The ALT-70 cellulitis model maintains predictive value at 24 and 48 hours after presentation

Area of Science:

• Emergency Medicine
• Clinical Diagnostics
• Predictive Modeling

Background:

• Cellulitis is frequently misdiagnosed, leading to increased hospitalizations and healthcare expenses.
• The ALT-70 predictive model provides an objective method for distinguishing cellulitis from similar conditions upon emergency department (ED) arrival.

Purpose of the Study:

• To assess the diagnostic performance of the ALT-70 predictive model at 24 and 48 hours after initial ED presentation.
• To determine if the ALT-70 model's accuracy is maintained over time for differentiating cellulitis.

Main Methods:

• A retrospective cohort review was conducted, incorporating data collected at 24 and 48 hours post-initial ED presentation.
• Classification metrics of the ALT-70 score were compared at three time points: initial ED presentation, 24 hours, and 48 hours.

Main Results:

• A statistically significant difference in median ALT-70 scores was observed between true cellulitis cases and mimickers at all evaluated time points.
• The sensitivity, specificity, positive predictive value, and negative predictive value of the ALT-70 score remained consistent across the initial presentation, 24-hour, and 48-hour assessments.

Conclusions:

• The ALT-70 model demonstrates sustained performance at 24 and 48 hours, comparable to its initial presentation accuracy.
• This consistent performance allows for broader application of the ALT-70 model in various clinical settings for cellulitis diagnosis.
• The single-center design may limit the generalizability of these findings.