Flexible scan statistic with a restricted likelihood ratio for optimized COVID-19 surveillance

Area of Science:

• Epidemiology
• Spatial Analysis
• Public Health

Background:

• Effective disease surveillance is vital for detecting new COVID-19 variants, as recommended by the WHO.
• Integrating COVID-19 surveillance with other respiratory diseases requires identifying high-risk areas, information often missing in developing nations.
• Ghana's routine analyses lacked spatial risk data for COVID-19 incidence and Case Fatality Rates (CFR).

Purpose of the Study:

• To uncover spatial patterns of COVID-19 incidence and CFR in Ghana using scan-statistic cluster analysis.
• To identify high-risk districts for targeted sentinel or genomic surveillance.
• To examine the influence of covariates on spatial clusters of COVID-19 incidence and CFR.

Main Methods:

• Employed flexible spatial scan statistic with restricted likelihood ratio for cluster analysis.
• Analyzed COVID-19 data from Ghana covering four pandemic waves (March 2020 - February 2022).
• Adjusted for covariates: distance to epicentre, population aged ≥ 65, male proportion, and urban proportion.

Main Results:

• Identified 56 significant spatial clusters for incidence and 26 for CFR across all four waves.
• Most Likely Clusters (MLCs) for incidence were in south-eastern Ghana; CFR clusters were in central and northeastern districts.
• Closeness to epicentre and high urban population proportion increased incidence; high proportion of ≥ 65 years increased CFR.

Conclusions:

• COVID-19 incidence and CFR are spatially clustered in Ghana, influenced by urban population, male proportion, elderly population, and proximity to epicentre.
• High-risk districts identified can serve as crucial sites for enhanced surveillance.
• Future control measures should integrate healthcare access improvements and urban population growth management.