Objective:

To clarify methodological and programmatic aspects of the study “Deterioration of Immunization Resilience: A Bayesian Machine Learning Assessment of the 2025-2026 Measles Outbreak in Mexico.”

Approach:

• Confidence vs credibility interval: The authors acknowledge that the reported Bayesian interval is a posterior credibility interval. Its width, 64.45% to 100.00%, reflects substantial uncertainty regarding immunization coverage recovery.
• Negative R²: The authors acknowledge that the out-of-sample R² of −0.36 indicates poor predictive fit for the 2025 data point and performance worse than a horizontal mean line.
• Predictive modeling vs programmatic perspective: The response explains that models based on historical data assume continuity and cannot readily account for abrupt interventions. The model was intended to establish a baseline risk scenario without extraordinary public health measures.

Key Findings:

• Complete immunization recovery remains mathematically possible because the interval includes the World Health Organization’s 95% threshold, but exceptional public health measures would be required.
• The 2025 outbreak represented a drastic departure from the preceding 10-year trend.
• Emergency catch-up campaigns, international collaborations, and geographically targeted interventions could alter the modeled trajectory.

Interpretation:

Predictive machine learning assessments should be integrated with programmatic realities when developing policies to address immunization gaps in Mexico.

Limitations:

• The model’s poor fit for the 2025 data point limits its reliability for forecasting 2026.
• Retrospective smoothing cannot adequately capture the 2025 outlier, contributing to the exceptionally wide credibility interval.

Conclusion:

The response clarifies that the model estimates baseline risk rather than the effects of extraordinary interventions intended to disrupt outbreak progression.

Sources:

• International Journal of Infectious Diseases