Multivariate strategies for screening evaluation of harmful drinking

Area of Science:

• Biostatistics
• Clinical Chemistry
• Public Health

Background:

• Harmful alcohol consumption poses a significant public health challenge.
• Current screening programs rely on traditional data interpretation methods for evaluating alcohol intake.
• There is a need for improved diagnostic strategies for identifying harmful drinking patterns in large populations.

Purpose of the Study:

• To introduce and evaluate the unequal dispersed classes (UNEQ) chemometric strategy for assessing harmful alcohol drinking.
• To compare the performance of the UNEQ approach with traditional methods in population screening.
• To determine the utility of five indirect biomarkers in identifying harmful alcohol consumption.

Main Methods:

• A cohort of 423 patients was analyzed.
• Five inexpensive indirect biomarkers were measured: aspartate aminotransferase, alanine aminotransferase, γ-glutamyltransferase, mean corpuscular volume, and carbohydrate-deficient transferrin.
• The chemometric UNEQ strategy was applied and compared to traditional interpretation methods.

Main Results:

• The multivariate UNEQ approach significantly improved the diagnostic performance of indirect biomarkers for harmful drinking.
• Reliable decision rules were established with a reduction in doubtful classifications.
• The UNEQ model demonstrated superior accuracy in distinguishing between low-risk and harmful drinkers.

Conclusions:

• The UNEQ model offers an innovative and generalizable approach for clinical evaluation.
• This strategy efficiently extracts information from individual biomarkers to create a synthetic multivariate parameter.
• The developed parameter can be directly integrated into diagnostic protocols for improved harmful drinking assessment.