | 0.253 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.004 | 0.008 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.004 | 0.002 |
| Open science | 0.003 | 0.005 |
| Research integrity | 0.001 | 0.001 |
| 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".