GeoMosquito data
1. Executive Summary: The “Centrality” Risk
The primary finding of this analysis is the significant role of urban centrality. Children living or attending school in “Central+” areas (characterized by high connectivity and density) show a substantially higher exposure to dengue compared to those in peripheral or “Central-” areas. The classification of “Central+” and “Central-“ areas is a way to categorize urban locations not just by their geography, but by their functional risk within the epidemic network.
Central+ Areas (High Connectivity & Density)
These areas represent the “hubs” of Bangkok’s mobility system. They are characterized by:
- High Population Density: Districts like Din Daeng, where the concentration of people creates an “explosive” transmission potential.
- Structural Connectivity: These locations act as primary bridges or anchors in the transportation network, meaning they have a high frequency of people moving in and out daily.
- Increased Risk: Children living in these areas have 1.78 times higher odds of having been exposed to dengue compared to other areas.
- Mosquito Pressure: These areas show a trend towards higher mosquito densities (average of 5.77 mosquitoes per home compared to ~4.0 in other categories).
Central- Areas (Low Connectivity & Sparse Density)
These areas are typically located in the suburban or residential “belts” surrounding the city core. They are characterized by:
- Lower Spatial Interaction: They are less “central” to the city’s daily movement, meaning the virus travels there less frequently through human mobility.
- Stochastic Data: Because population and transmission are lower, the data from these areas is often “sparse” or noisy, requiring different statistical models (like Poisson models) to estimate risk.
- Lower Seroprevalence: In these areas, only about 25% to 27% of children tested positive for IgG antibodies, compared to 30% to 33% in Central+ areas
Key Seroprevalence Results (IgG Positive)
The study utilized GLMM Binomial regression to identify risk factors for prior infection:
- Home Location: Living in a Central+ area increases the odds of being seropositive by 1.78 times.
- School Location: Attending a school in a Central+ area increases the odds of being seropositive by 1.77 times.
- Age: While age shows a positive trend with exposure (P=0.059), the location effects are much more statistically dominant (P<0.001).
2. Statistical Data Tables
The following tables provide the raw evidence used to build our predictive models.
Table 1: Risk Category Analysis (Home vs. School)
We combined Home and School categories to see where the highest risk overlaps.
| Category | Meaning | N | Proportion IgG+ | Upper/Lower CI |
| AA | Central Home & Central School | 241 | 0.297 | 0.058 |
| AC | Central Home & NOT Central School | 79 | 0.405 | 0.108 |
| CA | NOT Central Home & Central School | 172 | 0.180 | 0.057 |
| CC | NOT Central Home & NOT Central School | 445 | 0.293 | 0.042 |
Table 2: Comparative T-Tests between Groups
Significant differences were found when comparing specific living/schooling arrangements.
| Comparison | T-Value | P-Value | Significance |
| AA vs CA | 2.66 | < 0.01 | Significant |
| AC vs CA | 3.71 | < 0.001 | Highly Significant |
| CA vs CC | -2.80 | < 0.01 | Significant |
3. Mosquito Population Dynamics
We monitored Aedes aegypti density in both schools and homes to correlate vector presence with human infection.
School Environment
While weekly variation was highly significant (P<0.001), the differences between school types were less pronounced in the raw counts.
| Class | Mean Aedes Count | S.E.M. |
| Central+ / Env+ | 3.3 | 0.209 |
| Central+ / Env- | 4.21 | 0.207 |
| Central- / Env+ | 3.1 | 0.181 |
| Central- / Env- | 2.76 | 0.256 |
Home Environment
Data suggests that “Central+” homes tend toward higher mosquito densities, reaching near significance (P=0.075).
| Home Class | Means (Log) | Backtransformed Mean |
| A (High Centrality) | 1.752 | 5.768 |
| B | 1.398 | 4.048 |
| C | 1.372 | 3.943 |
| D | 1.395 | 4.036 |