Application of a Weighted Absolute Percentage Error-Based Method for Calculating the Aggregate Accuracy of Reported...

Area of Science:

• Public Health
• Epidemiology
• Health Informatics

Background:

• Accurate malaria surveillance data is crucial for effective control programs.
• Routine Data Quality Audits (RDQA) are standard for assessing data quality in health facilities.
• Current RDQA aggregate reporting accuracy methods have limitations in identifying all low-quality data.

Purpose of the Study:

• To highlight the limitations of existing aggregate reporting accuracy methods in malaria RDQA toolkits.
• To introduce and validate a novel Weighted Absolute Percentage Error-based Aggregate Data Reporting Accuracy (WADRA) approach.
• To demonstrate WADRA's ability to detect low-accuracy facilities missed by current methods.

Main Methods:

• Analysis of three example scenarios illustrating limitations of current malaria RDQA toolkits.
• Development and application of the WADRA method, utilizing register values as weighting factors.
• Comparison of WADRA's performance against existing aggregate reporting accuracy metrics.

Main Results:

• Current aggregate reporting accuracy methods can misclassify low-accuracy facilities as high-accuracy.
• The WADRA approach effectively identifies low-accuracy facilities by incorporating register values.
• WADRA enhances the sensitivity of data quality audits.

Conclusions:

• The WADRA method offers a significant improvement over existing aggregate reporting accuracy measures in malaria surveillance.
• Adoption of WADRA by country malaria programs can lead to more precise identification of data quality issues.
• Improved detection of low-accuracy facilities facilitates better resource allocation and targeted interventions for data quality improvement.