f0020: The relationship between the predicted rate of reported diarrheal cases and elevation in Ho Chi Minh City, 2008–2010. Scatterplot showing the estimated average reporting rate by district for 2008–2010 (with bars indicating 95% credible interval) and average district elevation in meters above sea-level. The numbers within the plot represent the point estimate for each of the corresponding districts. The dashed line represents the fitted linear model.

Mentions: In each year, the relative risk of reported diarrhea at the city level rose rapidly during the final weeks of the year, followed by a peak period lasting from 1 to 2 months. The citywide intercepts for each month of the study are highlighted in Fig. 3. Further, the elevation of each district was strongly related to the reported rate of diarrheal illness (Table 1 and Fig. 4). The RR of a one-meter increase in elevation was 0.95 (95%CI:0.92, 0.97). Therefore, the district with the lowest median elevation (0.32 m) experienced nearly double the risk of the highest district (7.95 m) (RR=0.57, 95%CI:0.31, 0.74). In further models incorporating both elevation and waterway coverage as fixed effects, the influence of waterways was non-significant, although this became significant when elevation was excluded. However, the inverse relationship between average district elevation and the proportion of district area covered by waterways (Pearson's R=−0.59, 95%CI:−0.80, −0.25) suggest that waterway coverage may explain some of the risk associated with lower elevation. Finally, district-level elevation was not associated with distance from CH1 (β=0.07, 95%CI=−2.76, 2.90), suggesting that adjustment for distance to the largest and most centrally located hospital does not confound the effect of elevation.