A model for assessing interference.

Most studies of analytical interference indicate the magnitude but not the type of interference. We developed a model for interference that assesses the magnitude of the following types of interference: independent of analyte concentration, dependent on analyte concentration, and a combination of the two. The experimental design for the model is an orthogonally arranged matrix with progressively increasing concentrations of analyte and interfering agent. Multiple regression with these independent variables (concentration of analyte, concentration of interfering agent, and the product of the two) is used to determine the magnitude, direction, and significance of each type of interference. Applying the model to study interference by hemoglobin with determinations of creatinine, direct bilirubin, and total bilirubin showed that hemoglobin interferes with determination of creatinine independent of the analyte concentration, interferes with total bilirubin dependent on the analyte concentration, and interferes with direct bilirubin by a combination of these two.