Detecting chronic kidney disease in population-based administrative databases using an algorithm of hospital encounter and physician claim codes
Bibliographic record
Abstract
BACKGROUND: Large, population-based administrative healthcare databases can be used to identify patients with chronic kidney disease (CKD) when serum creatinine laboratory results are unavailable. We examined the validity of algorithms that used combined hospital encounter and physician claims database codes for the detection of CKD in Ontario, Canada. METHODS: We accrued 123,499 patients over the age of 65 from 2007 to 2010. All patients had a baseline serum creatinine value to estimate glomerular filtration rate (eGFR). We developed an algorithm of physician claims and hospital encounter codes to search administrative databases for the presence of CKD. We determined the sensitivity, specificity, positive and negative predictive values of this algorithm to detect our primary threshold of CKD, an eGFR <45 mL/min per 1.73 m² (15.4% of patients). We also assessed serum creatinine and eGFR values in patients with and without CKD codes (algorithm positive and negative, respectively). RESULTS: Our algorithm required evidence of at least one of eleven CKD codes and 7.7% of patients were algorithm positive. The sensitivity was 32.7% [95% confidence interval: (95% CI): 32.0 to 33.3%]. Sensitivity was lower in women compared to men (25.7 vs. 43.7%; p <0.001) and in the oldest age category (over 80 vs. 66 to 80; 28.4 vs. 37.6 %; p < 0.001). All specificities were over 94%. The positive and negative predictive values were 65.4% (95% CI: 64.4 to 66.3%) and 88.8% (95% CI: 88.6 to 89.0%), respectively. In algorithm positive patients, the median [interquartile range (IQR)] baseline serum creatinine value was 135 μmol/L (106 to 179 μmol/L) compared to 82 μmol/L (69 to 98 μmol/L) for algorithm negative patients. Corresponding eGFR values were 38 mL/min per 1.73 m² (26 to 51 mL/min per 1.73 m²) vs. 69 mL/min per 1.73 m² (56 to 82 mL/min per 1.73 m²), respectively. CONCLUSIONS: Patients with CKD as identified by our database algorithm had distinctly higher baseline serum creatinine values and lower eGFR values than those without such codes. However, because of limited sensitivity, the prevalence of CKD was underestimated.
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 | 0.012 | 0.044 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.006 | 0.005 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.000 |
| 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".