A bayesian approach to laboratory utilization management.

Area of Science:

• Clinical laboratory science
• Health services research
• Epidemiology

Background:

• Laboratory utilization management aims to enhance healthcare value by refining laboratory test orders.
• Current rule-based systems for auditing lab orders are inflexible and can conflict with clinical decision-making and physician autonomy.
• Limitations of rule-based approaches include poor adaptation to medical ambiguity and focus on order context rather than patient outcomes.

Purpose of the Study:

• To develop and demonstrate an alternative to rule-based laboratory utilization management.
• To introduce a novel method for objective evaluation of laboratory test requests.
• To improve the efficiency and effectiveness of laboratory utilization audits.

Main Methods:

• Developed a model based on an epidemiological formula relating disease prevalence, test proportion positive, sensitivity, and specificity.
• Applied the model to estimate pretest probability (PTP) of disease in audited patient cohorts.
• Utilized PTP comparisons across providers and patient groups for objective laboratory request evaluation, demonstrated with enterovirus (EV) meningitis testing.

Main Results:

• The model successfully identified patient subpopulations with a low prevalence of disease.
• These low-prevalence groups exhibited demographic and seasonal factors associated with lower risk for EV meningitis.
• Findings suggest that laboratory test orders may be excessive for patients at low risk for EV.

Conclusions:

• Introduced a new, objective method for auditing laboratory services within laboratory utilization management programs.
• The proposed approach offers a more nuanced and data-driven alternative to traditional rule-based systems.
• This method facilitates better identification of appropriate laboratory test utilization.