A delta check failure, or alert, is defined as a difference between a patient’s present laboratory result and their previous laboratory result. This difference must exceed a predefined limit, within a predefined amount of time. As an example, if two sodium results differ by more than 13 mEq/L within a 3 day time span, a delta check alert would be triggered. The technologist would be prompted to review further and (hopefully) determine whether this difference was due to preanalytical sampling issues, a true biological change in the patient, or possible sample misidentification.
Preanalytical issues may include dilution with IV fluids, inappropriately submitting a sample collected with anticoagulant, circadian rhythm, posture, etc. Depending on the analyte, a true change in the patient’s status may also be due to recent medical intervention (surgery, blood products, etc.). Possible causes of sample misidentification include improper patient identification during phlebotomy, identical or similar names, or fraud.
Hands down, the two most popular questions regarding delta checks are…which analytes should have delta checks applied to them, and what limits should I use?! Unfortunately, there is not a "one size fits all" answer…
Two commonly used tools to help determine delta check limits are the reference change value (RCV) and the index of individuality (II) (1). Briefly, these two metrics take into account analyte-specific analytical and biological variation, with the II specifically addressing the ratio between within- and between-individual variation. Primarily, we want to focus our delta check efforts on analytes with low within-individual variation that are relatively tightly controlled by the body (e.g., creatinine, alkaline phosphatase, bilirubin, mean corpuscular volume [MCV]).
A recent publication by Strathmann et al. (2) uses a computer modeling approach with historical data to predict whether delta check rules adequately identify mislabeled specimens. They found that MCV had the highest positive predictive value for mislabeling, with the fewest false positives. Furthermore, they propose that delta check limits assigned to traditional analytes such as sodium and potassium are unlikely to identify mislabeling events and question their overall utility. Applying multiple rules simultaneously may provide the most diagnostic utility: failing multiple delta check rules increases the likelihood that a sample has been mislabeled.
Questioning which analytes to apply delta check limits to, and what those limits are, brings up another critical point: delta check limits will vary among institutions and the populations that they serve. Does your lab perform testing for a dialysis clinic? A cancer hospital? The limits and analytes that will be helpful in a pediatric hospital will vary widely from those used to serve a primarily geriatric population. Strathmann’s report also touches on this critical point, concluding that delta check rules were not commutable between different institutions. They give an example of hematocrit delta check rules not performing well in a cancer and/or transplant-focused facility. Similarly, delta check limits for renal function tests (creatinine, urea nitrogen) would not serve a dialysis or renal transplant population well.
Unfortunately, as you can see, there are no easy answers to the question "what limits should I use for my delta checks?" The issues outlined above are just a few that should be considered when determining the utility of delta check limits in your institution.