Predicting Leptospirosis with Machine Learning
Author Information
Author(s): Muacevic Alexander, Adler John R, Sengupta Mallika, Kundu Aditya, Mandal Saikat, Chatterjee Shiv Sekhar, Ghoshal Ujjala, Banerjee Sayantan, Mukhopadhyay Kaushik
Primary Institution: All India Institute of Medical Sciences, Kalyani, IND
Hypothesis
Can machine learning models and geographic mapping improve the diagnosis of leptospirosis using baseline laboratory tests?
Conclusion
Geographic mapping identified clusters of leptospirosis cases, but no significant association was found between individual laboratory parameters and the disease diagnosis.
Supporting Evidence
- 43 out of 325 patients tested positive for leptospirosis.
- Geographic mapping revealed case clusters in West Bengal.
- Machine learning models showed moderate predictive accuracy with KNN achieving 74% accuracy.
Takeaway
The study used computer algorithms to help find cases of leptospirosis, a disease that can be hard to diagnose, but it didn't find strong links between lab tests and the disease.
Methodology
An observational retrospective study analyzing patients suspected of leptospirosis over one year using laboratory tests and machine learning models.
Potential Biases
The study was conducted at a single center, which may limit the generalizability of the findings.
Limitations
The study lacked detailed clinical data, which may have affected the models' ability to differentiate leptospirosis from other febrile illnesses.
Participant Demographics
325 patients, mean age 35.9 years, 53.5% female.
Statistical Information
P-Value
0.032
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website