Extended Data Table 6 Summary statistics for the data

From: A clinically applicable approach to continuous prediction of future acute kidney injury

  1. A breakdown of training (80%), validation (5%), calibration (5%) and test (10%) datasets by both unique patients and individual admissions. Where appropriate, the percentage of total dataset size is reported in parentheses. The dataset was representative of the overall US Department of Veterans Affairs population for clinically relevant demographics and diagnostic groups associated with renal pathology.
  2. *Average age after taking into account exclusion criteria and statistical noise added to meet HIPAA Safe Harbour criteria.
  3. **Chronic kidney disease stage 1 is evidence of renal parenchymal damage with a normal glomerular filtration rate. This is rarely recorded in our dataset; instead, the numbers for stage-1 chronic kidney disease have been estimated from admissions that carried an ICD-9 code for chronic kidney disease, but for which the glomerular filtration rate was normal. For this reason, these numbers may underrepresent the true prevalence in the population.
  4. ***In total, 172 US Department of Veterans Affairs inpatient sites and 1,062 outpatient sites were eligible for inclusion. In addition, 130 data centres aggregated data from 1 or more of these facilities, of which 114 such data centres had data for inpatient admissions used in this study. Although the exact number of sites included was not provided in the dataset for this work, no patients were excluded on the basis of location.
Back to article page