Barriers to Effective Snakebite Treatment in Ghana
Author Information
Author(s): Nyarko Eric, Agyemang Edmund Fosu, Ameho Ebenezer Kwesi, Agyekum Louis, Gutiérrez José María, Fernandez Eduardo Alberto
Primary Institution: University of Ghana
Hypothesis
This study aimed to assess the barriers to effective snakebite treatment in Ghana using machine learning and statistical design.
Conclusion
The study identified key barriers to effective snakebite treatment in Ghana, including high costs of antivenoms and lack of access in remote areas.
Supporting Evidence
- Snakebite envenoming affects 2.5 million people and causes 81,000–138,000 deaths annually.
- The World Health Organization aims to halve the deaths related to snakebite by 2030.
- High costs and lack of access to antivenoms are significant barriers identified in the study.
- Machine learning models were used to predict barriers to effective treatment.
- The study involved 203 healthcare workers from Ghana.
- MaxDiff design was used to gauge the relative importance of barriers.
- Results showed no single machine learning model was superior, but Ridge regression performed best.
- Addressing identified barriers can enhance snakebite treatment effectiveness.
Takeaway
This study looked at why treating snakebites is hard in Ghana and found that things like expensive medicine and not having it available in remote areas make it tough.
Methodology
The study used a MaxDiff statistical experiment design and six supervised machine learning models to analyze data from healthcare workers.
Potential Biases
The study may not fully represent the views of all stakeholders involved in snakebite treatment.
Limitations
The study was limited to a specific geographical area and primarily focused on healthcare workers' perspectives.
Participant Demographics
Participants included community healthcare workers with experience in treating snakebite victims.
Statistical Information
P-Value
0.000
Confidence Interval
95% CI: 0.303, 0.486
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website