Table 5.
Effects of PM2.5, temperature, and relative humidity on household entry, host-seeking, household exit, and mortality
| Parameter | β (95% CI) | Δ in AIC | Δ in R2 | |
|---|---|---|---|---|
| Household entry | PM2.5 (µg/m3) | 1.2 (–1.8 to 4.1) | −1.2 | 0.00 |
| Temperature | −13.1 (–16.2 to −10.0)*** | −62.7 | 0.09 | |
| Relative humidity | −0.7 (–3.9 to 2.5) | −0.9 | 0.00 | |
| Intercept | 44.2 (38.9 to 49.4) | |||
| Host-seeking | PM2.5 | −5.4 (–7.4 to −3.4)*** | −27.5 | 0.04 |
| Temperature | 0.9 (–1.3 to 3.1) | −0.7 | 0.01 | |
| Relative humidity | −0.4 (–2.9 to 2.2) | −0.3 | 0.00 | |
| Intercept | 21.6 (16.0 to 27.2) | |||
| Household exit | PM2.5 | 1.1 (0.2 to 2.0)* | −4.3 | 0.03 |
| Temperature | −3.3 (–4.2 to −2.3)*** | −40.6 | 0.16 | |
| Relative humidity | 0.3 (–0.6 to 1.0) | 1.5 | 0.01 | |
| Intercept | 4.5 (2.4 to 6.7) | |||
| Mortality | PM2.5 | 1.3 (0.0 to 2.6)* | −1.0 | 0.04 |
| Temperature | 0.8 (–0.5 to 2.2) | 1.5 | −0.01 | |
| Relative humidity | −0.3 (–1.5 to 0.8) | 2.9 | 0.00 | |
| Intercept | 6.6 (2.6 to 10.5) |
PM2.5 = real-time fine particulate matter. β values represent the linear change in each outcome for a one standard deviation increase in each predictor variable, controlling for all other independent variables. Variables are scaled for comparison. Δ in Akaike information criteria (AIC) shows relative change in model fit when each variable is added last to a full model. Negative values indicate improved model fit. Δ in R2 is the proportion of variance explained by each variable, calculated as the change in the conditional R2 value when each variable is added last to the full model. *P < 0.05; **P < 0.01; ***P < 0.001.