Back

Prioritizing interventions for cholera control in Kenya, 2015-2020

Boru, W.; Xiao, S.; Amoth, P.; Kareko, D.; Langat, D.; Were, I.; Ali, M.; Sack, D. A.; Lee, E. C.; Debes, A. K.

2022-11-10 epidemiology
10.1101/2022.11.07.22281877 medRxiv
Show abstract

Kenya has experienced cholera outbreaks since 1971, with the most recent wave beginning in late 2014. Between 2015-2020, 32 of 47 counties reported 30,431 suspected cholera cases. The Global Task Force for Cholera Control (GTFCC) developed a Global Roadmap for Ending Cholera by 2030, which emphasizes the need to target multi-sectoral interventions to priority areas known as "cholera burden hotspots." This study utilizes the GTFCCs hotspot method to identify hotspots in Kenya at the county and sub-county administrative levels from 2015 through 2020. 32 of 47 (68.1%) counties reported cholera cases during this time while only 149 of 301 (49.5%) sub-counties reported cholera cases. The analysis identifies hotspots based on the mean annual incidence (MAI) over the past five-year period and choleras persistence in the area. Applying a MAI threshold of 90th percentile and the median persistence at both the county and sub-county levels, we identified 13 high risk sub-counties from 8 counties, including the 3 high risk counties of Garissa, Tana River and Wajir. This demonstrates that several sub-counties are high level hotspots while their counties are not. In addition, when cases reported by county versus sub-county hotspot risk are compared, 1.4 million people overlapped in the areas identified as both high-risk county and high-risk sub-county. However, assuming that finer scale data is more accurate, 1.6 million high risk sub-county people would have been misclassified as medium risk with a county-level analysis. Furthermore, an additional 1.6 million people would have been classified as living in high-risk in a county-level analysis when at the sub-county level, they were medium, low or no-risk sub-counties. This results in 3.2 million people being misclassified when county level analysis is utilized rather than a more-focused sub-county level analysis. This analysis highlights the need for more localized risk analyses to target cholera intervention and prevention efforts towards the populations most vulnerable.

Matching journals

The top 3 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.