Establishing an evidence base for critical laboratory value thresholds.

Area of Science:

• Clinical Pathology
• Laboratory Medicine
• Patient Safety

Background:

• Critical laboratory values signal potentially life-threatening patient conditions.
• The effectiveness of established critical value thresholds in predicting mortality requires rigorous evaluation.

Purpose of the Study:

• To assess the accuracy of critical laboratory value thresholds in identifying patients at risk of in-hospital death.
• To examine the relationship between the degree of test abnormality, time, and mortality.

Main Methods:

• Analysis of five years of inpatient admissions data (165,066 admissions, 872,503 test results).
• Evaluation of in-hospital mortality based on the time of critical or near-critical results and the severity of abnormality.

Main Results:

• Some critical value thresholds effectively identified mortality risk (e.g., elevated potassium), while others were suboptimal (e.g., elevated hematocrit, lactate).
• Mortality risk was often time-dependent, but not for all critical values (e.g., elevated activated partial thromboplastin time [APTT], international normalized ratio [INR], glucose).
• Prior hospital admission within 30 days was associated with lower mortality for certain critical values but not others.

Conclusions:

• A subset of laboratory critical value thresholds accurately predicts in-hospital mortality risk, with most demonstrating time urgency.
• The impact of prior hospital admissions on mortality risk varies across different critical values.