link.growkudos.com/1e4i5naqx34

Now published: journals.plos.org/plosntds/art...

Well done! Thanks @drleshan.bsky.social

📝 Read our full preprint for comprehensive insights into leptospirosis risk prediction and the complex interplay of environmental and socioeconomic factors driving outbreaks in Thailand: ⤵️ doi.org/10.1101/2025...

Our approach demonstrates how machine learning can help unravel complex disease drivers when traditional modeling approaches struggle with highly correlated factors and limited data resolution.

😷 We also documented how COVID-19 disrupted leptospirosis surveillance in Thailand, with model performance declining during the pandemic (2020-2021) but recovering in 2022. This suggests significant underreporting during the pandemic years.

🌧️ While previous studies focused heavily on rainfall, our analysis revealed more complex climate interactions. Vapor pressure, maximum temperature, and precipitation during the driest month all influence outbreak patterns in different ways.

👨‍👩‍👧‍👦 Beyond agriculture, larger household size emerged as a critical risk factor, indicating leptospirosis disproportionately affects rural communities. Understanding these socioeconomic dimensions is crucial for targeted interventions.

🌾 Surprisingly, we found that rice production factors were the strongest predictors of leptospirosis risk. Traditional farming practices appear more conducive to disease transmission compared to mechanized methods, highlighting agriculture's role in outbreak dynamics.

🦠 Leptospirosis poses a significant public health challenge in Thailand, with complex transmission patterns influenced by rice farming, climate, and socioeconomic conditions. Our XGBoost model achieved high predictive accuracy (AUC>0.93) in identifying high-risk provinces.

Community surveillance of discarded condoms might help track the mpox transmission: www.thelancet.com/journals/lan...