Skip to content

Latest commit

 

History

History
62 lines (31 loc) · 1.71 KB

README-OLD.md

File metadata and controls

62 lines (31 loc) · 1.71 KB
Raw

Team 5

Background and motivation

Renal failure happens for some physiological reason. Predict renal failure.

Ward level less populated with staff.

Serum creatinine - high level means kidney not working as well.

Those with blood tests usually show they are more sick.

We are Looking for people who develop AKI within the stay.

Inclusion Criteria

KDIGO Criteria.

Stage 1

Baseline = admission measurement.

  • If in any 48H, serum creatinine increases by > 26.5umol/L or 0.3mg/dL. NO flexibility for time.
  • Or within 7 days, serum creatinine increases x1.5 from the lowest value in that window.

Stage 2

  • Increases by x2 relative to the baseline that was used to trigger stage 1 x1.5 criteria.

Stage 3

  • Absolute value > 354umol/L

Or we can tweak criteria. If not enough regular measurements, getting measurements within same admission should be ok.

Or potentially used nadir creatinine (lowest measured value).

Exclusion

  • Exclude people with ESRD/ESKD/ESRF. If they are on dialysis, or ckd stitch who are not yet on dialysis.

  • Can analyze CKD people as a separate set. Later...

  • Looking for people who develop AKI within the stay. Looking for people who develop it from at least 48H after admission. So how to exclude those who got it within 48H?

Creatinine can improve with time, to imply function has improved. The lowest value in the entire admission can be a surrogate baseline. Can look back and see if the admission/any within 48H measured value triggers the x1.5 AKI criteria.

Questions

  • Prediction window

Covariates

  • Drugs
  • Acute conditions they are admitted for. ie. heart attack, acute myocardial infarction, pneumonia, cancer,
  • Age, sex, height, weight, race
  • Sepsis?