A computer program for the derivation of non-parametric reference ranges from patients' results

Area of Science:

• Clinical Chemistry
• Biostatistics
• Medical Informatics

Background:

• Establishing accurate reference ranges is crucial for diagnosing diseases.
• Current methods may be influenced by patient populations with varying health statuses.
• A robust method is needed to derive reference ranges from existing patient data.

Purpose of the Study:

• To present a novel computer program for deriving medical reference ranges.
• To develop a method that accounts for disease prevalence in patient data.
• To provide a non-parametric approach for calculating reference ranges.

Main Methods:

• A computer program modifies the two halves of a patient results frequency distribution curve.
• Linear regression extends the para-apical portion to represent disease-unaffected results.
• Non-parametric calculation derives the final reference range.

Main Results:

• The program successfully derives reference ranges from patient data.
• Calculated ranges closely matched existing laboratory ranges for common chemistries.
• The method demonstrated a way to adjust for disease prevalence.

Conclusions:

• The developed computer program offers a reliable method for deriving medical reference ranges.
• This approach provides a valuable tool for clinical laboratories.
• Further discussion on the method's advantages and limitations is warranted.